Scholarly article on topic 'A possible high Nb/Ta reservoir in the continental lithospheric mantle and consequences on the global Nb budget – Evidence from continental basalts from Central Germany'

A possible high Nb/Ta reservoir in the continental lithospheric mantle and consequences on the global Nb budget – Evidence from continental basalts from Central Germany Academic research paper on "Earth and related environmental sciences"

CC BY-NC-ND
0
0
Share paper
Keywords
{}

Abstract of research paper on Earth and related environmental sciences, author of scientific article — Jörg A. Pfänder, Stefan Jung, Carsten Münker, Andreas Stracke, Klaus Mezger

Abstract Compared to chondrites the accessible silicate reservoirs on Earth (i.e., mantle and continental crust) are depleted in Nb as expressed in their relatively low Nb/Ta. Although it was postulated that the “missing Nb” may be stored within a hidden reservoir in the mantle or within the Earth’s core, the role of the subcontinental lithospheric mantle in balancing the global Nb budget remains unclear. Continental basalts are pooled melts that tap the compositional spectrum of the subcontinental lithospheric mantle, and alkaline basalts from Central Germany are typical representatives of such melts. Here we present high-precision concentration data of Zr, Hf, Nb, Ta, and Lu determined by isotope-dilution MC–ICPMS along with Hf isotope compositions in a variety of intracontinental volcanic rocks from different locations in Central Germany. These rocks display Nb/Ta ratios (15.0–19.1) that still lie below the chondritic value (19.9±0.6) but are distinctly higher than in ocean-island basalts (15–16) that share similarly enriched mantle sources. They are also higher than Nb/Ta in the continental crust (12–13) and in the bulk-silicate Earth (BSE: ∼14), and therefore imply that the subcontinental lithospheric mantle also has high Nb/Ta and could potentially balance the Nb deficit observed in most terrestrial silicate reservoirs. Trace element modelling indicates that the HFSE composition of the continental basalts cannot be explained by simple melting of asthenospheric garnet or spinel peridotite sources, but requires the presence of metasomatised mantle domains that have been re-enriched by low-degree melts. Positively correlated Nb/Ta and Lu/Hf along with low Zr/Nb and Zr/Sm provide strong evidence that these low-degree melts may have a carbonatitic affinity and that the slight Nb-excess observed in continental basalts results primarily from carbonatite assimilation within the subcontinental lithospheric mantle. This is consistent with the evidence for carbonatite metasomatism found in xenoliths from Central Germany. Hafnium isotope modelling indicates that carbonatite metasomatism occurred long before the onset of Cenozoic magmatism (>100Ma). Our results suggest that the lithospheric mantle may host some of the “missing Nb”, but high-Nb/Ta domains are likely restricted to regions that have been affected by carbonatite metasomatism. Although Nb concentrations can be extremely high in carbonatites, such domains are probably a more local phenomenon and volumetrically too small to account as a whole for the global Nb deficit. Model calculations, however, indicate, that up to ∼30% of the “missing Nb” may be hosted in the subcontinental lithospheric mantle. The Earth’s core or any other hidden reservoir in the deep mantle remains significant in balancing the global Nb budget, but their role may be less important than previously thought.

Academic research paper on topic "A possible high Nb/Ta reservoir in the continental lithospheric mantle and consequences on the global Nb budget – Evidence from continental basalts from Central Germany"

Available online at www.sciencedirect.com

SciVerse ScienceDirect

Geochimica et Cosmochimica Acta 77 (2012) 232-251

www.elsevier.com/locate/gca

A possible high Nb/Ta reservoir in the continental lithospheric mantle and consequences on the global Nb budget - Evidence from continental basalts from Central Germany

Jorg A. Pfander a'c'*, Stefan Jungb, Carsten Münkera,d, Andreas Strackea,

Klaus MezgeraJ

a Institut für Mineralogie, Universität Münster, Corrensstr. 24, 48149 Münster, Germany Mineralogisch-Petrographisches Institut, Universität Hamburg, Grindelallee 48, 20146 Hamburg, Germany c Geologisches Institut, TU Bergakademie Freiberg, Bernhard-von-Cotta Str. 2, 09599 Freiberg, Germany d Institut für Geologie und Mineralogie, Universität zu Köln, Zülpicher Str. 49A, 50674 Köln, Germany

Received 11 January 2011; accepted in revised form 3 November 2011; available online 9 November 2011

Abstract

Compared to chondrites the accessible silicate reservoirs on Earth (i.e., mantle and continental crust) are depleted in Nb as expressed in their relatively low Nb/Ta. Although it was postulated that the "missing Nb" may be stored within a hidden reservoir in the mantle or within the Earth's core, the role of the subcontinental lithospheric mantle in balancing the global Nb budget remains unclear. Continental basalts are pooled melts that tap the compositional spectrum of the subcontinental lithospheric mantle, and alkaline basalts from Central Germany are typical representatives of such melts. Here we present high-precision concentration data of Zr, Hf, Nb, Ta, and Lu determined by isotope-dilution MC-ICPMS along with Hf isotope compositions in a variety of intracontinental volcanic rocks from different locations in Central Germany. These rocks display Nb/Ta ratios (15.0-19.1) that still lie below the chondritic value (19.9 ±0.6) but are distinctly higher than in ocean-island basalts (15-16) that share similarly enriched mantle sources. They are also higher than Nb/Ta in the continental crust (12-13) and in the bulk-silicate Earth (BSE: ~14), and therefore imply that the subcontinental lithospheric mantle also has high Nb/Ta and could potentially balance the Nb deficit observed in most terrestrial silicate reservoirs.

Trace element modelling indicates that the HFSE composition of the continental basalts cannot be explained by simple melting of asthenospheric garnet or spinel peridotite sources, but requires the presence of metasomatised mantle domains that have been re-enriched by low-degree melts. Positively correlated Nb/Ta and Lu/Hf along with low Zr/Nb and Zr/Sm provide strong evidence that these low-degree melts may have a carbonatitic affinity and that the slight Nb-excess observed in continental basalts results primarily from carbonatite assimilation within the subcontinental lithospheric mantle. This is consistent with the evidence for carbonatite metasomatism found in xenoliths from Central Germany. Hafnium isotope modelling indicates that carbonatite metasomatism occurred long before the onset of Cenozoic magmatism (>100 Ma).

Our results suggest that the lithospheric mantle may host some of the "missing Nb", but high-Nb/Ta domains are likely restricted to regions that have been affected by carbonatite metasomatism. Although Nb concentrations can be extremely high in carbonatites, such domains are probably a more local phenomenon and volumetrically too small to account as a whole for the global Nb deficit. Model calculations, however, indicate, that up to ~30% of the "missing Nb" may be hosted in the subcontinental lithospheric mantle. The Earth's core or any other hidden reservoir in the deep mantle remains significant in balancing the global Nb budget, but their role may be less important than previously thought. © 2011 Elsevier Ltd. All rights reserved.

* Corresponding author. Tel.: +49 3731 393811; fax: +49 3731 393599.

E-mail address: pfaender@tu-freiberg.de (J.A. Pfander). 1 Present address: Institut fur Geologie, Universitat Bern, Baltzerstrasse 1+3, 3012 Bern, Switzerland.

0016-7037/$ - see front matter © 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.gca.2011.11.017

1. INTRODUCTION

Subchondritic Nb/Ta (<19.9) in virtually all accessible silicate Earth reservoirs (i.e., mantle and continental crust), determined by high-precision high-field strength element (HFSE) concentration measurements, show that the silicate portion of the Earth is depleted in Nb relative to chondrites (Barth et al., 2000; Rudnick et al., 2000; Miinker et al., 2003; Weyer et al., 2003; Pfander et al., 2007; Willbold et al., 2009). Various models have been proposed to explain this so called "Nb-paradox" and they include hidden reservoirs of formerly subducted material in the silicate Earth (Barth et al., 2000; Rudnick et al., 2000; Nebel et al., 2010) or Nb fractionation into the metal core during or shortly after Earth's core formation (Wade and Wood, 2001; Jochum et al., 2002; Miinker et al., 2003; Mann et al., 2009). Niobium becomes slightly siderophile at high pressure (Wade and Wood, 2001) and thus the core is a potential reservoir for the missing Nb. Independent of which of these models is favoured, only little attention was paid to the HFSE budget of the subcontinental lithosphere, which may also play a role in balancing the global Nb budget. Evidence for high-Nb/Ta domains in the subcontinental lithospheric mantle (SCLM) comes from conventional trace element measurements of rutile-bearing eclogite and peridotite xenoliths, in which apparently elevated Nb/Ta may result from metasomatism within the lithospheric mantle (Ionov et al., 1993, 2002; Kalfoun et al., 2002; Aulbach et al., 2008).

Intracontinental volcanic rocks are important for studying the geochemical composition of the subcontinental lithospheric mantle and the underlying asthenosphere. Although continental basalt-hosted xenoliths provide direct petrological and geochemical information on the composition of the lithospheric mantle, this information is of very local significance and in most cases likely not directly linked to the melting region. In contrast, the volcanic rocks themselves, although potentially modified by post-melting processes during ascend, are pooled melts and thus provide a more robust information on the melting region and hence, indirectly, on average source compositions within the asthenosphere and lithosphere. This study presents high-precision HFSE and Lu concentration data along with Hf isotope compositions of continental basalts from the Rhon, Eifel and Vogelsberg regions in Central Germany. The origin of these mostly primitive continental volcanic rocks is commonly attributed to thermal anomalies in the astheno-spheric mantle beneath Central Europe (Wilson and Dow-nes, 1992; Hoernle et al., 1995; Goes et al., 1999; Wedepohl and Baumann, 1999; Ritter et al., 2001; Lustrino and Wilson, 2007) with a geochemical signature resembling that of ocean island basalts (Worner et al., 1986; Hegner et al., 1995; Jung, 1995; Jung and Masberg, 1998; Jung and Hoer-nes, 2000; Blusztajn and Hegner, 2002; Jung et al., 2005; Lustrino and Wilson, 2007). Continental basalts and their oceanic counterparts are therefore expected to have similar Zr/Hf and Nb/Ta, although crustal contamination can be important in some areas (Jung and Masberg, 1998; Jung and Hoernes, 2000; Bogaard and Worner, 2003; Haase et al., 2004; Jung et al., 2006). The HFSE, Lu/Hf, and Hf

isotope compositions of oceanic basalts are fairly well known (Pfander et al., 2007 and references therein). Any resolvable deviation in the HFSE budget between oceanic and continental basalts therefore constrains the processes that affected the composition of primary melts within the subcontinental lithospheric mantle and overlying crust and hence provides insights into the geochemical composition and evolution of the subcontinental lithospheric mantle over time. The high-precision data presented here also resolve the contributions of various asthenospheric and lithospheric mantle sources and constrain the roles of metasomatism and assimilation processes within the lithospheric mantle and overlying crust.

2. GEOLOGICAL FRAMEWORK AND SAMPLES

Volcanic activity in the Central European Volcanic Province (CEVP) took mostly place in the Miocene. In Central Germany, volcanic edifices from West to East include the Eifel, Siebengebirge, Westerwald, Vogelsberg, Hessian Depression, Rhon, Heldburg and Oberpfalz areas (Fig. 1), where predominantly basanites, alkali olivine basalts, olivine nephelinites, melilitites and their fractionation products erupted. In the Hessian Depression and the Vogelsberg area, olivine-tholeiites and quartz-tholeiites occur also. To the South of these large volcanic edifices, the volcanic region of Southern Germany with the Kaiserstuhl, Urach and Hegau volcanic fields produced more undersat-urated, bimodal melilititic-phonolitic suites. K/Ar ages indicate volcanic activity at around 18-10 Ma in the Vogelsberg region (Lippolt, 1982), which is the largest European volcanic edifice. More recent 40Ar/39Ar data constrain the age of the volcanic activity in the Vogelsberg to c.

17 Ma (Bogaard and Worner, 2003). In the Eifel, Tertiary volcanic activity is recorded between 44 and 35 Ma and around 25 Ma (Fekiacova et al., 2003) and in the Rhon area from 26 to 11 Ma with a major phase between 22 and

18 Ma (Lippolt, 1982). After volcanism terminated about 5 Ma ago, late Quaternary recurrence was restricted to the Eifel region between 650,000 and 10,800 years BP (Lippolt et al., 1990).

Based on geochemical data and seismic tomography, the magmatic activity in Europe is mostly attributed to litho-spheric extension and asthenospheric upwelling in the upper part of the mantle, expressed by low p-wave velocity zones that locally extend to depths of ~400 km beneath the Eifel and Massif Central (Granet et al., 1995; Hoernle et al., 1995; Goes et al., 1999; Grunewald et al., 2001; Ritter et al., 2001; Wilson and Patterson, 2001; Keyser et al., 2002; Lustrino and Wilson, 2007). Goes et al. (1999) have postulated that several small columnar thermal structures extend to upper mantle levels from a large-scale thermal anomaly in the lower mantle beneath Central Europe, suggesting that European intraplate volcanism locally taps lower mantle regions. On the basis of noble gas compositions of ultramafic xenoliths from alkaline volcanic rocks from the Pannonian basin and the Eifel, Buikin et al. (2005) inferred the presence of fairly undegassed, possibly deep mantle components in European intraplate volcanic rocks, consistent with the tomographic findings. Trace element and Nd-Sr-Pb

Vogelsberg 24-9

Westerwald 28-5 Siebengebirge 28-6

Eifel 45-24 0.5 - 0.01

Hessian Depression 20-8

Rhön 26-11

Upper Palatinate 29-19

Heldburg 42-11

Lippolt (1982).

isotope compositions of primitive lavas from the CEVP resemble those of ocean-island basalts (OIB). They are enriched in incompatible trace elements, have intermediate Nd-Sr-Pb isotope compositions similar to FOZO (see compilation of Lustrino and Wilson, 2007; FOZO: focal zone, see Stracke et al., 2005) and are commonly suggested to be of asthenospheric origin with minor lithospheric contributions (Worner et al., 1986; Blusztajn and Hart, 1989; Hegner et al., 1995; Jung and Masberg, 1998; Wedepohl and Baumann, 1999; Jung and Hoernes, 2000; Bogaard and Worner, 2003; Haase et al., 2004; Jung et al., 2006).

Although large scale decompression melting of astheno-spheric mantle is required to produce the voluminous occurrences of Cenozoic volcanic rocks in Central Germany (Jung et al., 2006), the role of the subcontinental lithosphere in the petrogenesis of these lavas is an ongoing matter of debate. For example, negative anomalies of Rb, K and partly Ba in mantle-normalised trace element diagrams observed in some lavas (e.g. from the Eifel, Vogelsberg and Rhon areas) have been interpreted as a lithospheric signature as they indicate the presence of residual hydrous phases such as phlogopite and amphibole (K-richterite) in the melt source region (Bogaard and Worner, 2003; Jung et al., 2006). These phases are stable in the relatively cold lithosphere up to pressures of ~6.5 GPa at a

maximum temperature of 1100 0C (Mengel and Green, 1986; Konzett and Ulmer, 1999), but not in a hot thermal plume (Class and Goldstein, 1997). For the melilitites of Southern Germany, a predominantly lithospheric origin has been proposed on the basis of prominent negative K and Rb anomalies as well as prevailing unradiogenic Os isotope compositions (Hegner et al., 1995; Blusztajn and Hegner, 2002). Therefore, a two stage mechanism involving primary melting in the asthenospheric mantle followed by subsequent melting of hydrated amphibole/phlogopite bearing metasomatised lithospheric mantle appears likely for most of the lavas of the CEVP (Wilson and Downes, 1992; Jung et al., 2005, 2011; Lustrino and Wilson, 2007).

Samples analysed during this study are from the Eifel, Rhon and Vogelsberg areas. Major- and trace element, as well as Nd-Sr and Os isotope data have been published previously (Jung and Masberg, 1998; Jung and Hoernes, 2000; Jung et al., 2005, 2006, 2011; see data compilation in the Electronic Annex). Rocks from the Eifel and Vogelsberg regions are basanites and nephelinites with MgO from 11.2 to 16.3 wt.% and Ni contents from 213 to 531 ppm (Jung and Masberg, 1998; Jung et al., 2006). The Vogelsberg samples also comprise olivine- and quartz-tholeiites with MgO from 7.4 to 10.1 wt.% and Ni contents from 88 to 230 ppm (Jung and Masberg, 1998). Samples from the Rhon comprise

basanites, nephelinites, alkali-olivine-basalts and two hornblende-basalts that range in MgO from 8.6 to 12.3 wt.% with Ni contents from 83 to 447 ppm (Jung and Hoernes, 2000). Basanites and nephelinites from all regions mostly represent near primary melts with some samples that underwent minor fractionation of olivine and clinopyroxene, whereas the tholeiites from the Vogelsberg are more evolved and crystallized olivine, clinopyroxene and minor plagioclase prior to eruption (Jung and Masberg, 1998; Bogaard and Worner, 2003).

3. ANALYTICAL METHODS

HFSE and Lu concentrations were determined by isotope-dilution (except Nb) using the Micromass Isoprobe multi-collector ICP-MS (inductively coupled plasma mass-spectrometer) at the University of Miinster. HFSE and Lu separation and purification followed the procedures described in Munker et al. (2001) and Weyer et al. (2002). Sample powders prepared in agate mills were spiked with a mixed 180Ta-180Hf-176Lu-94Zr tracer prior to digestion in concentrated HF-HNO3 on a hot plate (~120 0C) for several days. After drying down three times with concentrated HNO3 containing some trace HF (<0.05 M), and complete dissolution in hot 6 M HCl (containing trace HF), the evaporated chloride residues were redissolved in 3 M HCl for column separation on Ln-Spec® and anion-ex-change resins.

Niobium was determined by internal standardization from measured 93Nb/90Zr ratios in an aliquot taken after HFSE separation on a Ln-Spec® resin column. Calibration of the Zr/Nb ratio was done by measuring different Zr/Nb standards prepared from pure (>99.9%) Ames metal. The external reproducibility is better than ±1% (2r) for Ta, Zr, Lu and Hf concentrations, and ±4% for Nb concentrations, external reproducibility for Nb/Ta and Zr/Hf ratios is better than ±4% and ±0.6% (2r), respectively. Hafnium isotope ratio measurements were normalised to 179Hf/177Hf= 0.7325 and 176Hf/177Hf data are given relative to a JMC 475 value of 0.282160 (Blichert-Toft et al., 1997; for details see also Munker et al., 2001). Repeated measurements of the Ames metal Hf standard (isotopically indistinguishable from JMC 475) resulted in 176Hf/177Hf= 0.282160 ± 0.000018 (2r, n = 86) for concentrations between 24 and 80 ppb. Results of repeated measurements of reference materials BHVO-1 and BHVO-2, run during the course of our measurements are given in Pfander et al. (2007). Total procedural blanks were typically <5 ng for Zr, <100 pg for Ta, <50 pg for Hf, <20 pg for Nb, and <10 pg for Lu.

4. RESULTS

Niobium, Ta, Zr, Hf and Lu concentrations as well as Zr/Hf and Nb/Ta ratios are given in Table 1. Zr/Hf for all continental basalts range from 31.4 to 45.7 (Fig. 2a) and are, with the exception of two tholeiites from the Vogelsberg, higher than the chondritic value (34.3 ± 0.3; Munker et al., 2003). Nb/Ta in all samples range from 15.0 to 19.1 (Fig. 2b) and are consistently subchondritic

(<19.9 ± 0.6; Munker et al., 2003), but higher than in the bulk-silicate Earth (BSE; Nb/Ta = 14.0 ± 0.3; Munker et al., 2003) and along with some modern island arc lavas (Konig and Schuth, 2011) among the highest values measured so far in young terrestrial samples by high-precision (isotope dilution) methods. Compared to OIB and MORB, the continental basalts share an identical range in Zr/Hf, but Nb/Ta ratios tend to be higher (Fig. 2). Such elevated Nb/Ta, although at a larger scatter, have been also observed in volcanic rocks from other continental- and rift-related settings, e.g. in lavas from the East African Rift system (Furman and Graham, 1999) or in Mesozoic basalts from Northern China (Liu et al., 2008). In terms of Nb/Ta vs. Zr/Hf the samples define a slightly positive correlation that constrains the upper part of the terrestrial silicate frac-tionation trend (Fig. 2; see also Munker et al., 2003). Systematic differences in Zr/Hf and Nb/Ta between the different suites of samples are not observed and each suite covers nearly the whole range measured for both ratios during this study. Exceptions are samples from the Vogelsberg, where the variation in Nb/Ta is much larger for the basa-nites and nephelinites than for the tholeiites, the opposite of what is observed for Zr/Hf. This indicates that the processes that fractionated Zr/Hf during tholeiite genesis did not markedly affect Nb/Ta ratios.

In the Rhon samples, elevated Nb/Ta tend to correspond to elevated Ni and MgO contents (Fig. 3), i.e. the most primitive rocks display the highest Nb/Ta ratios. Basanites and nephelinites from the Vogelsberg span the whole range in Nb/Ta observed, but at a very restricted range of relatively high MgO and Ni contents. This suggests that variations in Nb/Ta are likely source controlled in these samples, whereas for the Rhon suite some Nb/Ta variation may be related to fractionation processes that also affect Ni and MgO contents.

Zr/Nb in continental basalts (2.22-4.22) overlap with the lower range of some HIMU-OIBs, but are generally lower than in other OIBs (Fig. 4), and are far below the chondritic value (13.5 ±0.6; Munker et al., 2003). Along with elevated Nb/Ta at overall higher Nb and Ta concentrations if compared to OIBs (average Nb and Ta concentrations in continental basalts and OIBs are 78.0 ppm vs. 60.2 ppm and 4.55 ppm vs. 3.75 ppm, respectively; OIB data from Pfander et al., 2007), this indicates a slight Nb-excess in continental basalts relative to OIBs (assuming similar degrees of partial melting that will strongly affect Zr/Nb ratios; see calculated melting lines in Fig. 4). Basanites and nephelinites from the Rhon area define a negative correlation in terms of Nb/Ta vs. Zr/Nb (Fig. 4) that most likely reflects different degrees of partial melting. Samples from the Eifel and Vogelsberg do not clearly correlate in this diagram, but plot close to the Rhon sample array with the exception of one Eifel and two Vogelsberg samples having distinctly lower Nb/Ta at a given Zr/Nb (Fig. 4). The Vogelsberg tholeiites have higher Zr/Nb (2.84-6.32) than basanites and nephelinites from this region, which are not correlated with Nb/Ta.

Lu/Hf and initial eHf values in the continental basalts (0.035-0.070 and +5.0 to +11.1 in basanites and nephelinites, 0.059-0.084 and +2.6 to +7.2 in Vogelsberg tholei-

Table 1

HFSE and Lu concentrations (ppm) as determined by isotope dilution (except Nb) and Lu-Hf isotope composition of intraplate basalts from Central Germany.

Sample Type Age MgO (wt.%) Lu Zr Nb Hf Ta Zr/Hf Nb/Ta 176Lu/177Hf 176Hf/177Hf±2a (176Hf/177l

21 Neph 20 10.08 0.354 378 112 8.78 6.60 43.0 17.0 0.0057 0.282944 ± 0.000007 0.282941

1a Bas 20 11.21 0.252 253 78.7 6.06 4.54 41.7 17.3 0.0059 0.283001 ± 0.000006 0.282998

1a (Dupl.) 0.251 6.06 4.53

1c Bas 20 10.34 0.254 257 70.3 6.29 4.30 40.8 16.4 0.0057 0.282969 ± 0.000009 0.282967

2 AOB 20 9.65 0.248 182 62.3 4.36 3.39 41.8 18.4 0.0081 0.282937 ±0.000013 0.282934

5 Bas 20 13.87 0.265 186 82.6 4.29 4.40 43.2 18.8 0.0087 0.283000 ± 0.000010 0.282997

5 (Dupl.) 0.263 4.28 4.37

6 AOB 20 9.69 0.284 221 68.0 5.05 3.94 43.7 17.2 0.0080 0.282937 ± 0.000009 0.282934

12 Neph 20 8.56 0.240 224 53.1 5.94 3.31 37.8 16.1 0.0057 0.282960 ± 0.000009 0.282958

10 Bas 20 8.86 0.281 280 71.1 6.83 4.31 41.0 16.5 0.0058 0.282967 ± 0.000007 0.282965

7 Hbl-Bas. 20 8.73 0.302 298 75.8 7.34 4.63 40.6 16.4 0.0058 0.283006 ± 0.000009 0.283004

18 Hbl-Bas 20 10.14 0.243 228 60.2 5.90 3.75 38.6 16.1 0.0058 0.282970 ± 0.000009 0.282967

19 Bas 20 11.67 0.252 163 72.2 3.96 3.84 41.2 18.8 0.0090 0.282962 ± 0.000007 0.282958

22 Bas 20 12.33 0.223 167 73.7 4.10 4.13 40.8 17.9 0.0077 0.282971 ± 0.000007 0.282968

Vögelsberg

VB-10 Bas 17 13.65 0.211 197 79.0 4.31 4.35 45.7 18.1 0.0069 0.282929 ± 0.000017 0.282927

VB-11 Bas 17 11.18 0.228 229 74.2 5.42 4.44 42.3 16.7 0.0060 0.283016 ±0.000013 0.283014

VB-20 Neph 17 14.56 0.246 196 88.3 4.60 4.63 42.7 19.1 0.0076 0.282974 ± 0.000009 0.282971

VB-23 Neph 17 13.20 0.268 230 103 5.50 5.74 41.8 18.0 0.0069 0.282976 ± 0.000010 0.282973

VB-28 Neph 17 14.13 0.221 231 81.7 6.00 5.37 38.5 15.2 0.0052 0.282905 ±0.000011 0.282903

VB-31 Neph 17 13.33 0.220 241 85.7 6.28 5.69 38.3 15.0 0.0050 0.282924 ±0.000011 0.282923

VB-13 Ol-Th 17 8.10 0.265 140 40.3 3.52 2.29 39.8 17.6 0.0107 0.282970 ± 0.000017 0.282966

VB-14 Ol-Th 17 8.40 0.266 139 38.6 3.49 2.29 39.7 16.8 0.0108 0.282969 ±0.000016 0.282966

VB-18 Ol-Th 17 10.09 0.249 132 46.5 4.20 2.71 31.4 17.1 0.0084 0.282857 ±0.000013 0.282854

VB-25A Ol-Th 17 7.93 0.243 135 29.5 3.63 1.78 37.1 16.6 0.0095 0.282842 ±0.000011 0.282839

VB-24A Qz-Th 17 7.96 0.242 130 28.5 2.88 1.72 45.1 16.6 0.0119 0.282909 ±0.000016 0.282906

VB-25B Qz-Th 17 7.82 0.254 134 28.7 3.62 1.76 37.1 16.3 0.0100 0.282847 ±0.000013 0.282844

VB-32 Qz-Th 17 7.78 0.227 101 24.6 2.96 1.37 34.2 17.9 0.0109 0.282902 ±0.000013 0.282899

VB-8 Qz-Th 17 7.42 0.268 164 24.9 4.12 1.44 39.8 17.3 0.0092 0.282837 ± 0.000014 0.282834

HEJ 5 Neph 25 13.22 0.348 229 100 5.63 6.48 40.7 15.5 0.0088 0.283060 ±0.000013 0.283056

HEJ 13 Bas 25 14.39 0.248 165 55.6 4.16 3.22 39.7 17.3 0.0085 0.283022 ± 0.000010 0.283018

HEJ 26 Bas 25 13.94 0.345 231 90.9 5.17 5.12 44.7 17.8 0.0095 0.283042 ± 0.000012 0.283038

HEJ 31 Neph 25 16.30 0.324 225 87.6 5.01 5.13 45.0 17.1 0.0092 0.282991 ±0.000010 0.282987

HEJ 59 Neph 25 11.65 0.335 207 85.9 4.78 4.59 43.4 18.7 0.0099 0.283063 ±0.000013 0.283058

HEJ 68 Bas 25 14.00 0.262 171 60.0 4.22 3.37 40.6 17.8 0.0088 0.283075 ± 0.000010 0.283070

Ages and MgO contents are taken from Jung and Masberg (1998), Jung and Hoernes (2000), and Jung et al. (2006) and from references therein. Lu, Hf,

6Lu/177Hf and 176

Vogelsberg samples are from Jung et al. (2005) and Jung et al. (2011). Errors represent in-run precision. Lu-Hf decay constant used: 1.865 x 10 11 a 1 (Scherer et al..

'Hf/1 2001).

7Hf of Rhön and

terrestrial silicate .

30 32 34 36 38 40 42 44 46 48

20-119 18 17 16 15 14

• Rhön (B&N)

♦ Elfel (B&N}

a Vogelsberg (B&N) A Vogelsberg (T)

• ч

К'*' *

15.9 ±0.6 (1<j)-

average OIB

30 32 34 36 38 40 42 44 46 48

Fig. 2. Nb/Ta vs. Zr/Hf in continental basalts from Central Germany. (a) Zr/Hf ratios are similar to values obtained in OIBs, whereas Nb/Ta ratios tend to be higher. Also shown are HFSE ratios measured in MORB samples (BUchl et al., 2002). CC = continental crust (Barth et al., 2000), BSE = bulk silicate Earth (Münker et al., 2003). (b) No differences are observed in Nb/Ta and Zr/Hf between different locations, but the variation in Nb/Ta in Vogelsberg tholeiites is markedly lower than in Vogelsberg basanites and nephelinites, which contrasts with the variations in Zr/Hf (B&N = Basanites and Nephelinites, T = Tholeiites; OIB data from Pfander et al., 2007).

ites) fall within the compositional range observed for most OIBs (Fig. 5). Exceptions are the Vogelsberg tholeiites that tend to have higher Lu/Hf but exhibit distinctly lower initial 8Hf. The highest initial eHf values are observed in Eifel samples (+11.1 to +8.1), intermediate values in Rhein samples (+8.7 to +6.2) and intermediate to low values in Vogelsberg samples (+7.5 to +5.0; Fig. 5b). In terms of 8Hf vs. eNd, all investigated samples plot slightly below the mantle array, and close to the composition of HIMU OIBs (Fig. 6; except Vogelsberg tholeiites). For the Vogelsberg basanites and nephelinites, eHf and eNd tend to be negatively correlated. With respect to Lu/Hf, samples from the Vogelsberg and the Eifel form distinct groups with higher ratios in the Eifel (>0.055) compared to the Vogelsberg samples (<0.055). Samples from the Rhon plot in-between, but can be separated into two groups, one with lower Lu/ Hf and Nb/Ta, and one with higher Lu/Hf and Nb/Ta (Fig. 5a). Elevated Nb/Ta therefore seem to be associated with elevated Lu/Hf in most basanites and nephelinites, but not in Vogelsberg tholeiites. This is particularly evident

from the highly correlated linear relationship between Nb/ Ta and Lu/Hf in the Vogelsberg basanites and nephelinites (Fig. 5a). No distinct correlations between Nb/Ta and eHf are observed, but in general, elevated eHf values seem to correspond to elevated Nb/Ta in basanites and nephelinites from the Vogelsberg and Eifel regions.

Relative to the investigated basanites and nephelinites, the Vogelsberg tholeiites show fairly different HFSE and trace element systematics. Unlike Nb/Ta and Nb/La that are within the range observed for basanites and nepheli-nites, Nb concentrations, Nb/Lu and La/Yb along with initial eHf values are distinctly lower, whereas Lu/Hf ratios tend to be higher (Fig. 5). This is consistent with previous suggestions, that the Vogelsberg tholeiites were derived from more depleted mantle sources at higher degrees of partial melting than the alkaline Vogelsberg rocks (Bogaard and Worner, 2003). The different Lu/Hf ratios may also reflect different melting depths (i.e. spinel vs. garnet stability fields).

5. DISCUSSION 5.1. Fractional crystallisation and crustal assimilation

Fractional crystallisation and concurrent crustal assimilation are post melting processes that may obliterate primary magma compositions, and thus may hamper a straightforward interpretation of trace element data with respect to the nature and composition of the melt source region (e.g. DePaolo, 1981; Spera and Bohrson, 2001). Evaluation of these processes is thus required before conclusions on the origin and source of basaltic magmas can be drawn from trace element data.

5.1.1. Effect of fractional crystallisation on Nb/Ta and Zr/Hf

Basanites and nephelinites from the Rhon, Vogelsberg and Eifel volcanic fields represent near primary melts that underwent only limited crystal fractionation (Jung and Masberg, 1998; Jung and Hoernes, 2000; Bogaard and Worner, 2003; Jung et al., 2006). In contrast, the alkali-olivine basalts from the Rhon and the olivine- and quartz-tholeiites from the Vogelsberg fractionated clinopy-roxene + olivine + amphibole (alkali basalts), olivine + orthopyroxene (tholeiites), and partly plagioclase (qtz-tholeiites) (Jung and Masberg, 1998; Jung and Hoernes, 2000; Bogaard and Worner, 2003). The alkaline basalts from the Vogelsberg also fractionated Fe-Ti oxides as is evident from decreasing TiO2 with decreasing MgO. In Fig. 3, Nb/Ta and Zr/Hf are plotted against MgO and Ni concentrations. The most primitive samples (MgO >12 wt.%, Ni >350 ppm) tend to have the highest Nb/Ta and Zr/Hf ratios, although two samples from the Vogelsberg plot at lower Nb/Ta for given Ni concentrations and samples from the Eifel scatter. In general, however, interpreting the positive correlations between Nb/Ta and Ni and MgO observed in the Rhon samples (Fig. 3b and d) as fractionation trends would require a fractionating assemblage with a bulk DNb/Ta >1. Olivine and orthopyroxene do not markedly shift Nb/Ta during crystallisation due to extremely low partition coefficients for both elements (see

504642-

C " 38-

34 30 (b)

20 18 16 1412

A T075^ ♦

A*.....

\ ▲ A

o Rhön (B&N) ♦ Eifel (B&N) A Vogelsberg (B&N) A Vogelsberg (T)

10 12 14 MgO [wt%]

20! 18" ■ 1614-

♦1 a °0

o Rhön (B&N) ♦ Eifel (B&N) A Vogelsberg (B&N) A Vogelsberg (t)

0 100 200 300 400 500 Ni [ppm]

Fig. 3. Zr/Hf and Nb/Ta vs. MgO and Ni concentrations in continental basalts from Central Germany. For Rhon samples (dashed fields), higher MgO correlate with slightly higher Nb/Ta (b). The most primitive lavas tend to have the highest Zr/Hf (c) and Nb/Ta ratios (d), and particularly in the Rhon samples distinct positive correlations are observed. In Vogelsberg samples, the basanites and nephelinites display a higher variation in Nb/Ta at higher Ni contents than the tholeiites (d; Ni concentrations are from Jung (1995), Jung and Masberg (1998), and Jung et al. (2006)).

24 22 20-

Si 18H

16 1412

o Rhön (B&N) ♦ Eifel (B&N) A Vogelsberg (B&N) A Vogelsberg (t)

^ other ■ oib OIB

HIMU average OIB cont.

4 5 6 7 Zr/Nb

Fig. 4. Nb/Ta vs. Zr/Nb in continental basalts. Zr/Nb of continental basalts plot at the lower end of the field defined by OIBs, with a broad overlap with HIMU OIB values, whereas Nb/Ta tend to be higher than in most OIB (OIB data from Pfander et al., 2007). Also shown are calculated melt compositions of typical astheno-spheric garnet (gt) and spinel (sp) peridotite sources (black lines; see text and Appendix A). Dashed line: Melt composition resulting from partial melting of a metasomatised refractory spinel peridotite (numbers denote the degree of melting) that has been modified by the addition of 1% metasomatic melt (see Appendix A for details).

Electronic Annex). Clinopyroxene has DNb/Ta <1 as evident from most published partitioning data (as is the case for Zr/ Hf, see Electronic Annex) and thus will produce melts with higher Nb concentrations and higher Nb/Ta as well as

Zr/Hf during crystallisation. The same effect on Nb/Ta, but at decreasing Nb concentrations, will be caused by the fractionation of Ti-bearing minerals (rutile, sphene, ilmenite, Ti-magnetite) that also have DNb/Ta <1 in mafic systems but absolute values for DNb and DTa are »1 (Green and Pearson, 1987; Foley et al., 2000; Schmidt et al., 2004; Klemme et al., 2005; Xiong et al., 2005; see Electronic Annex). Decreasing Nb/Ta with decreasing MgO and Ni concentrations (Fig. 3b and d) along with increasing TiO2 (that rule out fractionation of Ti-bearing phases; not shown) can therefore only be explained by crystallisation processes if low-Mg amphibole having DNb/Ta >1 (Tiepolo et al., 2000) dominates the fractionating assemblage along with olivine (and possibly clinopyroxene). Calculated DNb/Ta ratios of amphibole phenocrysts from the Rhon and Eifel areas are >1 (using their Mg# and Ti contents as given in Jung and Hoernes (2000) and Jung et al. (2006), and the algorithm of Tiepolo et al. (2000), which in principle supports this assumption. Scandium is compatible in amphibole and clinopyroxene in basanitic melts (Adam and Green, 2006). In the Rhon samples, Sc concentrations first increase with decreasing MgO and then strongly decrease from ~10 wt.% MgO downwards. In contrast, Sc contents in basanites and nephelinites from the Ei-fel and Vogelsberg consistently increase with decreasing MgO. Amphibole fractionation, likely following olivine, is therefore restricted to the most evolved basanites and nephelinites from the Rhon area, and may have slightly lowered Nb/Ta in these samples. The occurrence of

21 20 19

carbonatite

21 20 19 18 17 16 15 14 13

0.06 Lu/Hf

Vogelsberg Tholeiites

^ crustal ^ contamination

Vogelsberg

carbonatite assimilation

Rurutu

I.U.U.UV».. Ruiuiu 1. V

Tubuai(HIMU) young (b)

I • Rhön »Eifel Vogelsberg: A B&N ATh

6 8 SHf(t)

Fig. 5. Nb/Ta vs. Lu/Hf and £Hf(t) in continental basalts. (a) Elevated Nb/Ta at elevated Lu/Hf (except for the Vogelsberg tholeiites and not distinct for the Eifel samples) can be explained by carbonatite assimilation in the lithospheric mantle (symbols as above, details see text). Also shown is a calculated mixing line between the sample with the lowest Lu/Hf and Nb/Ta (VB31) and a model carbonatite (composition see Electronic Annex; numbers denote degree of carbonatite assimilation in percent). (b) In the Vogelsberg and most Eifel samples, Nb/Ta is also positively correlated with £Hf(t) values. Also shown are compositions of typical HIMU basalts from Rurutu and Tubuai (grey field, darker regions correspond to higher sample densities; data from Pfander et al., 2007). Samples termed "Rurutu young" originated from the remelting of metasomatised oceanic lithosphere (Chauvel et al., 1997) that had no effect on Nb/Ta, in marked contrast to metasomatism of the continental lithosphere (symbols as before).

amphibole phenocrysts in some Rhön basalts supports this suggestion (Jung and Hoernes, 2000). For the majority of investigated basanites and nephelinites, however, fractional crystallisation of mafic phases including amphibole and Ti-bearing minerals is of minor importance in explaining the observed Nb/Ta as well as Zr/Hf variations and their correlations with major and other trace elements. Instead, complexities in the source region of these magmas and post-melting processes other than simple fractional crystallisation might dominate the observed geochemical patterns.

5.1.2. Effect of crustal assimilation on Nb/Ta

Compared to the continental crust with extremely high Re/Os and highly radiogenic 187Os/188Os (crustal average ~1.7; Asmerom and Walker, 1998) the asthenospheric

• Rhön (B&N)

♦ Eifel (B&N) a Vogelsberg (B&N) A Vogelsberg (T)

Hawaii & , /» !;•<

EM2 (Samoa) ..s1**.

Fig. 6. eHf vs. eNd in continental basalts from Central Germany. Also shown are the compositional fields of MORB, EMI, EM2 and HIMU basalts (figure and data from Pfander et al., 2007 and references therein). Most of the basanite and nephelinite samples plot within or close to the field of HIMU OIBs and tend to lie slightly below the mantle array. BSE = Bulk silicate Earth (Blic-hert-Toft and Albarede, 1997).

and (non-metasomatised) lithospheric mantle both have relatively low time-integrated Re/Os and subchondritic (litho-spheric mantle) to about chondritic (asthenospheric mantle) 187Os/188Os ratios (60.127; Walker et al., 1989; Pearson et al., 1995a,b). The Os isotope composition of astheno-spheric melts, even if derived from mantle plumes that contain recycled crust having elevated 187Os/188Os, will thus be relatively low (up to ~0.16; e.g. Reisberg et al., 1993) if compared to the continental crust. Interaction of such low Os melts with the (oceanic) lithosphere having high Os concentrations but subchondritic 187Os/188Os may further lower 187Os/188Os in such melts and may produce an overall larger variation in 187Os/188Os (e.g. Widom et al., 1999; Class et al., 2009). The same effect is expected if asthenospheric melts interact with the aged continental lithospheric mantle that is likely also subchondritic in 187Os/188Os (Walker et al., 1989; Ellam et al., 1992; Pearson et al., 1995b; Shirey and Walker, 1998). For the SCLM beneath Central Germany, a range of 187Os/188Os between 0.124 and 0.142 in xenoliths has been documented (Schmidt and Snow, 2002), most likely reflecting low initial 187Os/188Os overprinted by radiogenic ingrowth caused by late-stage metasomatic Re addition. In contrast to astheno-sphere melt-lithosphere interaction and despite the extremely low Os concentration in the continental crust, assimilation of continental crustal material with its extremely radiogenic Os may significantly increase 187Os/188Os (>~0.16) in asthenospheric and lithospheric melts. Radiogenic Os in continental basalts from Germany and from elsewhere has therefore been interpreted as the result of crustal contamination (Hegau volcanic field: Blusztajn and Hegner, 2002; Rhon: Jung et al., 2005; Vogelsberg basanites and nephelinites and particularly Vogelsberg tholeiites: Jung et al., 2011; see also Fig. 7; other locations worldwide: Ellam et al., 1992; Hart et al., 1997; Chesley and Ruiz, 1998; McBride et al., 2001).

Basanites and nephelinites have the largest variation in Nb/Ta at restricted and relatively low 187Os/188Os ranging from 0.126 to 0.194 with the lowest values similar to the

21 20 19

Z 17 16 15 14

(187Os/188Os)i

Fig. 7. Nb/Ta vs. initial 187Os/188Os in continental basalts. Samples with lower Nb/Ta tend to have higher initial 187Os/188Os ratios (except tholeiites from the Vogelsberg). Assimilation of continental crust strongly increases 187Os/188Os without a resolvable change in Nb/Ta, whereas carbonatite assimilation enlarges Nb/Ta without a significant change in 187Os/188Os. Light grey field: range of initial 187Os/188Os in primary melilitites from the Urach and Hegau volcanic fields in Southern Germany (data from Blusztajn and Hegner, 2002); dark grey field: range of 187Os/188Os in xenoliths from the Eifel and Vogelsberg regions (data from Schmidt and Snow, 2002). For both groups of samples, Nb/Ta are not available. Black star: Inferred composition of primary magmas (before assimilation). 187Os/188Os in continental crust ~1.7 (Asm-erom and Walker, 1998), Nb/Ta ~12-13 (Barth et al., 2000). Os isotope data from Jung et al. (2005) and Jung et al. (2011).

chondritic ratio (~0.127, Asmerom and Walker, 1998; Fig. 7). Furthermore, the highest Nb/Ta tend to be related to the least radiogenic 187Os/188Os. This indicates that the primary magmas, either from asthenospheric or lithospheric sources, already shared highly variable Nb/Ta before they reached the continental crust which has radiogenic 187Os/188Os (and low Nb/Ta: ~12-13; Barth et al., 2000; Rudnick and Gao, 2003; Willbold et al., 2009). In other words, elevated and variable Nb/Ta in continental basalts at low 187Os/188Os are a source feature and not related to crustal assimilation.

Besides the low 187Os/188Os samples, some of the alkaline rocks have radiogenic 187Os/188Os at relatively low Nb/Ta (Fig. 7). As suggested previously, some of the volcanic rocks of the Rhon and Vogelsberg areas have assimilated crustal material (Jung and Masberg, 1998; Jung and Hoernes, 2000; Bogaard and Worner, 2003; Jung et al., 2005). Particularly the Vogelsberg tholeiites underwent significant lower crustal contamination (4-8%), expressed in moderately radiogenic Sr at unradiogenic Nd isotope compositions (Jung and Masberg, 1998; Bogaard and Worner, 2003), unradiogenic Hf isotope compositions relative to the alkaline rocks (eHf down to +2.6; this study; Fig. 5b), elevated Zr/Nb (Jung and Masberg, 1998, and Fig. 4) and extremely radiogenic 187Os/188Os (up to ~0.75; Jung et al., 2011). In contrast, their Nb/Ta are fairly constant (16.3-17.9) and the non-existing decrease in Nb/Ta with increasing 187Os/188Os indicates that the assimilation of crustal material has a negligible effect on Nb/Ta in the tholeiites. This is likely due to the relatively low Nb and Ta

concentrations in the continental crust (~8 ppm and ~0.7 ppm in bulk continental crust, respectively; Barth et al., 2000; Rudnick and Gao, 2003). As the Nb and Ta concentrations in the basanites and nephelinites are markedly higher than in the tholeiites, and as these rocks assimilated, if at all, less crustal material than the tholeiites, the low Nb/Ta in some alkaline basalts can not be caused by crustal assimilation even for samples having high 187Os/188Os. It is more likely that the Nb/Ta-187Os/188Os relationship in the alkaline basalts indicates the presence of different source components with differing 187Os/188Os and Nb/Ta, and a multi-stage process during magma genesis. Asthenospheric melts with OIB-like Nb/Ta (~ 15-16) could attain elevated Nb/Ta during assimilation processes within the lithospheric mantle, where metasomatised domains with unradiogenic Os but high Nb/Ta may exist. Such domains could be carbonatite-rich veins, as has been suggested previously on the basis of Pb-Hf-Os isotope compositions of continental basalts from this region (Jung et al., 2011). Pyroxene- and amphibole/mica-rich veins and layers from former melt injections are also a constituent of the lithospheric mantle and in contrast they may have radiogenic Os isotope compositions (Carlson and Nowell, 2001). Preferred melting of such veins due to a lower solidus than the ambient lithosphere, and mixing of these melts with the ascending high-Nb/Ta melts may then produce elevated 187Os/188Os during a second-stage process, possibly at shallower levels in the lithosphere. Subsequent crustal assimilation may further increase 187Os/188Os, but will have a negligible effect on Nb/Ta.

5.2. Source constraints from melting calculations

Previous studies have proposed that at least two distinct mantle sources, an asthenospheric and a lithospheric mantle component are required to explain the petrological and geochemical characteristics of the alkaline volcanism in Central Europe (e.g., Wilson and Downes, 1992; Jung et al., 2006; Lustrino and Wilson, 2007). A similar conclusion was made by Jung et al. (2006) for samples from the Eifel region. Based on trace element, Nd-Sr and O isotope systematics, Jung and Hoernes (2000) suggested that the Rhon volcanic rocks were also mostly derived from two distinct sources, a plume related asthenospheric source with FOZO-like trace element and isotope signatures (FOZO, see Stracke et al., 2005) and a lithospheric mantle source. Using the high-precision trace element data of this study allows to test these hypotheses.

The effects of melting asthenospheric garnet- and spinel peridotite as well as eclogite bearing mantle sources on the HFSE budget of silicate melts have previously been discussed by Pfander et al. (2007). It was shown that OIB sample suites from individual ocean islands form positive linear melting trends in plots of Nb/Ta vs. Nb concentration. The different slopes of these correlations reflect slightly different bulk-DNb/Ta of the mantle sources and the absence of crystal fractionation and assimilation processes that may affect the HFSE budget of OIBs (Pfander et al., 2007). In contrast, no correlations between Nb/Ta and Nb concentrations are observed in the continental basalts, suggesting

more heterogeneous source compositions or more complicated melt generation processes involving different sources. Variations in Nb/Lu, La/Yb and Nb concentrations in primitive mantle melts are strongly influenced by the degree of partial melting (Fig. 8). Therefore, at a given source com-

gt\ mixing lines 0.5 0__\

\ ...... /.^A _AJ—---—

) ......'...........£r:

2/ 3,/i 5,/ti-f -1-10 30 1 H-group melts (lithospheric meltst

0 20 40 60 80 100 120 140 Nb [ppm]

lithospheric melts

60 50403020-

- L-group 41.4 melts

H-group

melts v I \ ! f CU

(a a .

5.0 JJ^'-P \ A A *

O Rhon (B&N) ♦ Eifel (B&N) A Vogelsberg (B&N) A Vogelsberg (T)

refractory

O 0.5 ♦A 1 _ 0 o . »gt

sp asthenos-

0.5 pheric melts more fertile

(cpx rich)

1.5 Nb/La

Fig. 8. Nb/Lu vs. Nb concentration and La/Yb vs. Nb/La in continental basalts along with calculated melt compositions derived from asthenospheric spinel- and garnet-peridotite sources ('sp' and 'gt', respectively) with a primitive mantle starting composition (Hofmann, 1988). Also shown are melting lines calculated from the modal and trace-element composition of refractory spinel-perido-tite xenoliths from the Hessian Depression (data from Hartmann and Wedepohl, 1990: H-group = highly enriched, L-group = moderately enriched in highly-incompatible trace elements relative to PRIMA, see text and Electronic Annex for details). These xenoliths have been taken as being representative for the composition of the lithospheric mantle beneath Central Germany. (a) Continental basalts are mixtures between melts derived from a refractory lithospheric, but secondary enriched spinel peridotite and a more fertile asthenospheric garnet peridotite. Vogelsberg tholeiites have lower Nb concentrations and Nb/Lu and La/Yb due to higher degrees of melting and a different source composition. (b) Elevated Nb/La in all continental basalts can also be explained by addition of asthenospheric melts derived from spinel or garnet peridotite sources with high Nb/La to partial melts from lithospheric sources. Grey field: Range of melt compositions that result from partial melting of the lithospheric mantle (calculated from spinel peridotite xenolith compositions as given by Hartmann and Wedepohl (1990); numbers denote degrees of melting in percent). La and Yb concentrations from Jung and Masberg (1998), Jung and Hoernes (2000) and Jung et al. (2006) (see Electronic Annex).

position, these ratios allow to infer the degree of melting involved during magma genesis. In comparison, Nb and La are similarly incompatible elements and thus Nb/La are less influenced by partial melting than Nb/Lu at melting degrees that exceed ~1%, and more closely mirror the chemical and mineralogical composition of their source.

High GdN/YbN (2.3-4.8, except one with 7.0) in the continental basalts indicate that garnet was a residual phase during melt formation. As is evident from Fig. 8a, calculated melt compositions assuming a primitive mantle-like starting composition (Hofmann, 1988; see Appendix A for details on melting calculations) require at least 3-4% modal garnet in an asthenospheric source and low melting-degrees (<2%) to produce Nb/Lu in the observed range, but in this case strongly underestimate the Nb concentrations as well as La/Yb in the samples (Fig. 8, lines termed 'gt'). Assuming an asthenospheric spinel peridotite source with the same starting composition (lines 'sp' in Fig. 8), the Nb concentrations and Nb/Lu are markedly lower in the calculated melts than in the samples (Fig. 8a), as are the La/Yb ratios (Fig. 8b; except for the Vogelsberg tholei-ites, for which the compositions are well matched in both plots). Therefore, melting primitive-mantle like astheno-spheric spinel or garnet peridotite sources cannot explain the observed trace element signatures in basanites and nep-helinites (Fig. 8). In contrast, melts calculated from the averaged modal composition and trace element abundances of xenoliths from the Hessian Depression (Hartmann and Wedepohl, 1990; see Appendix A and Electronic Annex) yield markedly higher Nb concentrations and La/Yb at similar Nb/Lu (lines termed 'L-group' and 'H-group' in Fig. 8, grey field in Fig. 8b). These xenoliths are spinel per-idotites found in alkali basalts and nephelinites from the Hessian Depression, northeast of the Vogelsberg, and are thought to represent samples from the subcontinental litho-spheric mantle (SCLM). They are predominantly clinopy-roxene-poor, refractory harzburgites and minor lherzolites that have undergone previous melting, but were re-enriched by metasomatic alteration. Hartmann and Wedepohl (1990) distinguished L-group (low), I-group (intermediate) and H-group (high) xenoliths with different degrees of secondary enrichment in highly incompatible trace elements (see enrichment factors in the Electronic Annex). A striking feature of these xenoliths is, that partial melting of them leads to a stronger fractionation of La/Yb and Nb/Lu with overall higher values than melting of asthenospheric spinel- and even garnet peridotite sources (Fig. 8). The reason for this is the refractory mineralogy, but in particular the secondary enrichment process that produced very high Nb/Lu and La/Yb in these xenoliths. A combination of these high source ratios with the refractory mineralogy and low melting degrees produces the observed variations and suggests that partial melting of metasomatically re-enriched spinel-bearing lithospheric mantle, sampled by the xenoliths, played a pivotal role during the petrogenesis of these lavas.

In contrast to Nb/Lu and La/Yb, Nb/La in melts calculated from the spinel-bearing xenoliths do not match the observed compositions and are significantly lower (Fig. 8b). As DNb/La in clinopyroxene is less than 0.5 (see Electronic Annex), a higher modal abundance of

clinopyroxene in the xenoliths will increase Nb/La in the calculated melts, and may account for some variation in the data. It appears more likely, however, that the elevated Nb/La in the continental basalts reflect mixing between two melts from two different sources. A low-degree (<3%) asthenospheric melt derived from a clinopyroxene-rich garnet peridotite ('gt' in Fig. 8b) and a lithospheric melt from a refertilized clinopyroxene-poor spinel-bearing refractory harzburgite as represented by the xenoliths ('H-group' in Fig. 8b). This interpretation is in very good agreement with the Nb/Lu vs. Nb data where all samples plot between the melting lines of asthenospheric and lithospheric sources (Fig. 8a, dashed lines). In particular the alkaline lavas from the Eifel form a well defined linear trend in both, Nb/Lu vs. Nb as well as La/Yb vs. Nb/La space (Fig. 8), which can be explained as a mixing relationship. This implies that the lithospheric mantle endmember involved during melting is similar to the H-group xenolith composition of Hartmann and Wedepohl (1990). Notably, samples having the highest contribution from the lithospheric source tend to have the highest Nb concentrations and Nb/Ta ratios, underlining that elevated Nb/Ta are a feature of the subcontinental lithospheric mantle. About 3-4% partial melting of the lithospheric source and about 2-5% partial melting of the asthenospheric source and subsequent magma mixing is required to produce the compositional variation observed in lavas from the Eifel region. Lower degrees of melting and a greater contribution of the asthenospheric melt are required to explain the Vogelsberg alkaline lavas and some of the Rhon lavas. The composition of the Vogelsberg tho-leiites requires the highest contribution of asthenospheric melts (>90%). This is compatible with the inferred depleted mantle source for the tholeiites as suggested by Bogaard and Worner (2003), although estimated melting degrees are lower than previously suggested (5-7%, Jung and Masberg, 1998; 5-15%, Bogaard and Worner, 2003). Bogaard and Worner (2003) postulated the involvement of three distinct sources during the evolution of the Vogelsberg lavas (an enriched plume-like source, a depleted source, and a metasomatised lithospheric mantle source). The trace element systematics presented here confirm these findings and indicate an asthenospheric primitive mantle source and a re-enriched lithospheric mantle source for the alkaline lavas and a predominantly asthenospheric depleted to primitive mantle source for the tholeiites.

5.3. Carbonatite metasomatism

5.3.1. Evidence for a carbonatite signature in continental basalts

Metasomatism of the lithospheric mantle can account for a replenishment in highly incompatible trace elements (McKenzie, 1989; Rudnick et al., 1993; Wilson et al., 1995; Bodinier et al., 1996; Bedini et al., 1997; Beccaluva et al., 2001; Downes, 2001; Witt-Eickschen, 2007; Tappe et al., 2008; Pilet et al., 2008). The overall chemical and iso-topic enrichment pattern depends on the nature of the metasomatising liquid, which may be a hydrous fluid, a low-degree silica-undersaturated or a carbonatitic melt. Pervasive metasomatism within the SCLM is widespread

beneath Central Europe and produced phlogopite and amphibole bearing spinel peridotites, amphibolite veins, phlogopite bearing wehrlites and clinopyroxenites (Stosch and Seck, 1980; Stosch and Lugmair, 1986; Hartmann and Wedepohl, 1990; Zinngrebe and Foley, 1995; Downes, 2001; Witt-Eickschen, 2007; Jung et al., 2011). Carbonatite metasomatism has also been described to take place in the continental and oceanic lithosphere (Yaxley et al., 1991; Dupuy et al., 1992; Hauri et al., 1993; Ionov et al., 1993; Rudnick et al., 1993; Schiano et al., 1994; Hoernle et al., 2002; Jorgensen and Holm, 2002; Rabinowicz et al., 2002; Dasgupta et al., 2004; Moine et al., 2004; van Achterbergh et al., 2004; Tappe et al., 2008; Rosenthal et al., 2009; Dou-celance et al., 2010), and experimental results indicate that carbonate phases (calcite, dolomite, magnesite) coexist with amphibole, phlogopite, spinel and garnet under subsolidus conditions in peridotite assemblages in the pressure range from ~10 kbar to more than 40 kbar (Brey et al., 1983; Olafsson and Eggler, 1983). Evidence for carbonatite metasomatism in the European lithospheric mantle, particularly beneath Central Germany, comes from xenoliths, the occurrence of carbonatite rocks in the Kaiserstuhl region, Southwestern Germany (Keller, 1981), and from the occurrence of carbonatite fragments in tuffs from the Laacher See region of the West Eifel (Lloyd and Bailey, 1969). Compared to other SCLM xenoliths in Central and Western Europe, mantle xenoliths from the Northern Hessian Depression have some of the lowest and most consistently subchondritic Ti/Eu at low Ti concentrations (Hartmann and Wedepohl, 1990; Downes, 2001). This is regarded as evidence for carbonatite metasomatism as partial melting will have an only limited effect on Ti/Eu in mantle xeno-liths. Carbonatite metasomatism will cause a strong shift in this ratio due to extremely low Ti/Eu in carbonatites (Yaxley et al., 1998; Downes, 2001).

Elevated Nb/Ta if compared to OIBs (and MORB) are a specific feature observed in continental basalts (but also some HIMU OIBs) and correlate with very low Zr/Nb (Fig. 4) and high Nb concentrations (Fig. 8a). These characteristics indicate a slight Nb excess with respect to other HFSE in basaltic rocks from continental settings. Melting of a lithospheric mantle source metasomatised by low-degree silicate melts can principally explain these features (Fig. 4, dashed line), but will produce melting lines where decreasing Nb/Ta are correlated with increasing Lu/Hf. This is in marked contrast to the observed trends, where a positive co-variation between Nb/Ta and Lu/Hf is evident for primitive samples from all locations (except Vogelsberg tholeiites; Fig. 5a). Particularly in the basanites and nephe-linites from the Vogelsberg, Nb/Ta and Lu/Hf are well correlated and this suggests that the variation in these ratios is controlled by two-component mixing (or assimilation; see calculated mixing line in Fig. 5a). The composition of two nephelinites from the Vogelsberg with the lowest Nb/Ta and Lu/Hf (VB28 and VB31; 15.2 and 15.0, and 0.037 and 0.035, respectively) is very close to the composition of many mantle derived magmas, particularly OIBs that have Nb/Ta ~15-16 and Lu/Hf ~0.03-0.05 (Pfander et al., 2007). Assuming that these samples (VB28 and VB31) represent the least modified near-primary melts,

the assimilated component is required to have high Nb/Ta (>20) and high Lu/Hf (>0.1). Such a feature is observed in carbonatites (Hoernle et al., 2002; Bizimis et al., 2003; Cha-khmouradian, 2006; see also Electronic Annex), which have been previously suggested to play a role during the genesis of the alkaline lavas from this location (Jung et al., 2011). Owing to the extreme depletion in Ti, Zr and Hf relative to HREE but only moderate depletion in Nb and Ta relative to LREE, carbonatites combine high Lu/Hf with high Zr/Hf and high Nb/Ta but extremely low Zr/Nb (Rudnick et al., 1993; Hoernle et al., 2002; Bizimis et al., 2003; see also Electronic Annex). Carbonatite assimilation, or remelting of a carbonated mantle source, is therefore a viable process to explain high Nb/Ta and low Zr/Nb (Fig. 4) coupled to slightly elevated Zr/Hf and, most importantly, elevated Lu/Hf in alkaline lavas (Figs. 2 and 5a; note that in the following we solely consider assimilation, but remelt-ing of a carbonated peridotite source will yield similar results). During ascent, migration of such alkaline melts in the relatively cold lithospheric mantle will be predominantly along pre-existing structures and networks, where metasomatic phases such as carbonatites with superchon-dritic Nb/Ta are likely to be present from former melting or melt infiltration events. In particular low-volume carbonatite liquids with their restricted heat content (McKenzie, 1989) may be concentrated along such zones, from where they can be assimilated by later stage alkaline melts. As elevated Nb/Ta are also observed in lavas from the East African Rift system (Furman and Graham, 1999), where carbonatite occurrences are widespread (e.g. Rosenthal et al., 2009), such a process may be of general importance in continental- and rift-related settings. In contrast to Nb/ Ta and Lu/Hf, the effect of carbonatite assimilation on Zr/Hf is less pronounced owing to the fact that the Zr and Hf concentrations in carbonatites are in the order of those in continental basalts and OIBs, whereas the Nb and Ta concentrations tend to be much higher. This is consistent with the observation that the Zr/Hf ratios in continental basalts overlap those of OIBs, whereas Nb/Ta ratios are systematically higher (Fig. 2). Although there is an extreme variation of HFSE concentrations in carbona-tites from different settings (see compilation in Chakhmou-radian, 2006 and Electronic Annex), calculations using averaged concentration estimates confirm that Nb/Ta is more readily affected by carbonatite assimilation than Zr/ Hf. For example, taking the average carbonatite composition of Bizimis et al. (2003), addition of 5% carbonatite shifts the Zr/Hf in a sample (e.g. the Vogelsberg basanite VB31) from 38.3 to 38.7, but increases the Nb/Ta drastically from 15.1 to 18.8. These variations are in good agreement with the data shown in Fig. 2. Using the average carbonatite compositions of Hoernle et al. (2002) yields similar results, although the shift in Nb/Ta is less dramatic owing to the lower Nb concentrations and higher Zr/Nb in these carbonatites. Carbonatites also have very low Zr/Sm (Bizimis et al., 2003). Therefore, carbonatite assimilation during the evolution of the alkaline lavas described here is also evident from the subchondritic Zr/Sm ratios (<25.3; Sun and McDonough, 1989) that correlate to high Lu/Hf within nearly all samples from the Eifel and

45 403530 2520 15

EMI &EM2

__chondritic_ ratio

О Rhön (B&N) ♦ Eifel (B&N) A Vogelsberg (B&N) A Vogelsberg (T)

250 Zr [ppm]

Fig. 9. Zr/Sm vs. Zr concentration in continental basalts compared to OIBs. Predominantly subchondritic Zr/Sm in continental basalts contrast those observed in OIBs and most likely result from carbonatite assimilation in the subcontinental lithospheric mantle, or from the remelting of a carbonated lithospheric mantle source. Chondritic ratio from Sun and McDonough (1989), OIB data from Willbold and Stracke (2006).

Vogelsberg regions, and which are in marked contrast to predominantly superchondritic Zr/Sm in OIBs (Fig. 9).

5.3.2. Age of carbonatite metasomatism

Similar to Lu/Hf, initial eHf values in the Vogelsberg and Eifel alkaline rocks are positively correlated with Nb/Ta (Fig. 5b). As carbonatite rich domains in the SCLM with high Lu/Hf will produce elevated 176Hf/177Hf ratios over time (e.g. Tappe et al., 2007; Prelevic et al., 2010), the variations in 176Hf/177Hf suggest that carbonatite metasomatism in the SCLM occurred long before assimilation and magma eruption. Assuming that the correlations between Nb/Ta vs. Lu/Hf (Fig. 5a) and eHf (Fig. 5b) in Vogelsberg basanites and nephelinites predominantly reflect carbonatite assimilation, the variations in 176Hf/177Hf permit the calculation of the time interval between carbonatite metasomatism and subsequent assimilation or remelting of these carbonatites by rising alkaline melts (i.e. the average storage or life time for carbonatite veins or patches in the SCLM). This was done by calculating the time required to generate a sufficiently high 176Hf/177Hf ratio in carbonatite patches or veins in the litho-spheric mantle (at a given Lu/Hf and assuming a primitive mantle starting composition) such that subsequent assimilation of a defined portion of these carbonatites by uprising melts can account for the Hf isotope variation observed in the Vogelsberg basanites and nephelinites (as defined by the maximum difference in 176Hf/177Hf in the Vogelsberg samples, i.e. between sample VB20 [eHf = 7.4, Hf = 4.6 ppm] and sample VB28 [eHf = 5.0, Hf = 5.0 ppm] where the latter is assumed to represent the least contaminated melt; see Appendix A and Electronic Annex for details). Clearly, this approach results in a minimum age of carbonatite metasomatism, as the real spread in 176Hf/177Hf in the Vogelsberg lavas is likely larger than observed in the restricted number of samples investigated here. The calculated metasomatism model age is shown in Fig. 10 as a function of the Lu/Hf in the assimilated carbonatite (assuming a Hf concentration of 1 ppm) and the degree of carbonatite assimilation by rising

(Л .2

1600 1400 1200 1000 800 600 400 200 0

Lu/Hf in carbonatite

Fig. 10. Model age of carbonatite metasomatism calculated as a function of the Lu/Hf ratio in the metasomatising carbonatite and for different degrees of carbonatite assimilation by basanitic and nephelinitic melts. Calculations are based on Hf concentrations, Lu/Hf and 176Hf/177Hf data of basanites and nephelinites from the Vogelsberg and assume that the variation of 176Hf/177Hf in Vogelsberg lavas results from different degrees of assimilated carbonatite with high Lu/Hf and radiogenic 176Hf/177Hf (see text and Appendix A for details). The older the metasomatising event (i.e. the longer the time interval between metasomatism and subsequent assimilation), and the higher the Lu/Hf in the meta-somatising carbonatite, the less carbonatite assimilation by rising basanitic and nephelinitic melts is required to produce the observed shift in

176Hf/177Hf ratios in the Vogelsberg basanites and nephelinites. Even assuming relatively high degrees of assimilation (>5%) and fairly high Lu/Hf in carbonatites (>5) require that metasomatism occurred several hundred Ma prior to assimilation (numbers on curves denote degree of assimilation, assumed concentration of Hf in the carbonatite = 1 ppm). Also shown is the age of the Hercynian orogeny, during which the European lithospheric mantle was metasomatically overprinted. Average carbonatite Lu/Hf and standard deviation calculated from data given by Hoernle et al. (2002) and Bizimis et al. (2003).

magmas. Although the maximal variation in 176Hf/177Hf in the Vogelsberg samples is relatively small (~2.5 eHf units), the time span between carbonatite metasomatism and assim-ilation/remelting must exceed several 100 Ma, even if relatively high Lu/Hf and high assimilation rates are assumed (Fig. 10). The reason for this is the comparatively low Hf concentration in carbonatites. Even by assuming a carbonatite Hf concentration of ~2-3 ppm, which is at the upper limit of the published range for oceanic and continental carbona-tites (Hoernle et al., 2002; Bizimis et al., 2003; see Electronic Annex), calculated model ages exceed 100 Ma. A shorter time span between metasomatism and assimilation/remelt-ing of carbonatites would require both higher Lu/Hf ratios and unrealistically high Lu and Hf concentrations at assimilation rates greater than 5%. Carbonatite metasomatism of parts of the lithosphere beneath Central Germany is therefore assumed to have occurred long before the onset of Ceno-zoic magmatism.

5.4. Nb/Ta in the subcontinental lithosphere and consequences for the global Nb budget

Overall subchondritic Nb/Ta in most accessible terrestrial silicate reservoirs (i.e. mantle and crust; Barth et al.,

2000; Rudnick et al., 2000; Münker et al., 2003; Weyer et al., 2003; Münker et al., 2004; Pfander et al., 2007; König et al., 2008; Liu et al., 2008; Goss and Kay, 2009) lead to the suggestion, that the silicate Earth is missing some of its Nb. This "missing Nb" may be stored in the Earth's core or within a so far unsampled ("hidden") reservoir in the mantle. The core model is supported by experimental evidence, claiming a more siderophile behaviour of Nb compared to Ta during core formation (Wade and Wood, 2001; Jochum et al., 2002; Mann et al., 2009), and by a systematic variation of Nb/Ta between the silicate portions of Earth, Mars, Moon and asteroids (Münker et al., 2003). The claim for a hidden mantle reservoir with high Nb/Ta such as deeply subducted eclogitic or enriched Hadean crust is based on rutile analyses, Nb/Ta estimates for the continental crust, and analyses of Proterozoic anorthosites (Barth et al., 2000; Kamber and Collerson, 2000; Rudnick et al., 2000; Nebel et al., 2010). High-precision analyses of HFSE in orogenic eclogites have shown, however, that many of these plot on or slightly above the silicate differentiation line in terms of Nb/Ta vs. Zr/Hf, intersecting the chondritic Zr/Hf (34.3 ± 0.3; Munker et al., 2003) at subchondritic Nb/Ta (John et al., 2004; Schmidt et al., 2009). This indicates a Nb deficit in subduction related eclogites and is consistent with the observation that OIBs tapping recycled oceanic crust have also subchondritic Nb/Ta (~14-16; Fig. 2; Pfander et al., 2007). Unlike orogenic eclogites, whose Nb-Ta systematics are comparable to modern MORB (Nb/Ta ~13-16; Buchl et al., 2002; Fig. 2), eclogite xenoliths from cratonic settings tend to be different in their HFSE budget (e.g. Rudnick et al., 2000; Aulbach et al., 2008).

The role of the subcontinental lithosphere in balancing the global Nb budget has been considered previously (Kal-foun et al., 2002; Aulbach et al., 2008). Aulbach et al. (2008) have proposed that ancient (>2.9 Ga) metasomatism led to a slight enrichment of Nb over Ta in some portions of the SCLM, reflected by superchondritic Nb/Ta in rutiles with unradiogenic 176Hf/177Hf ratios. Along with the data presented here (elevated Nb/Ta relative to Zr/Hf and low Zr/Nb as well as on average elevated Nb and Ta contents in continental basalts if compared to OIBs) and in agreement with the inferred superchondritic Nb/La of the continental lithospheric mantle (McDonough, 1991) this provides evidence that the SCLM may host some of the Earth's "missing Nb", most likely selectively added by metasomatic processes (note that some additional portions of the "missing Nb" may be stored in the oceanic lithospheric mantle; see Doucelance et al., 2010). Although the contribution of different metasomatic agents (fluids or low degree melts with carbonatitic to silicic compositions) in the Nb enrichment of the lithospheric mantle is difficult to assess, Nb and Ta are efficiently transported by carbon-atitic melts. These can have extremely high concentrations of Nb and Ta (up to 10,000-times chondritic) that are an order of magnitude higher than those of other incompatible trace elements (Hoernle et al., 2002; Bizimis et al., 2003; Chakhmouradian, 2006), in agreement with their highly incompatible behaviour between mantle phases and car-bonatitic liquids (Dasgupta et al., 2009). Potential Nb stor-

age in the SCLM is therefore most likely associated to car-bonatite phases having also superchondritic Nb/Ta. For example, model calculations indicate that the spread in Nb/Ta observed in the continental basalts is consistent with the assimilation of up to ~7% of a carbonatite with a super-chondritic Nb/Ta of ~260 and a Nb concentration of 300 ppm (see Fig. 5a and Electronic Annex).

Gaillard et al. (2008) suggested that high electrical conductivities in the asthenospheric mantle could be explained by assuming the presence of <0.35 vol.% carbonatitic melt. Their inferred occurrence in the depth range of ~70-200 km agrees with experimental results, in which carbona-tite melts are the first phase on the solidus of CO2-bearing lherzolite at 3 GPa (Dasgupta et al., 2007; Foley et al., 2009). Carbonatite liquids in the asthenospheric mantle may thus be a widespread phenomenon and hence carbon-atite metasomatism of the overlying lithosphere may play a much larger role than so far assumed. The lithospheric mantle, particularly beneath the continents where deep lithospheric roots may prevent enhanced melting, may therefore be a geochemical reservoir that underwent significant carbonatite metasomatism and thus may host a significant amount of the "missing Nb". This raises the following questions: (i) does the lithospheric mantle have a sufficiently high volume (mass) to account for "missing Nb" assuming reasonable degrees of metasomatism, and (ii) if not, which proportion of the "missing Nb" may be potentially stored in the lithospheric mantle? To place constraints on these questions, mass balance calculations have been done as described in Appendix A and by using the parameters given in the Electronic Annex.

The Earth's core was considered by several authors as hosting the "missing Nb" on Earth (e.g. Wade and Wood, 2001; Jochum et al., 2002; Miinker et al., 2003). Based on this assumption, and balancing Nb and Ta between terrestrial and lunar samples and meteorites, Munker et al. (2003) calculated the Nb concentration in the core to be ~0.47 ppm. This corresponds to a total mass of 9.12 x 1020g, which is thus an estimate for the mass of the "missing Nb" (see Appendix A and data in the Electronic Annex) and accounts for ~35% of the total Nb inventory of the Earth (2.61 x 1021 g, calculated from the averaged Nb concentration of 0.437 ppm in chondrites and the mass of the Earth; see Appendix A and Electronic Annex). This estimate agrees well with the amount of "missing Nb" calculated from a mass balance considering the silicate reservoirs of the Earth (continental and oceanic crust, upper and lower mantle) which yields a total Nb deficit of ~37% in the accessible (silicate) portion of the Earth (see also Fig. 11 and data in the Electronic Annex). The latter estimate, however, is critical to the Nb concentration of the lower portion of the mantle. It hosts ~45% of the Nb inventory of the bulk Earth if a Nb concentration of 0.4 ppm is assumed for this reservoir (McDonough and Sun, 1995; Willbold and Stracke, 2006; varying this concentration between 0.2 and 0.6 ppm, the Nb-deficit varies between 59% and 14%).

Making the very simplified assumption that the whole "missing Nb" (~9.12 x 1020 g) is stored in carbonatite components in the continental lithospheric mantle, and the car-

Fig. 11. Sketch illustrating the most important parameters used in modelling the Nb inventory of the subcontinental lithospheric mantle and the potential of the lithospheric mantle to host the "missing Nb" (details see text, Appendix A and Electronic Annex). Nb concentration in the oceanic crust calculated from 3.5 ppm in MORB (Hofmann, 1988) and assuming that the oceanic crust consists of 50% extrusives (sheeted dikes and basalts represented by MORB) and of 50% cumulates (i.e. gabbros) having ~0.35 ppm Nb (see above-mentioned references for further information).

bonatites having 200-500 ppm Nb (see Electronic Annex), ~1.8-4.6% of the SCLM would need to consist of carbon-atites to host all of the "missing Nb". Although this amount in principle seems reasonable if the entire volume of the continental lithosphere (2.78 x 1019 m3; Artemieva, 2006) is considered, it is highly unlikely that carbonatite metasomatism can pervasively affect the whole lithospheric mantle up to the MOHO. Despite the low viscosity of carbonatite melts, their small volume and limited heat content most likely will restrict carbonatite metasomatism to deeper regions within the lithosphere, likely to the thermal boundary layer where such melts rapidly cool and precipitate (McKenzie, 1989). In Fig. 12a, the portion (volume) of metasomatised lithosphere that is required to store all the "missing Nb" on Earth is shown as a function of the Nb concentration in the metasomatising carbonatite melt and the degree of metasomatism. It is evident, that if carbona-tite metasomatism is assumed to be restricted to the lowermost 10-15% of the lithospheric mantle (grey in Fig. 12a), either unrealistically high Nb concentrations (>1500 ppm) or unrealistically high degrees of metasomatism (>5-7%) are required to store all of the "missing Nb" in the subcontinental lithospheric mantle.

To estimate the percent fraction of the "missing Nb" that may be potentially stored in the subcontinental litho-spheric mantle under reasonable conditions, this fraction was calculated as a function of the Nb concentration in the metasomatising melt and the degree of metasomatism, and by assuming that only 10% of the SCLM is affected by metasomatism. The result is shown in Fig. 12b. Given that continental carbonatites may contain up to several thousand ppm of Nb (Bizimis et al., 2003), and taking the average Nb concentration in carbonatites from Bizimis et al. (2003) (i.e. ~1000 ppm) as an upper limit, the subcon-

50 ■c z

CÜ CT 45 .c c

in'"» 40 g .2

■E E 35

o Œ c ® .2 o

0 c OL (5 E

o o û- ~

500 1000

Nb in carbonatite [ppm]

500 1000

Nb in carbonatite [ppm]

Fig. 12. (a) Portion of the continental lithospheric mantle that needs to be affected by metasomatism to account for the whole global Nb deficit as a function of the Nb concentration in the metasomatising carbonatitic melt and the degree of metasomatism (numbers on curves; details see text). (b) Portion of "missing Nb" that may be stored in the metasomatised lithospheric mantle as a function of the Nb concentration in the carbonatite and the degree of metasomatism (numbers on curves). The calculation assumes that only 10% of the lithospheric mantle are affected by metasomatism. Note that up to 30% of the "missing Nb" may be stored in the lithospheric mantle (details see text). Range of Nb concentrations estimated from data given by Bizimis et al. (2003).

of the subcontinental lithospheric mantle in balancing the global Nb budget. The data reveal that the HFSE composition of intraplate volcanic rocks from Central Germany cannot be explained by simple partial melting of common spinel or garnet peridotite sources, but require involvement of metasomatically enriched domains in the subcontinental lithospheric mantle.

Trace element modelling indicates that the observed geo-chemical features can be explained by pooling melts from garnet bearing asthenospheric mantle sources along with melts from lithospheric mantle sources that have been re-enriched in highly incompatible trace elements by metaso-matic overprinting. The lithospheric sources involve amphibole and phlogopite bearing domains. Specific trace element characteristics of the alkaline lavas, such as elevated Nb/Ta and Lu/Hf at extremely low Zr/Nb along with elevated Nb concentrations if compared to OIBs also call for a role of carbonatite metasomatism that affected the subcontinental lithospheric mantle beneath Central Germany. Radiogenic ingrowth due to high Lu/Hf in carbona-tites can account for elevated 176Hf/177Hf in some continental basalts, and constrains the metasomatic event to have taken place significantly before the onset of alkaline volcanism.

Model calculations indicate that up to one third of the Earth's "missing Nb" can be stored in metasomatic domains in the lithospheric mantle overprinted by carbona-tites. The role of this geochemical reservoir may thus be much more important in balancing the global Nb budget than previously thought.

ACKNOWLEDGMENTS

Erik Scherer is thanked for sharing his expertise in using the Münster Micromass Isoprobe. Many thanks to Heide-Marie Baier for lab assistance. Constructive reviews by Vincent Salters, Sonja Aulbach and Olli Nebel, and comments by associate editor Steven Shirey improved the manuscript and are highly appreciated.

tinental lithospheric mantle may account for up to about 30% of the "missing Nb" on Earth if the degree of metasomatism is estimated to be ~3% (grey field in Fig. 12b). Thus, the role of the Earth's core or any other (hidden) reservoir in the mantle in balancing the Nb budget on Earth may be less important than previously estimated.

In summary, it can be demonstrated here that the subcontinental lithospheric mantle can probably not host all of the Earth's "missing Nb". Model calculations, however, indicate that substantial amounts of the "missing Nb" (up to about one third) may be stored in the continental litho-spheric mantle, associated with carbonatite-rich phases that are a result of pervasive metasomatism.

6. CONCLUSIONS

High-precision HFSE and Lu concentration data of continental basalts from different locations in Central Germany are presented in order to assess their potential mantle sources and magmatic evolution, and to constrain the role

APPENDIX A. PARTITION COEFFICIENTS

The selection of appropriate partition coefficients is crucial in trace element modelling of petrogenetic processes. In particular for highly incompatible trace elements such as Nb or Ta published mineral-melt partitioning data may vary an order of magnitude, depending on mineral and melt composition, pressure, temperature and redox conditions during the experimental runs. Values used in this study (given in blue in the Electronic Annex) are identical to those used in Pfander et al. (2007) and are from McDade et al. (2003) for olivine, orthopyroxene and clinopyroxene (except for La in olivine: Kelemen et al., 2003), from Horn et al. (1994) for spinel, and from van Westrenen et al. (1999) for pyrope rich garnet (Py84). Partition coefficients used for amphibole are from Adam and Green (2003) and those for phlogopite from Green et al. (2000). These datasets, in contrast to many others frequently used, provide values for all HFSE including Ta, and have been determined under anhydrous conditions at pressures of

1.5 GPa for olivine, orthopyroxene and clinopyroxene (McDade et al., 2003) and 3.0 GPa for garnet (van Westrenen et al., 1999; for a detailed discussion and comparison of different sets of partition coefficients with an emphasis on HFSE see Pfander et al., 2007).

APPENDIX B. MELTING AND METASOMATISM CALCULATIONS

Melting of the asthenosphere in the garnet and spinel stability field was calculated as non-modal batch melting using the equation of Shaw (1970) and the modes and melting reactions as given in the Electronic Annex. Initial trace element concentrations used are primitive mantle values taken from Hofmann (1988; see Electronic Annex) where the Nb concentration has been reduced to 0.491 ppm to account for the inferred bulk silicate Earth Nb deficit. This deficit is expressed by a subchondritic Nb/Ta of ~14 established in the silicate portion of the Earth after core formation (for details see Münker et al., 2003).

Melting of the lithospheric mantle in the spinel stability field was modelled by using the averaged modal composition and trace element abundances of refractory, fluid-metasomatised spinel peridotites from the Hessian Depression (Central Germany) as given by Hartmann and Wedepohl (1990; modes and trace element concentrations see Electronic Annex).

The melting line of the metasomatised refractory spinel peridotite shown in Fig. 4 was calculated by non-modal batch melting of a source whose trace element and modal composition corresponds to the averaged composition of the L-group peridotite xenoliths (plus 6% amphibole and 3% phlogopite) from the Hessian Depression (Hartmann and Wedepohl, 1990; see Electronic Annex) to which 1% of a metasomatic melt was added. The metasomatic melt has been assumed to be a 0.01% melt fraction produced by non-modal batch melting from a primitive garnet peri-dotite source using modes and initial trace element abundances as given by Salters (1996) and Hofmann (1988), respectively (see Electronic Annex).

APPENDIX C. MODEL AGE OF CARBONATITE METASOMATISM

Model ages of carbonatite metasomatism in the litho-spheric mantle were calculated on the basis of the maximum spread in 176Hf/177Hf in the Vogelsberg samples and by assuming that this spread is produced by assimilation of carbonatites having high Lu/Hf and high 176Hf/177Hf due to radioactive ingrowth over time (for simplicity only assimilation is considered here, but remelting of a carbona-tite bearing peridotite source will yield similar results). The age of metasomatism was then calculated as a function of the Lu/Hf in the carbonatite (i.e. the metasomatised mantle domain) as the time span required to produce a sufficiently high 176Hf/177Hf in the carbonatite such that the assimilation of a defined portion (2%, 5% and 10% in Fig. 10) of it shifts the Hf isotope composition of the least radiogenic sample (VB20) to that of the most radiogenic (VB28) sam-

ple. The Hf concentration of the carbonatite was set to 1 ppm and the 176Hf/177Hf ratio at the time of metasomatism was assumed to be that of the primitive mantle (i.e. the chondritic value at the time of metasomatism; chondritic 176Lu/177Hf today = 0.03 32, 176Hf/177Hf= 0.282772 (Blichert-Toft and Albarede, 1997), Lu-Hf decay constant = 1.865 x 10~u a-1 (Scherer et al., 2001)). This approach is critical to the Hf concentration and hence the Lu/Hf in the assumed model carbonatite (for composition see Electronic Annex). Lower Hf concentrations result in higher model ages. If reasonable degrees of assimilation are assumed (<2-3%), relatively high Lu/Hf (>2-3) are required for the carbonatite or a very old (>1.5 Ga) metasomatism event to explain the spread in 176Hf/177Hf in the Vogelsberg samples.

APPENDIX D. NB BUDGET IN THE LITHOSPHERE

Balancing the Nb inventory of the Earth between different Earth reservoirs and estimating the potential of the subcontinental lithospheric mantle to host parts of the Earth's "missing Nb" was done using the parameters given in the Electronic Annex. The Nb inventory of the bulk Earth was calculated from their mass and the average Nb concentration in chondrites (excluding CV chondrites; Münker et al., 2003), yielding 2.61 x 1021 g. The amount of "missing Nb" was then calculated by two approaches. First, by assuming that all the "missing Nb" is stored in the Earth's core (e.g., Münker et al., 2003). Taking the Nb concentration of the core and its mass, this corresponds to 9.12 x 1020g of "missing Nb", corresponding to 34.9% with respect to the total Nb inventory of the Earth. In a second approach, Nb was balanced between the bulk Earth and the Earth's silicate reservoirs (continental crust, oceanic crust, depleted MORB mantle and enriched mantle), which in sum host 1.65 x 1021 g Nb, yielding a deficit of 9.60 x 1020 g (36.8%) with respect to the total Nb inventory of the Earth, in good agreement to the first approach. By assuming different degrees of metasomatism (3%, 5%, 10% and 20% in Fig. 12a), we then calculated the fraction of the subcontinental lithospheric mantle that needs to be metasoma-tised to host all of the "missing Nb" as a function of the Nb concentration in the carbonatite. In a second approach, the percent fraction of the "missing Nb" that can be potentially stored in the subcontinental lithospheric mantle was calculated as a function of the degree of metasomatism and the Nb concentration in the carbona-tite, and making the assumption that only 10% of the (lowermost) subcontinental lithosphere is affected by metasomatism. Note that the mass of the continental lithosphere was calculated from the volume (2.78 x 1019m3; Artemieva, 2006) and the average density of the upper mantle (3.58 g/cm3; Dziewonski and Anderson, 1981).

APPENDIX E. SUPPLEMENTARY DATA

Supplementary data associated with this article can be found, in the online version, at doi:10.1016/ j.gca.2011.11.017.

REFERENCES

Adam J. and Green T. (2003) The influence of pressure, mineral composition and water on trace element partitioning between clinopyroxene, amphibole and basanitic melts. Eur. J. Mineral. 15, 831-841.

Adam J. and Green T. (2006) Trace element partitioning between mica- and amphibole-bearing garnet lherzolite and hydrous basanitic melt: 1. Experimental results and the investigation of controls on partitioning behaviour. Contrib. Mineral. Petrol. 152, 1-17.

Albarede F. (2003) Geochemistry: An Introduction. Cambridge University Press, Cambridge.

Artemieva I. M. (2006) Global 1°x1° thermal model TC1 for the continental lithosphere: implications for lithosphere secular evolution. Tectonophysics 416, 245-277.

Asmerom Y. and Walker R. J. (1998) Pb and Os isotopic constraints on the composition and rheology of the lower crust. Geology 26, 359-362.

Aulbach S., Griffin W. L., Pearson N. J. and O'Reilly S. Y. (2008) Subcontinental lithospheric mantle origin of high Nb/Ta in eclogite xenoliths. Nat. Geosci. 1, 468-472.

Barth M. G., McDonough W. F. and Rudnick R. L. (2000) Tracking the budget of Nb and Ta in the continental crust. Chem. Geol. 165, 197-213.

Beccaluva L., Bianchini G., Coltorti M., Perkins W. T., Siena F., Vaccaro C. and Wilson M. (2001) Multistage evolution of the European lithospheric mantle: new evidence from Sardinian peridotite xenoliths. Contrib. Mineral. Petrol. 142, 284-297.

Bedini R. M., Bodinier J.-L., Dautria J.-M. and Morten L. (1997) Evolution of LILE-enriched small melt fractions in the litho-spheric mantle: a case study from the East African Rift. Earth Planet. Sci. Lett. 153, 67-83.

Bizimis M., Salters V. J. M. and Dawson J. B. (2003) The brevity of carbonatite sources in the mantle: evidence from Hf isotopes. Contrib. Mineral. Petrol. 145, 281-300.

Blichert-Toft J. and Albarede F. (1997) The Lu-Hf isotope geochemistry of chondrites and the evolution of the mantle-crust system. Earth Planet. Sci. Lett. 148, 243-258.

Blichert-Toft J., Chauvel C. and Albarede F. (1997) Separation of Hf and Lu for high-precision isotope analysis of rock samples by magnetic sector-multiple collector ICP-MS. Contrib. Mineral. Petrol. 127, 248-260.

Blusztajn J. and Hart S. (1989) Sr, Nd, and Pb isotopic character of Tertiary basalts from southwest Poland. Geochim. Cosmochim. Acta 53, 2689-2696.

Blusztajn J. and Hegner E. (2002) Osmium isotopic systematics of melilitites from the Tertiary Central European Volcanic Province in SW Germany. Chem. Geol. 189, 91-103.

Bodinier J. L., Merlet C., Bedini R. M., Simien F., Remaidi M. and Garrido C. J. (1996) Distribution of niobium, tantalum, and other highly incompatible trace elements in the lithospheric mantle: the spinel paradox. Geochim. Cosmochim. Acta 60, 545550.

Bogaard P. J. F. and Worner G. (2003) Petrogenesis of basanitic to tholeiitic volcanic rocks from the Miocene Vogelsberg, Central Germany. J. Petrol. 44, 569-602.

Brey G., Brice W. R., Ellis D. J., Green D. H., Harris K. L. and Ryabchikov I. D. (1983) Pyroxene-carbonate reactions in the upper mantle. Earth Planet. Sci. Lett. 62, 63-74.

Buchl A., Munker C., Mezger K. and Hofmann A. W. (2002) High-precision Nb/Ta and Zr/Hf ratios in global MORB. Geochim. Cosmochim. Acta (Spec. Suppl.) 66, A108.

Buikin A., Trieloff M., Hopp J., Althaus T., Korochantseva E., Schwarz W. H. and Altherr R. (2005) Noble gas isotopes suggest deep mantle plume source of late Cenozoic mafic

alkaline volcanism in Europe. Earth Planet. Sci. Lett. 230, 143162.

Carlson R. W. and Nowell G. M. (2001) Olivie-poor sources for mantle-derived magmas: Os and Hf isotopic evidence from potassic magmas of the Colorado Plateau. Geochem. Geophys. Geosys. 2.

Chakhmouradian A. R. (2006) High-field-strength elements in carbonatitic rocks: geochemistry, crystal chemistry and significance for constraining the sources of carbonatites. Chem. Geol. 235, 138-160.

Chauvel C., McDonough W. F., Guille G., Maury R. and Duncan R. (1997) Contrasting old and young volcanism in Rurutu Island, Austral chain. Chem. Geol. 139, 125-143.

Chesley J. T. and Ruiz J. (1998) Crust-mantle interaction in large igneous provinces: implications from the Re-Os isotope systematics of the Columbia River flood basalts. Earth Planet. Sci. Lett. 154, 1-11.

Class C. and Goldstein S. L. (1997) Plume-lithosphere interactions in the ocean basins: constraints from the source mineralogy. Earth Planet. Sci. Lett. 150, 245-260.

Class C., Goldstein S. L. and Shirey S. B. (2009) Osmium isotopes in Grande Comore lavas: a new extreme among a spectrum of EM-type mantle endmembers. Earth Planet. Sci. Lett. 284, 219227.

Dasgupta R., Hirschmann M. M. and Withers A. C. (2004) Deep global cycling of carbon constrained by the solidus of anhydrous, carbonated eclogite under upper mantle conditions. Earth Planet. Sci. Lett. 227, 73-85.

Dasgupta R., Hirschmann M. M. and Smith D. R. (2007) Partial melting experiments of peridotite + CO2 at 3 GPa and genesis of alkalic ocean island basalts. J. Petrol. 48, 2093-2124.

Dasgupta R., Hirschmann M. M., McDonough W. F., Spiegelman M. and Withers A. C. (2009) Trace element partitioning between garnet lherzolite and carbonatite at 6.6 and 8.6 GPa with applications to the geochemistry of the mantle and of mantle-derived melts. Chem. Geol. 262, 57-77.

DePaolo D. (1981) Trace element and isotopic effects of combined wallrock assimilation and fractional crystallization. Earth Planet. Sci. Lett. 53, 189-202.

Doucelance R., Hammouda T., Moreira M. and Martins J. C. (2010) Geochemical constraints on depth of origin of oceanic carbonatites: the Cape Verde case. Geochim. Cosmochim. Acta 74, 7261-7282.

Downes H. (2001) Formation and modification of the shallow subcontinental lithospheric mantle: a review of geochemical evidence from ultramafic xenolith suites and tectonically emplaced ultramafic massifs of western and central Europe. J. Petrol. 42, 233-250.

Dupuy C., Liotard J. M. and Dostal J. (1992) Zr/Hf fractionation in intraplate basaltic rocks: carbonate metasomatism in the mantle source. Geochim. Cosmochim. Acta 56, 2417-2423.

Dziewonski A. M. and Anderson D. L. (1981) Preliminary reference Earth Model. Phys. Earth Planet. Inter 25, 297-356.

Ellam R. M., Carlson R. W. and Shirey S. B. (1992) Evidence from Re-Os isotopes for plume-lithosphere mixing in Karoo flood basalt genesis. Nature 359, 718-721.

Fekiacova Z., Mertz D. F. and Renne P. (2003) Geodynamic setting of the Hocheifel volcanism, western Germany. Geophys. Res. Abstr. 5, 10355.

Foley S. F., Barth M. G. and Jenner G. A. (2000) Rutile/melt partition coefficients for trace elements and an assessment of the influence of rutile on the trace element characteristics of subduction zone magmas. Geochim. Cosmochim. Acta 64, 933-938.

Foley S., Yaxley G. M., Rosenthal A., Buhre S., Kiseeva E. S., Rapp R. P. and Jacob D. E. (2009) The composition of near-

solidus melts of peridotite in the presence of CO2 and H2O between 40 and 60 kbar. Lithos 112, 274-283.

Furman T. and Graham D. (1999) Erosion of lithospheric mantle beneath the East African Rift system: geochemical evidence from the Kivu volcanic province. Lithos 48, 237-262.

Gaillard F., Malki M., Iacono-Marziano G., Pichavant M. and Scaillet B. (2008) Carbonatite melts and electrical conductivity in the asthenosphere. Science 322, 1363-1365.

Goes S., Spakman H. and Bijwaard H. (1999) A lower mantle source for Central European volcanism. Science 286, 19281932.

Goss A. R. and Kay S. M. (2009) Extreme high field strength element (HFSE) depletion and near-chondritic Nb/Ta ratios in Central Andean adakite-like lavas (28 S, 68 W). Earth Planet. Sci. Lett. 279, 97-109.

Granet M., Achauer U. and Wilson M. (1995) Nachweis eines Mantelplumes unter Frankreichs Zentralmassiv durch seismische Sondierung. Spektr. d. Wissensch., 28-33.

Green T. H. and Pearson N. J. (1987) An experimental study of Nb and Ta partitioning between Ti-rich minerals and silicate liquids at high pressure and temperature. Geochim. Cosmochim. Acta 51, 55-62.

Green D. H., Blundy J. D., Adam J. and Yaxley G. M. (2000) SIMS determination of trace element partition coefficients between garnet, clinopyroxene and hydrous basaltic liquids at 2-7.5 GPa and 1080-12008 °C. Lithos 53, 165-187.

Grunewald S., Weber M. and Kind R. (2001) The upper mantle under Central Europe: indications for the Eifel plume. Geophys. J. Int. 147, 590-601.

Haase K. M., Goldschmidt B. and Garbe-Schonberg C.-D. (2004) Petrogenesis of tertiary continental intra-plate lavas from the Westerwald region. J. Petrol. 45, 883-905.

Hart S., Carlson R. W. and Shirey S. B. (1997) Radiogenic Os in primitive basalts from the northwestern USA: implications for petrogenesis. Earth Planet. Sci. Lett. 150, 103-116.

Hartmann G. and Wedepohl K. H. (1990) Metasomatically altered peridotite xenoliths from the Hessian Depression (Northwest Germany). Geochim. Cosmochim. Acta 54, 71-86.

Hauri E., Shimizu N., Dieu J. J. and Hart S. R. (1993) Evidence for hotspot-related carbonatite metasomatism in the oceanic upper mantle. Nature 365, 221-226.

Hegner E., Walter H. J. and Satir M. (1995) Pb-Sr-Nd isotopic compositions and trace element geochemistry of megacrysts and melilitites from the Tertiary Urach volcanic field: source composition of small volume melts under SW Germany. Contrib. Mineral. Petrol. 122, 322-335.

Hoernle K., Zhang Y. S. and Graham D. (1995) Seismic and geochemical evidence for large-scale mantle upwelling beneath the eastern Atlantic and western and central Europe. Nature 374, 34-39.

Hoernle K., Tilton G., Le Bas M. J., Duggen S. and GarbeSchonberg D. (2002) Geochemistry of oceanic carbonatites compared with continental carbonatites: mantle recycling of oceanic crustal carbonate. Contrib. Mineral. Petrol. 142, 520542.

Hofmann A. W. (1988) Chemical differentiation of the Earth: the relationship between mantle, continental crust and oceanic crust. Earth Planet. Sci. Lett. 90, 297-314.

Horn I., Foley S. F., Jackson S. E. and Jenner G. A. (1994) Experimentally determined partitioning of high-field strength-elements and selected transition-elements between spinel and basaltic melt. Chem. Geol. 117, 193-218.

Ionov D., Dupuy C., O'Reilly S. Y., Kopylova M. G. and Genshaft Y. S. (1993) Carbonated peridotite xenoliths from Spitsbergen: implications for trace element signature of mantle carbonate metasomatism. Earth Planet. Sci. Lett. 119, 283-297.

Ionov D., Bodinier J.-L., Mukasa S. B. and Zanetti A. (2002) Mechanisms and sources of mantle metasomatism: major and trace element compositions of peridotite xenoliths from Spitsbergen in the context of numerical modelling. J. Petrol. 43, 2219-2259.

Jochum K. P., Hofmann A. W., Seufert M. and Stoll B. (2002) Niobium in planetary cores: consequences for the interpretation of terrestrial Nb systematics. Geochim. Cosmochim. Acta (Spec. Suppl.) 66, A368.

John T., Scherer E. E., Haase K. and Schenk V. (2004) Trace element fractionation during fluid-induced eclogitization in a subducting slab: trace element and Lu-Hf-Sm-Nd isotope systematics. Earth Planet. Sci. Lett. 227, 441-456.

Jorgensen J. O. and Holm P. M. (2002) Temporal variation and carbonatite contamination in primitive ocean island volcanics from Sao Vicente, Cape Verde Islands. Chem. Geol. 192, 249267.

Jung S. (1995) Geochemistry and petrogenesis of rift-related Tertiary alkaline rocks from the Rhön area (central Germany). N. Jb. Mineral. Abh. 169, 193-226.

Jung S. and Masberg P. (1998) Major- and trace-element systematics and isotope geochemistry of Cenozoic mafic volcanic rocks from the Vogelsberg (central Germany). Constraints on the origin of continental alkaline and tholeiitic basalts and their mantle sources. J. Volc. Geotherm. Res. 86, 151-177.

Jung S. and Hoernes S. (2000) The major- and trace-element and isotope (Sr, Nd, O) geochemistry of Cenozoic alkaline rift-type volcanic rocks from the Rhon area (Central Germany): petrology, mantle source characteristics and implications for asthenosphere-lithosphere interactions. J. Volc. Geotherm. Res. 99, 27-53.

Jung S., Pfander J. A., Brügmann G. and Stracke A. (2005) Sources of primitive alkaline volcanic rocks from the central European volcanic province (Rhön, Germany) inferred from Hf, Os and Pb isotopes. Contrib. Mineral. Petrol. 150, 546-559.

Jung C., Jung S., Hoffer E. and Berndt J. (2006) Petrogenesis of tertiary mafic alkaline magmas in the Hocheifel, Germany. J. Petrol. 47, 1637-1671.

Jung S., Pfander J. A., Brauns M. and Maas R. (2011) Crustal contamination and mantle source characteristics in continental intra-plate volcanic rocks: Pb, Hf and Os isotopes from central European volcanic province basalts. Geochim. Cosmochim. Acta 75, 2664-2683.

Kalfoun F., Ionov D. and Merlet C. (2002) HFSE residence and Nb/Ta ratios in metasomatised, rutile bearing mantle perido-tites. Earth Planet. Sci. Lett. 182, 1-17.

Kamber B. and Collerson K. D. (2000) Role of 'hidden' deeply subducted slabs in mantle depletion. Chem. Geol. 166, 241-254.

Kelemen P. B., Yogodzinski G. M. and Scholl D. W. (2003) Along-strike variation in the Aleutian Island Arc: genesis of high Mg# andesite and implications for continental crust. In Inside the Subduction Factory, Geophys. Monogr. Series,, vol. 3 (ed. J. Eiler). AGU, Washington, DC, pp. 223-276.

Keller J. (1981) Carbonatitic volcanism in the Kaiserstuhl alkaline complex: evidence for highly fluid carbonatitic melts at the Earth's surface. J. Volc. Geotherm. Res. 9, 423-431.

Keyser M., Ritter J. R. R. and Jordan M. (2002) 3D shear-wave velocity structure of the Eifel plume, Germany. Earth Planet. Sci. Lett. 203, 59-82.

Klemme S., Prowatke S., Hametner K. and Gunther D. (2005) Partitioning of trace elements between rutile and silicate melts: implications for subduction zones. Geochim. Cosmochim. Acta 69, 2361-2371.

König S. and Schuth S. (2011) Deep melting of old subducted oceanic crust recorded by superchondritic Nb/Ta in modern island arc lavas. Earth Planet. Sci. Lett. 301, 265-274.

König S., Munker C., Schuth S. and Garbe-Schönberg D. (2008) Mobility of tungsten in subduction zones. Earth Planet. Sci. Lett. 274, 82-92.

Konzett J. and Ulmer P. (1999) The stability of hydrous potassic phases in lherzolite mantle: an experimental study to 9.5 GPa in simplified and natural bulk compositions. J. Petrol. 40, 629652.

Lippolt H. J. (1982) K/Ar determinations and the correlation of Tertiary volcanic activity in Central Europe. Geologisches Jahrbuch Reihe D: Mineralogie, Petrographie, Geochemie, Lagerstättenkunde D52, 113-135.

Lippolt H. J., Troesch M. and Hess J. C. (1990) Excess argon and dating of quaternary Eifel volcanism, IV. Common argon with high and lower-than-atmospheric 40Ar/36Ar ratios in phonolitic rocks, East Eifel, F.R.G. Earth Planet. Sci. Lett. 101, 19-33.

Liu Y., Gao S., Kelemen P. B. and Xu W. (2008) Recycled crust controls contrasting source compositions of Mesozoic and Cenozoic basalts in the North China Craton. Geochim. Cosmochim. Acta 72, 2349-2376.

Lloyd F. E. and Bailey D. K. (1969) Carbonatite in the tufts of the West Eifel, Germany. Contrib. Mineral. Petrol. 23, 136-139.

Lustrino M. and Wilson M. (2007) The circum-mediterranean anorogenic Cenozoic igneous province. Earth Planet. Sci. Lett. 81, 1-65.

Mann U., Frost D. J. and Rubie D. C. (2009) Evidence for high-pressure core-mantle differentiation from the metal-silicate partitioning of lithophile and weakly-siderophile elements. Geochim. Cosmochim. Acta 73, 7360-7386.

McBride J. S., Lambert D. D., Nicholls I. A. and Price R. C. (2001) Osmium isotopic evidence for crust-mantle interaction in the genesis of continental intraplate basalts from the Newer Volcanics Province, Southeastern Australia. J. Petrol. 42, 1197-1218.

McDade P., Blundy J. D. and Wood B. J. (2003) Trace element partitioning on the Tinaquillo Lherzolite solidus at 1.5 GPa. Phys. Earth Planet. Inter. 139, 129-147.

McDonough W. F. (1991) Partial melting of subducted oceanic crust and isolation of its residual eclogitic lithology. Philos. Trans. R. Soc. Lond. 335, 407-418.

McDonough W. F. and Sun S. S. (1995) The composition of the earth. Chem. Geol. 120, 223-253.

McKenzie D. (1989) Some remarks on the movement of small melt fractions in the mantle. Earth Planet. Sci. Lett. 95, 53-72.

Mengel K. and Green D. H. (1986) Stability of amphibole and phlogopite in metasomatized peridotite under water-saturated and water-undersaturated conditions. In Kimberlites and Related Rocks 1 (eds. J. Ross, A. L. Jaques, J. Ferguson, D. H. Green, S. Y. O'Reilly, R. V. Danchin and A. J. A. Janse). Blackwell Scientific, Oxford, pp. 571-581.

Moine B. N., Grégoire M., O'Reilly S. Y., Delpech G., Sheppard S. M. F., Lorand J. P., Renac P., Giret A. and Cottin J. Y. (2004) Carbonatite melt in oceanic upper mantle beneath the Kergue-len Archipelago. Lithos 75, 239-252.

Münker C., Weyer S., Scherer E. and Mezger K. (2001) Separation of high field strength elements (Nb, Ta, Zr, Hf) and Lu from rock samples for MC-ICPMS measurements. Geochem. Geo-phys. Geosys. 2 (Paper No. 2001GC000183).

Münker C., Pfander J. A., Weyer S., Büchl A., Kleine T. and Mezger K. (2003) Evolution of planetary cores and the Earth-Moon system from Nb/Ta systematics. Science 301, 84-87.

Münker C., Worner G., Yogodzinski G. and Churikova T. (2004) Behaviour of high field strength elements in subduction zones: constraints from Kamchatka-Aleutian arc lavas. Earth Planet. Sci. Lett. 224, 275-293.

Nebel O., van Westrenen W., Vroon P. Z., Wille M. and Raith M. M. (2010) Deep mantle storage of the Earth's missing niobium

in late-stage residual melts from a magma ocean. Geochim. Cosmochim. Acta 74, 4392-4404.

Olafsson M. and Eggler D. H. (1983) Phase relations of amphibole-carbonate and phlogopite-carbonate peridotite: petrologic constraints on the asthenosphere. Earth Planet. Sci. Lett. 64, 305-315.

Pearson D. G., Carlson R. W., Shirey S. B., Boyd F. R. and Nixon P. H. (1995a) Stabilisation of Archaean lithospheric mantle: a Re-Os isotope study of peridotite xenoliths from the Kaapvaal craton. Earth Planet. Sci. Lett. 134, 341-357.

Pearson D. G., Shirey S. B., Carlson R. W., Boyd F. R., Pokhilenko N. P. and Shimizu N. (1995b) Re-Os, Sm-Nd and Rb-Sr isotope evidence for thick Archaean lithospheric mantle beneath the Siberian craton modified by multi-stage metasomatism. Geochim. Cosmochim. Acta 59, 959-977.

Pfander J. A., Münker C., Stracke A. and Mezger K. (2007) Nb/Ta and Zr/Hf in ocean island basalts - implications for crust-mantle differentiation and the fate of Niobium. Earth Planet. Sci. Lett. 254, 158-172.

Pilet S., Baker M. B. and Stolper E. M. (2008) Metasomatized lithosphere and the origin of alkaline lavas. Science 320, 916919.

Prelevic D., Stracke A., Foley S. F., Romer R. L. and Conticelli S. (2010) Hf isotope compositions of Mediterranean lamproites: mixing of melts from asthenosphere and crustally contaminated mantle lithosphere. Lithos 119, 297-312.

Rabinowicz M., Ricard Y. and Grégoire M. (2002) Compaction in a mantle with a very small melt concentration: implications for the generation of carbonatitic and carbonate-bearing high alkaline mafic melt impregnations. Earth Planet. Sci. Lett. 203, 205-220.

Reisberg L. C., Zindler A., Marcantonio F., White W., Wyman D. and Weaver B. (1993) Os isotope systematics in ocean island basalts. Earth Planet. Sci. Lett. 120, 149-167.

Ritter J. R. R., Jordan M., Christensen U. R. and Achauer U. (2001) A mantle plume below the Eifel volcanic fields, Germany. Earth Planet. Sci. Lett. 186, 7-14.

Rosenthal A., Foley S. F., Pearson D. G., Nowell G. M. and Tappe S. (2009) Petrogenesis of strongly alkaline primitive volcanic rocks at the propagating tip of the western branch of the East African Rift. Earth Planet. Sci. Lett. 284, 236-248.

Rudnick R. L. and Gao S. (2003) Composition of the continental crust. In Treatise on Geochemistry, vol. 3 (eds. H. D. Holland and K. K. Turekian). Elsevier Ltd., pp. 1-64.

Rudnick R. L., McDonough W. F. and Chappell B. W. (1993) Carbonatite metasomatism in the northern Tanzanian mantle: petrographic and geochemical characteristics. Earth Planet. Sci. Lett. 114, 463-475.

Rudnick R. L., Barth M., Horn I. and McDonough W. F. (2000) Rutile-bearing refractory eclogites: missing link between continents and depleted mantle. Science 287, 278-281.

Salters V. J. M. (1996) The generation of mid-ocean ridge basalts from the Hf and Nd isotope perspective. Earth Planet. Sci. Lett. 141, 109-123.

Salters V. J. M. and Stracke A. (2004) Composition of the depleted mantle. Geochem. Geophys. Geosyst. 5, Q05004. doi:10.1029/ 2003GC000597.

Scherer E., Münker C. and Mezger K. (2001) Calibration of the Lutetium-Hafnium Clock. Science 293, 683-687.

Schiano P., Clocchiatti R., Shimizu N., Weis D. and Mattielli N. (1994) Cogenetic silica-rich and carbonate-rich melts trapped in mantle minerals in Kerguelen ultramafic xenoliths: implications for metasomatism in the oceanic upper mantle. Earth Planet. Sci. Lett. 123, 167-178.

Schmidt G. and Snow J. (2002) Os isotopes in mantle xenoliths from the Eifel volcanic field and the Vogelsberg (Germany): age

constraints on the lithospheric mantle. Contrib. Mineral. Petrol. 143, 694-705.

Schmidt M. W., Dardon A., Chazot G. and Vannucci R. (2004) The dependence of Nb and Ta rutile-melt partitioning on melt composition and Nb/Ta fractionation during subduction processes. Earth Planet. Sci. Lett. 226, 415-432.

Schmidt A., Weyer S., John T. and Brey G. P. (2009) HFSE systematics of rutile-bearing eclogites: new insights into subduction zone processes and implications for the earth's HFSE budget. Geochim. Cosmochim. Acta 73, 455-468.

Shaw D. M. (1970) Trace element fractionation during anatexis. Geochim. Cosmochim. Acta 34, 237-243.

Shirey S. B. and Walker R. J. (1998) The Re-Os isotope system in cosmochemistry and high-temperature geochemistry. Annu. Rev. Earth Planet. Sci. 26, 423-500.

Spera F. J. and Bohrson W. A. (2001) Energy-constrained open-system magmatic processes I: General model and energy-constrained assimilation and fractional crystallization (EC-AFC) formulation. J. Petrol. 42, 999-1018.

Stosch H.-G. and Seck H. A. (1980) Geochemistry and mineralogy of two spinel peridotite suites from Dreiser Weiher, West Germany. Geochim. Cosmochim. Acta 44, 457-470.

Stosch H.-G. and Lugmair G. W. (1986) Trace element and Sr and Nd isotope geochemistry of peridotite xenoliths from the Eifel (West Germany) and their bearing on the evolution of the subcontinental lithosphere. Earth Planet. Sci. Lett. 80, 281-298.

Stracke A., Hofmann A. W. and Hart S. R. (2005) FOZO, HIMU and the rest of the mantle zoo. Geochem. Geophys. Geosyst. 6.

Sun S.-S. and McDonough W. F. (1989) Chemical and isotopic systematics of oceanic basalts: implications for mantle compositions and processes. In Magmatism in Ocean Basins, vol. 42 (eds. A. D. Saunders and M. J. Norry). Geological Society, Special Publications, London, pp. 313-345.

Tappe S., Foley S., Stracke A., Romer R. L., Kjarsgaard B. A., Heaman L. M. and Joyce N. (2007) Craton reactivation on the Labrador Sea margins: 40Ar/39Ar age and Sr-Nd-Hf-Pb isotope constraints from alkaline and carbonatite intrusives. Earth Planet. Sci. Lett. 256, 433-454.

Tappe S., Foley S. F., Kjarsgaard B. A., Romer R. L., Heaman L. M., Stracke A. and Jenner G. (2008) Between carbonatite and lamproite - diamondiferous Torngat ultramafic lamprophyres formed by carbonate-fluxed melting of cratonic MARID-type metasomes. Geochim. Cosmochim. Acta 72, 3258-3286.

Tiepolo M., Vannucci R., Oberti R., Foley S., Bottazzi P. and Zanetti A. (2000) Nb and Ta incorporation and fractionation in titanian pargasite and kaersutite: crystal-chemical constraints and implications for natural systems. Earth Planet. Sci. Lett. 176, 185-201.

van Achterbergh E., Griffin W. L., Ryan C. G., O'Reilly S. Y., Pearson N. J., Kivi K. and Doyle D. J. (2004) Melt inclusions from the deep Slave lithosphere: implications for the origin and evolution of mantle-derived carbonatite and kimberlite. Lithos 74, 461-474.

van Westrenen W., Blundy J. and Wood B. (1999) Crystal-chemical controls on trace element partitioning between garnet and anhydrous silicate melt. Am. Mineral. 84, 838-847.

Wade J. and Wood B. J. (2001) The Earth's 'missing' Niobium may be in the core. Nature 409, 75-78.

Walker R. J., Carlson R. W., Shirey S. B. and Boyd F. R. (1989) Os, Sr, Nd and Pb isotope systematics of southern African peridotite xenoliths: implications for the chemical evolution of subcontinental mantle. Geochim. Cosmochim. Acta 53, 15831595.

Wedepohl K. H. and Baumann A. (1999) Central European Cenozoic plume volcanism with OIB characteristics and indi-

cations of a lower mantle source. Contrib. Mineral. Petrol. 136, 225-239.

Wedepohl K. H., Gohn E. and Hartmann G. (1994) Cenozoic alkali basaltic magmas of western Germany and their products of differentiation. Contrib. Mineral. Petrol. 115, 253-278.

Weyer S., Münker C., Rehkamper M. and Mezger K. (2002) Determination of ultra low Nb, Ta, Zr, and Hf concentrations and precise Nb/Ta and Zr/Hf ratios by isotope dilution analyses with multiple collector ICP-MS. Chem. Geol. 187, 295-313.

Weyer S., Munker C. and Mezger K. (2003) Nb/Ta, Zr/Hf and REE in the depleted mantle: implications for the differentiation history of the crust-mantle system. Earth Planet. Sci. Lett. 205, 309-324.

Widom E., Hoernle K. A., Shirey S. B. and Schmincke H. E. (1999) Os isotope systematics in the Canary Islands and Madeira; lithospheric contamination and mantle plume signatures. J. Petrol. 40, 279-296.

Willbold M. and Stracke A. (2006) Trace element composition of mantle end-members: implications for recycling of oceanic and upper and lower continental crust. Geochem. Geophys. Geosyst. 7, Q04004. doi:10.1029/2005GC001005.

Willbold M., Hegner E., Stracke A. and Rocholl A. (2009) Continental geochemical signatures in dacites from Iceland and implications for models of early Archaean crust formation. Earth Planet. Sci. Lett. 279, 44-52.

Wilson M. and Downes H. (1992) Mafic alkaline magmatism associated with the European Cenozoic rift system. Tectono-physics 208, 173-182.

Wilson M. and Patterson R. (2001) Intra-plate magmatism related to hot fingers in the upper mantle: evidence from the tertiary-quaternary volcanic province of western and central Europe. In Mantle Plumes: their Identification through Time, vol. 352 (eds. R. Ernst and K. Buchan). Geological Society of America, Special Papers, pp. 37-58.

Wilson M., Rosenbaum J. M. and Dunworth E. A. (1995) Melilitites: partial melts of the thermal boundary layer? Contrib. Mineral. Petrol. 119, 181-196.

Witt-Eickschen G. (2007) Thermal and geochemical evolution of the shallow subcontinental lithospheric mantle beneath the Eifel: constraints from mantle xenoliths, a review. In Mantle Plumes (eds. J. R. R. Ritter and U. R. Christensen). Springer, pp. 323-338.

Worner G., Zindler A., Staudigel H. and Schmincke H.-U. (1986) Sr, Nd, and Pb isotope geochemistry of tertiary and quaternary alkaline volcanics from West Germany. Earth Planet. Sci. Lett. 79, 107-119.

Xiong X. L., Adam T. J. and Green T. H. (2005) Rutile stability and rutile/melt HFSE partitioning during partial melting of hydrous basalt: Implications for TTG genesis. Chem. Geol. 218, 339-359.

Yaxley G. M., Crawford A. J. and Green D. H. (1991) Evidence for carbonatite metasomatism in spinel peridotite xenoliths from western Victoria, Australia. Earth Planet. Sci. Lett. 107, 305317.

Yaxley G. M., Green D. H. and Kamenetsky V. (1998) Carbonatite metasomatism in the southeastern Australian lithosphere. J. Petrol. 39, 1917-1939.

Zinngrebe E. and Foley S. F. (1995) Metasomatism in mantle xenoliths from Gees, West Eifel, Germany: evidence for the genesis of calc-alkaline glasses and metasomatic Ca-enrichment. Contrib. Mineral. Petrol. 122, 79-96.

Associate editor: Steven B. Shirey