Geochemistry, Geophysics, Geosystems

Theistareykir revisited



[1] Postglacial basalts from Theistareykir (younger than 10,000 years), northern Iceland, define the depleted end of the spectrum of chemical and isotopic compositions observed in Icelandic volcanics but extend to some of the most enriched chemical and isotopic compositions found in Icelandic tholeiites. A spectacular feature of these basalts is the impressive correlations observed between radiogenic isotope ratios (Sr, Nd, Hf, and Pb) and almost the entire spectrum of major and trace element concentrations and ratios. The radiogenic isotope and major and trace element compositions are little affected by crystal fractionation and are essentially unaffected by interaction with the preexisting crust. The Theistareykir basalts must therefore be relatively close in composition to primary melts from the mantle. Consequently, their chemical and isotopic compositions provide a unique opportunity to investigate the nature of melting beneath Iceland and the geochemical character and origin of the mantle source. Large variations in incompatible element abundances require source heterogeneity, as well as variable extents of melting, to be important factors in determining the final chemical composition of the melts. Melting integrates over a large pressure range and is dominated by melting a depleted peridotite similar to the ambient depleted North Atlantic mantle. The isotopically enriched component is of relatively minor abundance and probably has a lower solidus temperature compared to the depleted component. More than one isotopically enriched component must be involved, but it is difficult to identify the end-member compositions using those of the lavas because of preeruptive averaging and damping of the enriched isotopic signals by mixing with the ambient depleted mantle or melts thereof, suggesting that the isotopic signals in Icelandic melts represent a somewhat muted isotopic signal of the enriched component(s) in the Icelandic source mantle. Comparison of the isotopic arrays of Icelandic basalts with those of global OIB suggests that the dominant enriched component may have a HIMU (high μ) affinity rather than representing a component similar to the enriched end of the Iceland isotopic arrays, and small amounts of an enriched component similar to enriched mantle (EMI)- type OIB sources are probably also involved.

1. Introduction

[2] From a geochemical perspective, the principal goal for studies of ocean island and mid-ocean ridge basalt (OIB and MORB) petrogenesis is to quantify the competing effects of pressure, temperature, source composition, as well as melting and melt aggregation style with regard to the final chemical and isotopic composition of the erupted melt products. Although numerous studies have documented isotopic variability in MORBs and OIBs on different spatial and temporal scales, especially the relationship between partial melting and sampling of isotopically different portions of the source remains poorly understood. This is partly because elemental concentrations are, in contrast to isotopic ratios, affected by closed system fractionation. Both isotopic ratios and elemental concentrations can also be altered by crustal assimilation. It is therefore only possible to use elemental concentrations to constrain the influence of melting and source heterogeneity if the effects of fractionation and assimilation are understood. Additional complexity in interpreting elemental concentrations is added by polybaric fractional melting and by mixing of the resulting magmas from different depths, which can cause large compositional variations in the melt, even if the source is homogeneous before melting starts. These various effects have been combined by different authors in different ways to produce a large variety of models that can account for the geochemical observations in MORBs and OIBs, sometimes even in case of only a single locality. Iceland, the type locality for plume-ridge interaction [Hart et al., 1973; Schilling, 1973], is a typical example. Moreover, most data sets, for Iceland or any other MORB or OIB locality, contain a limited set of isotopic ratios and elemental concentrations which are often determined on samples from locations that are widely separated in space and time. It is, however, clear that different types of measurements provide different constraints and that the processes involved in melt extraction and evolution may change in both space and time, but neither the spatial scales nor the timescales of such changes are yet known. For these reasons, only integration of more complete and diverse sets of geological and geochemical observations can lead to better-defined models for melting and source heterogeneity, and recent efforts have indeed taken a first step in this direction (see e.g., Hauri [1997], Hauri and Kurz [1997], Sims et al. [1999], Stracke et al. [1999], White et al. [1993], White and Duncan [1995], Albarède et al. [1997], Spiegelman [1996], Spiegelman and Reynold, [1999], and the combined results of the Hawaii Scientific Drilling project (HSDP)).

[3] In the study presented here, we chose to measure a large variety of elemental concentrations and isotopic ratios on the same suite of samples, which have been collected from a single ridge segment and erupted in a relatively short period of time. We chose the Theistareykir segment in northeast Iceland [Slater, 1996; Slater et al., 1998, 2001] (Figure 1), located just south of the Tjörnes fracture zone and furthest away from the inferred Icelandic plume center (Figure 1). The samples chosen for this study erupted within only 10,000 years after the last glaciation. Restricting our study to lavas erupted in such a limited time interval and limited spatial region should make the interpretation of the geochemical observations easier. The samples have been thoroughly characterized in terms of field relations and occurrence by Grönvold [see O'Nions et al., 1976] and major and trace element compositions by Slater and McKenzie [Slater, 1996; Slater et al., 1998, 2001]. The investigated lavas are quite variable, both elementally and isotopically. Most samples are olivine-tholeiites with MgO contents greater than 7–8 wt%; some of the larger flows have MgO contents between 9 and 13 wt% and should be described as picrites according to the latest International Union of Geological Sciences (IUGS) classification [Le Bas, 2000]. Some small volume flows consist of rocks with up to 22 wt% MgO and contain as much as 20% olivine phenocrysts. These lavas have traditionally been called picrites, a term which is retained here (Tables 1 and 2), though they would classify as komatiites according to the latest IUGS classification [Le Bas, 2000]. In this study, we use major element concentrations reported by Slater [1996] and report additional trace element concentrations of 33 elements and Sr, Nd, Pb, and Hf isotope ratios on 43 samples (Tables 1 and 2). Oxygen isotope values have previously been determined on a subset of the samples by Eiler et al. [2000]. This comprehensive geochemical data set thus allows investigation of a variety of key processes including magma evolution, crustal interaction, partial melting, and the composition and evolution of the Icelandic source mantle from a variety of perspectives.

Figure 1.

Geological map of Iceland simplified after the geological map published by the University of Edinburgh ( Red isopleths show crustal thickness in km [Staples et al., 1997; Darbyshire et al., 1998, 2000a, 2000b]. The region with the largest crustal thickness corresponds to the inferred center of the Iceland plume [Wolfe et al., 1997]. Regional He isotope variations [Kurz et al., 1985; Breddam et al., 2000] are plotted as white numbers.

Table 1. Major and Trace Element Concentrations of the Theistareykir Samplesa
Phenocrystsaphyricpl.20%pl.10%pl.7%aphyricaphyricol ± pl2%ol ± pl1%ol ± pl ± cpx5%
  • a

    Major element compositions and estimates of phenocryst abundance are taken from Slater [1996]. FeO are based on a Fe2+/Fe3+ = 0.9, and Mg# = (Mg2+/(Mg2+ + Fe2+)).

  • All trace element compositions were determined by ICP-MS at the NHMFL and are reported in ppm, except the Pb concentrations (Pb*) which were determined by isotope dilution. Reproducibility for the rare earth elements (REE) is generally better than 2%, whereas reproducibility for Nb, Ta, Zr, Y, Hf, Rb, Sr, Ba and Th is generally better than 3%, and Sc, Ti, V, Cr, Co, Ni and U reproduced to better than 5–6%. The reference values for the BIR standard are taken from Eggins et al. [1997, and references therein].

Phenocrystsaphyricol.4%ol ± cpx ± pl2%ol.2%ol ± pl2%ol ± pl ± cpx7%ol ± p ± cpx2%  
Phenocrystsaphyricol ± pl ± cpx5%ol ± px5%aphyricol ± px5%aphyricaphyricol ± pl ± cpx4% 
Phenocrystsol.1%ol.5%ol.1%ol.15%ol.3%ol.2%ol.5%ol ± pl3% 
Phenocrystspl.10%pl.1%ol.20%ol.12%n = 8ol.20%ol.22%  
SiO249.349.5746.846.03 46.5146.36  
Al2O316.6114.8312.0311.04 11.6411.58  
Fe2O31.141.240.990.98 0.990.98  
FeO9.2510.058.017.96 7.987.91  
Fe2O3total11.1412.419.899.83 9.869.77  
MgO7.428.0718.6721.51 20.1220.1  
CaO12.6712.2311.2310.23 10.6710.65  
Na2O2. 1.171.22  
K2O0.1050.1180.0120.01 0.0110.014  
TiO21. 0.460.46  
MnO0. 0.160.16  
P2O50. 0.030.02  
Mg#59598183 8282  
Sum99.899.799.699.5 99.799.5  
FlowAsbyrgi (NE of Theist.)Draugarhraun (Krafla)    
Sample932293239324936693961 σ (%)BHVO1 σ (%)BIR
Phenocrystsaphyricaphyricaphyricaphyricaphyricn = 8 n = 8(ref.conc.)
Table 2. Sr, Nd, Pb, and Hf Isotope Ratios in the Theistareykir Basaltsa
  • a

    The 87Sr/86Sr ratios are normalized to 87Sr/86Sr = 0.1194, and 143Nd/144Nd ratios are normalized to 146Nd/144Nd = 0.7219 and 176Hf/177Hf ratios are normalized to 179Hf/177Hf = 0.7325. The in-run precision of 87Sr/86Sr and 143Nd/144Nd is generally <±0.000008 (2σm), and <±0.000012 (2σm) for 176Hf/177Hf. The E&A SrCO3 standard (n = 16) averaged 87Sr/86Sr = 0.708009 ± 17 (2σ) and the La Jolla Nd standard (n = 32) averaged 143Nd/144Nd = 0.511848 ± 11 (2σ) during the measurement period. 176Hf/177Hf ratios are reported relative to 176Hf/177Hf = 0.282160 of the JMC-475 Hf standard, which averaged 176Hf/177Hf = 0.282154 ± 28 (2σ, n = 23) during the analyses period. The NBS 981 Pb standard (n = 33) averaged 206Pb/204Pb = 16.9385 ± 38 (2σ) 207Pb/204Pb = 15.4909 ± 45 (2σ), 208Pb/204Pb = 36.6982 ± 137 (2σ) during the measurement period.

9385- 0.70298980.51310570.283243918.05103615.40043237.665375
 - - - 0.2832229- - - 
9389  0.70297970.51308580.283240718.05503515.40592937.683371
94101- 0.70311980.51305370.2832301118.14722715.41712237.790357
94102- 0.70310970.51306560.2832131018.12883615.40662837.762572
 - - 0.51305911- - - - 
9332- 0.70304470.51305960.2832181618.06332215.41002137.696844
 - - - - 18.07012315.41862137.722351
 - - 0.513065110.28318712- - - 
9499- 0.70304670.51307070.283223618.05893115.40932637.696463
 - - 0.51305212- - - - 
 - - 0.51304610- - - - 
 - - 0.5130678- - - - 
 - - - 0.28319011- - - 
-- - - 0.28325318- - - 
 - - 0.51310110- - - - 
 - - 0.51307815- - -   
9375  0.70301680.51309060.2832531418.04573315.40872837.680772
9409- 0.70302880.51307570.2832601118.04554915.39634237.6611108
 - - - - - - - 
9435  0.70299480.51310060.2832691418.02514415.40503937.657597
 - - - 0.2832791318.01812415.39712237.630556
9353- 0.70312380.513033100.2831881018.05421915.40941637.715940
 - 0.70311780.5130347- - - - 
 - - - 0.2832642518.01322415.40201937.649246
 - - 0.51310410- - - - 
 - - 0.5131066- - - - 
9393- 0.70310080.513040±880.2832231218.07251115.41781037.745926
 - - 0.5130380.2831976- - - 
 - - 0.523039±66- - - - 
9394- 0.70302070.51309660.283251918.03654715.41184137.679099
9412- 0.70306380.51307570.28321312- - - 
9416- 0.70305970.51306480.283234718.05742315.40652037.693753
 - - 0.51305513- - - - 
94112- 0.70302470.51309780.2832561018.01352615.40022337.642660
 - - 0.513091160.2832739- - - 
9371- 0.70294270.51310290.283256818.04562615.40322337.665754
 - - 0.5130859- - - - 
 - - - 0.2832368- - - 
 0.7029347- 0.51311114- - - - 
 - - - 0.28323616- - - 
9377- 0.70291570.51310560.283265818.00803015.39962937.621073
9395- 0.70289770.51310970.2832561218.01872715.40292237.622553
9476- 0.70292370.51311480.283260818.01453815.39253137.610777
 - - 0.51311810- - - - 
 - - 0.51308310- - - - 
 - - 0.5130669- - - - 
 - - - - 17.92134615.38204437.4977106
 - 0.7028508- - - - - 
 - 0.70284770.513145120.28327725- - - 
 - - 0.51314611- - - - 
93970.70289970.70285080.51314470.28329413- - - 
 - 0.70284780.51315870.28331815- - - 
TH29- 0.70284970.5131537- 17.90318615.36667237.4512172
Samples adjacent to Theistareykir Asbyrgi (NE of Theistareykir)
 0.70320170.7031828- - 18.34171415.45011238.032132
 0.7031777- - - - - - 
 0.7031888- - - - - - 
93230.70321570.703208 0.51302580.283170618.37402215.44221938.032446
 - -7- 0.2831486- - - 
9324  0.70319470.51302870.283185618.37661215.44541038.040728
Draugarhraun (Krafla)
 0.7031958- - - - - - 

2. Results

2.1. Major Elements

[4] The major element concentrations of the Theistareykir samples are presented in Table 1 and are taken from Slater [1996]. Compositions range from olivine-tholeiites with MgO contents between ∼8 and 16 wt% to picrites with up to ∼22 wt% MgO. Other major element concentrations and ratios are inversely correlated with MgO (or Mg#; see Figure 2) but show an increased scatter at MgO∼8 wt%. CaO/Al2O3, Al2O3/TiO2, and Na2O/TiO2 ratios are positively correlated with MgO (or Mg#; see Figure 2). The samples from Asbyrgi and Draugarhraun, locations adjacent to Theistareykir, fall off these general trends. The Asbyrgi samples have lower SiO2 and CaO and higher TiO2 and FeO compared to samples from Theistareykir with similar MgO contents. The Draugarhraun samples, which belong to the Krafla volcanic system, are highly differentiated quartz-tholeiites with MgO of ∼5.5 wt% and Mg# = 42. (Figure 2).

Figure 2.

Whole rock major element variations in the Theistareykir basalts plotted versus MgO as an index of the relative amount of fractional crystallization. Increasing major element abundances (SiO2, FeO, Na2O, CaO, Al2O3, TiO2) with decreasing MgO indicate control by olivine fractionation and/or accumulation. Colored arrows indicate the effect of 10% of fractional crystallization of olivine (ol), plagioclase (plag) and clinopyroxene (cpx) on a melt with about 14 wt% MgO similar to samples 9370 or 9394. Decreasing CaO/Al2O3 and Al2O3/TiO2 ratios with decreasing MgO, however, indicate that ol fractionation alone cannot explain the major element variations and are explained as a result of partial melting rather than fractional crystallization of ol, cpx and plag (see sections 3.1 and 3.2; Figure 9). Modeling fractional crystallization of the high-MgO basalts using MELTS [Ghiorso, 1994; Ghiorso and Sack, 1995] also shows that fractionation of the high-MgO basalts (black curve) is dominated by crystallization of olivine within the first 10% of crystallization. Continued fractionation of ol-plag-cpx does not lead to melts with MgO < 7–8 wt% even with combined ol-cpx-plag crystallization levels greater than 70%, suggesting that the evolved quartz-tholeiites from Draugarhraun (MgO of ∼5.5 wt%) do not lie on one common liquid line of descent with the high-MgO Theistareykir basalts (>8 wt% MgO). The black curve represents the fractionation path of sample 9394, which according to the criteria of Korenaga and Kelemen [2000] is close to a “primary” melt (see section 3.1 and Figure 10). The crystallization sequence at 1kbar is ol (first 10%) −> ol + plag −> ol + plag-cpx; ticks marks are for 10, 21, 32, 41, 49, 62, and 71% of fractional crystallization. The gray curve shows the fractional crystallization path at 1kbar for the composition given by Ghiorso and Carmichael [1985], which is similar to those of the Asbyrgi basalts (tick marks are for 15, 33, 46, 58, 70, and 83% of crystallization; see Ghiorso and Carmichael [1985] for details of the crystallization sequence). Contrary to the high-MgO Theistareykir basalts, fractional crystallization of such evolved melts leads to compositions that are within the range of typical Icelandic quartz-tholeiites.

[5] Because the Draugarhraun and Asbyrgi basalts clearly have different parental magmas than basalts from Theistareykir (see appendix A), they have been excluded from the discussion in section 3. Furthermore it should be noted that both the spatial [Slater, 1996] and the genetic relationship of the Asbyrgi basalts to Theistareykir is unclear. The compositions of the Draugarhraun and Asbyrgi basalts is shown for reference in the figures, and their relationship to the Theistareykir basalts is discussed in appendix A.

2.2. Trace Elements

[6] The primitive character of the Theistareykir basalts as indicated by their major element compositions (Figure 2, Table 1) is confirmed by high concentrations of compatible trace elements (e.g., Cr up to 2500 ppm and Ni up to 800 ppm; Table 1). Despite this, the Theistareykir basalts display a large range of light-rare earth-element (LREE) and very-incompatible-element (VICE) enrichments (Figure 3, Table 1). They range from strongly LREE and VICE depleted to slightly LREE and VICE enriched [see also Slater, 1996; Slater et al., 1998, 2001]. The variable LREE enrichment is reflected in the large range of (La/Sm)N ratios, which range from 0.33 to 1.27 (where the subscript N denotes chondrite normalized values [Anders and Grevesse, 1989]). The basalts from Draugarhraun are more enriched with (La/Sm)N = 1.49, and the basalts from Asbyrgi have average (La/Sm)N of ∼1.12. The heavy rare earth element (HREE) part of the trace element patterns in the Theistareykir and Draugarhraun basalts is mostly flat as indicated by (Dy/Yb)N ratios that range from 1.0 to 1.1. The basalts from Asbyrgi have slightly sloped HREE patterns and (Dy/Yb)N ratios between 1.22 and 1.29 (Figure 3). The range in REE and VICE enrichment in the Theistareykir basalts covers almost the entire range reported for Icelandic tholeiites [e.g., Hemond et al., 1993; Chauvel and Hemond, 1999]. Only alkali basalts from the southeastern rift zone and intermediate to acidic volcanic rocks from evolved volcanic centers such as Krafla are more enriched than the most enriched Theistareykir basalts.

Figure 3.

Chondrite normalized REE patterns [Anders and Grevesse, 1989] and primitive upper mantle (PUM) normalized extended REE patterns [McDonough and Sun, 1995], arranged by individual flows as mapped by Slater [1996]. The Theistareykir samples display a wide range in VICE and LREE enrichment ranging from LREE and VICE depleted samples to slightly VICE and LREE enriched samples [see, also, Slater, 1996; Slater et al., 1996, 1998]. Note that not all of the Theistareykir samples have been analyzed for Pb concentrations.

[7] Compared to most OIBs and MORBs, Theistareykir and Icelandic basalts in general are unusually depleted (Figure 4, Table 1). In a MORB-normalized plot (Figure 4), the REE (Ce-Lu) abundances are lower, whereas the VICE (Cs-La) have a similar range as primitive MORB, resulting in a steeper multielement pattern of the Theistareykir basalts compared to primitive MORB with similar VICE abundances. Sr and Ba are enriched, Nb and Ta are moderately enriched compared to La, and Zr, Hf, U, and Pb are depleted compared to neighboring REE in the Theistareykir basalts relative to the most primitive MORB. These features appear to be characteristic of Icelandic basalts in general [see Hemond et al., 1993; Chauvel and Hemond, 1999]. High Sr/Nd ratios in Theistareykir basalts are accompanied by only slightly positive Eu anomalies (Eu/Eu* = 2*EuN/(SmN + GdN) = 0.99–1.04). The quartz-tholeiites from Draugarhraun have low Sr/Nd and negative Eu anomalies.

Figure 4.

(a) MORB normalized extended REE patterns and (b) REE patterns of the Theistareykir samples compared to those in primitive MORB (MgO > 8wt%, (La/Sm)N < 0.7) compiled from the LDEO petrological database as well as average N-MORB [Hofmann, 1988; Sun and McDonough, 1989]. Normalizing values are the average concentrations of the compiled primitive MORB and are given in Figure 4a in ppm. Note the distinct REE depletions (with the exception of La) of the Theistareykir samples relative to N-MORB while VICE (Cs-La) are within the range of primitive MORB, resulting in steeper multielement patterns than MORB. Zr and Hf are depleted and Ba, Sr, Nb, and Ta are enriched compared to most primitive MORB. Note also that not all of the Theistareykir samples have been analyzed for Pb concentrations. Figures 4c and 4d) Shown is the field for primitive MORB plus the average picritic melt corrected for 20% ol-accumulation, which is remarkably similar to high-degree melts of the depleted mantle (DM-melt, F ≥ 20%, source composition is the depleted mantle as listed on the at, 2000, as personal communication to S. Jacobsson. Concentrations in ppm are as follows: Cs = 0.0005, Rb = 0.041, Ba = 0.45, Th = 0.0059, U = 0.0022, Nb = 0.112, Ta = 0.0063, La = 0.12, Ce = 0.54, Pb = 0.018, Pr = 0.11, Nd = 0.74, Sr = 12.9, Zr = 6.2, Hf = 0.167, Sm = 0.30, Eu = 0.12, Gd = 0.43, Tb = 0.080, Dy = 0.56, Ho = 0.13, Y = 3.66, Er = 0.38, Lu = 0.061). Two different high-degree melts are shown: 1) a simple single-stage melt (green crosses), which is modeled by non-modal dynamic melting [Zou, 1998] with a porosity of 1%. Calculations assume the following modal mineral compositions: 55% ol, 25% opx, 18% cpx, 2% spinel (sp). Melting modes are ol: opx: cpx: sp = 0.05:0.05:0.45:0.45 [Johnson et al., 1990]. Sources of partition coefficients are: ol and opx: Zindler and Jagoutz [1988]; Kennedy et al. [1993], and Beattie [1993]. Cpx: Hart and Dunn [1993], Johnson [1998], Hauri et al. [1994a], and Halliday et al. [1995]. Cs and Ta are assumed to have the same partitioning behavior as Rb and Nd, respectively. 2) As the partition coefficients used in the calculation of the single stage melt are more appropriate to basaltic than to peridotitic systems [Salters and Longhi, 1999], a polybaric melt (red circles) using partition coefficients and pressure dependent melt reactions that are more appropriate for melting at mid-ocean ridges [Salters, 1996; Salters and Longhi, 1999] is also shown. Polybaric melting produces almost identical results at largely similar melt fractions (about 22 compared to 24%), showing that these calculations are relatively insensitive to the chosen melting parameters. Polybaric melting is modeled by incremental melting with small (0.1%) melt increments and a residual porosity of 0.1%. Melt reactions for the polybaric melting calculations are taken from Salters [1996], partition coefficients are taken from Salters and Longhi [1999] and Salters personal communication (2001) (see Table 4 for the values of the partition coefficients used).

Figure 4.


[8] The macroscopic freshness of the Theistareykir basalts is confirmed by their trace element compositions. U and the alkali elements (Cs, Rb, and K) are sensitive indicators of surficial weathering processes. K/U and Cs/Rb ratios fall within the narrow range of values commonly observed in pristine mantle melts. K/U and 1000*Cs/Rb are between 10,000 and 16,000 and 9.4 and 13, respectively (Table 1), compared to ≈13,000 ± 3000 and 13 ± 3 for MORB and OIB source mantle [Hofmann and White, 1983; Jochum et al., 1983]. Th/U ratios are tightly grouped at 3.2 ± 0.3 (Table 1), and together with the narrow range and pristine nature of K/U and alkali element ratios, show that the trace element compositions of the Theistareykir samples has not been markedly affected by surficial weathering or pre-eruptive assimilation of weathered materials.

[9] Incompatible trace element concentrations and ratios in the Theistareykir basalts (e.g., (La/Sm)N, (Sm/Yb)N, Nb/Zr, and δ(Sm/Nd), δ(Lu/Hf)) generally correlate well with major element concentrations and ratios (e.g., Na2O, K2O, TiO2, FeO, and MgO and CaO/Al2O3, CaO/Na2O, Al2O3/TiO2, Na2O/TiO2, and K2O/TiO2; see Figure 5). Samples from the Asbyrgi flow are discrepant in terms of VICE ratios (e.g., Ba/La, Nb/Th) compared to samples with similar major element characteristics from Theistareykir. (Note that δ(Sm/Nd) and δ(Lu/Hf) values are an estimate of the fractionation of these two ratios during melting or other recent mantle modification; they are calculated as the fractional deviation of the Sm/Nd and Lu/Hf ratios in the basalt from those in the source assuming that the source was derived in a single fractionation event from an undifferentiated reservoir 2 Ga ago [Salters and Hart, 1989; Salters, 1996]).

Figure 5.

Representative correlations between major and trace element concentrations and ratios. FeO, δ(Lu/Hf), and (Sm/Yb)N increase while SiO2/FeO and Al2O3/TiO2 decrease with increasing pressure of melting. With increasing degree of melting, incompatible element abundances and ratios such as La, SiO2, K2O, TiO2 and Nb/Zr, (La/Sm)N, δ(Sm/Nd) decrease while CaO/Al2O3 increases with increasing degree of melting. The Theistareykir basalts are, therefore, qualitatively consistent with trends expected from continuous melting of a homogeneous source with the incompatible element enriched samples representing the low-F, high-P melts and the incompatible element depleted picrites representing the high-F, low-P melts. Symbols as in Figure 2, excluding the Draugarhraun basalts.

2.3. Radiogenic Isotopes

[10] Measured variations in Sr, Nd, Pb, and Hf isotope ratios in the Theistareykir samples (87Sr/86Sr = 0.702847–0.703208, 143Nd/144Nd = 0.513025–0.513144, 206Pb/204Pb = 17.8950–18.3766, 207Pb/204Pb = 15.3664–15.4454, 208Pb/204Pb Pb = 37.4425–38.0407, and 176Hf/177Hf = 0.283170–0.283294; Table 2, Figure 6) reflect much of the range of the isotopic variability known for Icelandic volcanics. Despite the limited temporal and spatial variation of the sample suite, our Sr and Nd data show a relatively large range compared to the total range observed for Icelandic basalts and agree well with literature data [Elliott et al., 1991; Hemond et al., 1993] (Figure 6a). Our Hf isotope data also fall well within the range of Hf isotope data previously reported for Icelandic volcanics [Nowell et al., 1998; Salters and White, 1998; Hanan et al., 2000; Kempton et al., 2000] (Figure 6b). However, compared to results of Elliott et al. [1991] our Pb isotope data show a greater range and have higher 207Pb/204Pb for a given 206Pb/204Pb (Figures 6c and 6d). Our results therefore support the argument made by Hanan and Schilling [1997] and Thirlwall [2000] that the Pb isotope data of Elliott et al. [1991] are inaccurate due to mass fractionation artifacts. Our Pb data are similar to Pb isotope ratios found in MORB from the adjacent Kolbeinsey ridge [Mertz et al., 1991; Mertz and Haase, 1997; Schilling et al., 1999], although the Theistareykir data lie to the low 207Pb/204Pb side of the Kolbeinsey array (Figure 6d).

Figure 6.

Correlations between the radiogenic isotope ratios in the Theistareykir melts compared to the variations observed for Icelandic volcanics in general. (a) 87Sr/86Sr versus 143Nd/144Nd; (b) 143Nd/144Nd versus 176Hf/177Hf, and (c, d) 206Pb/204Pb versus 207Pb/204Pb and 208Pb/204Pb, (e) 206Pb/204Pb versus 143Nd/144Nd, (f) 206Pb/204Pb versus 87Sr/86Sr. The Theistareykir melts show excellent linear correlations in Figures 6a–6d and extend from the most depleted toward intermediate compositions found in Icelandic volcanics. The isotopically most depleted basalts from Theistareykir are similar to the basalts from the adjacent Kolbeinsey ridge. Although simple linear trends are observed for the Theistareykir and Icelandic basalts as a whole in Figures 6a–6d, more complex relationships are observed in Sr-Pb and Nd-Pb isotope plots in Figures 6e and 6f. For a given Sr and Nd isotope composition, the Pb isotope composition of Icelandic volcanics are most radiogenic in volcanics from the southeastern volcanic zone (SEVZ), least radiogenic in volcanics from the northeastern volcanic zone (NEVZ), and intermediate in volcanics from central-Iceland and the western volcanic zone. See Figure 1 for locations. Symbols for the Theistareykir basalts as in Figure 2. Data sources Figures 6a–6d: Theistareykir, literature: Elliott et al. [1991] and Hemond et al. [1993]. Reykjanes Peninsula: Zindler et al. [1979], Elliott et al. [1991], Hemond et al. [1993], Salters and White [1998], Condomines et al. [1983], Sun and Jahn [1975], Dupré and Allègre [1980], Gee et al. [1998a, 1998b]. Northeastern volcanic zone (NEVZ): Hemond et al. [1988, 1993], Condomines et al. [1981, 1983], Nicholson et al. [1991]. Central Iceland: Hemond et al. [1988, 1993], Condomines et al. [1981, 1983], Sun and Jahn [1975]. Katla: Hemond et al. [1988, 1993]; Condomines et al. [1981, 1983]; Sigmarsson et al. [1992b]; Furman et al. [1991], Park [1990]. Torfajoekull: Stecher et al. [1999]. Hekla: Hemond et al. [1988, 1993], Condomines et al. [1981, 1983], Sigmarsson et al. [1992a], Furman et al. [1991], Park [1990]. Veidivoetn: Park [1990]. Vestmann Islands: Hemond et al. [1988, 1993], Condomines et al. [1981], Furman et al. [1991], Park [1990]. Snaefellsness Peninsula: Hemond et al. [1988, 1993], Condomines et al. [1981]. Snaefell: Hards et al. [1995]. Tertiary basalts: Hanan and Schilling [1997], Hardason et al. [1997]. Additional Pb and Hf isotope data are from Chauvel and Hemond [1999] and Kempton et al. [2000], and data for the basalts from the Kolbeinsey ridge are from Schilling et al. [1999], Mertz et al. [1991] and Mertz and Haase [1997]. Data sources Figures 6e and 6f: Iceland: [Park, 1990; Elliott et al., 1991; Hemond et al., 1993; Hards et al., 1995; Hardarson et al., 1997; Gee et al., 1998a; Chauvel and Hemond, 1999; Stecher et al., 1999; Kempton et al., 2000]. Kolbeinsey Ridge: [Mertz et al., 1991; Mertz and Haase, 1997; Schilling et al., 1999]. Reykjanes ridge: [Taylor et al., 1997].

Figure 6.


Figure 6.


[11] Sr, Nd, Pb, and Hf isotope ratios in the Theistareykir and Icelandic volcanics in general extend from compositions typical of north Atlantic MORB toward the center of the OIB array, or, in fact, toward the “enriched” end of the MORB array (Figure 7). Thus the Icelandic trend is distinct from those of most other individual OIB isotopic arrays, which do not include the geochemically depleted compositions found in Iceland, and tend to range from compositions intermediate between MORB and HIMU toward more enriched compositions (e.g., HIMU, EMI, EMII [Zindler and Hart, 1986]; Figure 7).

Figure 7.

Radiogenic isotope variations in oceanic basalts, compared to those in the Theistareykir basalts. (a) 143Nd/144Nd versus 176Hf/177Hf, symbols for the Theistareykir basalts as in Figure 2. (b–g) 143Nd/144Nd and 87Sr/86Sr versus 206Pb/204Pb, symbols as indicated. Figures 7b and 7e show a compilation of ocean island basalts, Figures 7c and 7f show only those OIB that show predominant mixing between MORB and HIMU type sources plus the LOMU (low 206Pb/204Pb) Cook-Austral islands, Figures 7d and 7g show only those ocean island basalts that show significant mixing toward EMI and EMII type compositions. Data sources include only studies that provide Sr, Nd and Pb isotope data on the same samples. These are too numerous to be cited here but are available from the first author upon request. Cook-Australs islands (LOMU): Aitutaki, Atiu, Mauke, Rapa, Marotiri, McDonald. Cook-Austral islands (HIMU): Tubaii, Rurutu, Rarotonga, Mangaia. See text for further discussion.

Figure 7.


Figure 7.


[12] Sr, Nd, Hf, and Pb isotope ratios are highly inter-correlated (Figures 6 and 7, Table 3a). The predominantly linear trends are similar to the relationships observed between radiogenic isotope ratios in basalts from other spatially restricted localities in Iceland [e.g., Park, 1990; Elliott et al., 1991; Hemond et al., 1993; Hards et al., 1995; Hardarson et al., 1997; Gee et al., 1998a]. Linear trends are also observed in Sr-Nd and Pb-Pb isotope ratio plots for Icelandic basalts as a whole (Figures 6a6d). In contrast, the Sr-Pb and Nd-Pb isotope data in Icelandic volcanics [Park, 1990; Elliott et al., 1991; Hemond et al., 1993; Hards et al., 1995; Hardarson et al., 1997; Gee et al., 1998a; Chauvel and Hemond, 1999; Stecher et al., 1999; Kempton et al., 2000] show more complex inter-relationships (Figures 6e and 6f). At approximately constant 87Sr/86Sr and 143Nd/144Nd, Pb isotope ratios are most radiogenic in volcanics from the southeastern volcanic zone (SEVZ), least radiogenic in volcanics from the northeastern volcanic zone (NEVZ), and intermediate in volcanics from central-Iceland and the western volcanic zone (Figures 6e and 6f). He isotope systematics appear to show a similar geographic dependence [Kurz et al., 1985] and to a good first approximation, the highest He isotope ratios appear to be associated with more enriched radiogenic isotope ratios and occur in volcanics erupting close to the inferred center of the plume (see Figure 1).

Table 3. Correlation Coefficient r2 for Correlations Between Isotope Ratios and Major and Trace Element Concentrations and Ratiosa
  • a

    For perfect linear correlations r2 = 1, with the degree of scatter increasing as r2 approaches 0. All r2 are calculated excluding the samples from Draugarhraun as these belong to the Krafla volcanic system (see section 3.1). r2 is given in absolute values, i.e. without a + or − sign indicating a positive or negative correlation, and r2 below 0.4 are given in italics. The asterisk indicates that r2 is calculated excluding the samples from the Asbyrgi region (see section 2.2 and 2.3).

Correlation Coefficients (r2) for Isotope Ratios and Major Elements
Correlation Coefficients (r2) for Isotope Ratios and Trace Elements

[13] Good correlations are observed between radiogenic isotope ratios and major and trace element concentrations and ratios in the Theistareykir basalts (see Tables 3a and 3b and Figure 8). CaO is, however, an exception showing no correlation with the isotope ratios. No correlation with the isotopic composition is observed for Sc and statistically weak correlations (r2 < about 0.45) are observed for Li, Y, Co, and the HREE (Gd-Lu; Table 3b).

Figure 8.

Representative correlations between radiogenic isotope ratios (143Nd/144Nd and 206Pb/204Pb) and major element and trace element concentrations (Na2O, FeO, Ba, Nb) and ratios (CaO/Al2O3, Al2O3/TiO2, (La/Sm)N, (Sm/Yb)N, δ(Lu/Hf), δ(Sm/Nd)). The isotopically most enriched samples (highest 206Pb/204Pb and lowest 143Nd/144Nd) are derived from the highest mean pressure of melting (highest FeO, (Sm/Yb)N, δ(Lu/Hf) and lowest Al2O3/TiO2) and are the lowest degree melts (highest Na2O, (La/Sm)N, δ(Sm/Nd) and lowest CaO/Al2O3). Symbols as in Figure 2, excluding the Draugarhraun basalts.

3. Discussion

3.1. Crustal Processes

[14] The major element systematics of the high-MgO Theistareykir magmas (> 8 wt% MgO) (Figure 2) are affected by crystal fractionation and/or accumulation of olivine (ol). Combined major and trace element relationships (Figures 2 and 9) show that clinopyroxene (cpx) or combined ol-cpx fractionation [Maclennan et al., 2001], in general, does not exert a major control on magma composition. Also, significant amounts of plagioclase (plag) fractionation (> 5%) (Figure 9b) have not occurred in samples with MgO > 8 wt%, although the increasing scatter in variation diagrams (Figure 2) observed for samples with MgO of about 8 wt% may be attributed to either plag accumulation (samples with plag phenocrysts and high Al2O3: sample 9359, 9385, 9389 from Storavitishraun and sample 9383 from Theistareykirhraun) or incipient plag and cpx (along with ol) fractionation (samples with low Al2O3 and CaO contents: sample 9353 and 9393 from Langavitihraun; Figure 2, Table 1).

Figure 9.

(a) CaO/Al2O3 versus V and (b) Al2O3/TiO2 versus Sr. Both V and Sr behave as incompatible elements during melting (Sr somewhat more so), and are compatible in clinopyroxene and plagioclase, respectively. Experimental evidence shows that CaO/Al2O3 increases with increasing degree of melting and decreasing pressure of melting [see, e.g., [Falloon et al., 1988; Kinzler and Grove, 1992b; Hirose and Kushiro, 1993; Kushiro, 1996]. Thus, the negative correlation between CaO/Al2O3 and V and the lack of correlation between CaO/Al2O3 and Sc (not shown, Sc ranges from 35 to 52ppm; Table 1) show that partial melting rather than clinopyroxene crystallization (at any pressure) is responsible for the positive MgO-CaO/Al2O3 relationship (Figure 2d). Although compatible with the major element systematics alone, fractional crystallization of wherlite (orange arrow) [Maclennan et al., 2001], cannot explain the inverse relationship between CaO/Al2O3 and V and can only be explained by relatively large amounts of combined ol and cpx fractionation (> 30% ol), which is inconsistent with the major element systematics (Figure 2). Al2O3/TiO2 decreases with increasing pressure of melting and similar to CaO/Al2O3 versus V, the negative relationship between Al2O3/TiO2 and Sr at Theistareykir suggest that plagioclase fractionation (or combined olivine-plagioclase and olivine-plagioclase-clinopyroxene fractionation) is expected to be of minor importance, and that the Al2O3/TiO2 versus Sr relationship is controlled by melting, not fractionation. Arrows indicate fractional crystallization of olivine (ol), plagioclase (plag) and clinopyroxene (cpx), starting from a lava very similar to sample 9394 (90 ppm Sr, 240 ppm V). Dol from Kennedy et al. [1993], Dcpx from Hart and Dunn [1993], Dplag from White [1999] and references therein. Symbols as in Figure 2, excluding the Draugarhraun basalts.

[15] The size (up to 3 mm) and abundance of ol phenocrysts (up to 20%; see Table 1 and Slater [1996]) in the most MgO-rich lavas (MgO > 18 wt%) as well as Ni contents greater than those in most “primary” magmas (Figure 10a) suggest that at least these magmas have accumulated ol. Petrographic evidence for ol accumulation in the other samples (< 18 wt% MgO) is elusive, and ol phenocryst compositions [Slater, 1996] are similar to calculated equilibrium compositions. The whole rock MgO-FeO systematics also do not clearly support ol accumulation for the basalts with less than 18 wt% MgO (Figure 10b; see also Albarède et al. [1997]). However, the amount of accumulated or fractionated ol ultimately depends on the inferred or estimated composition of the parental or “primary” magma, which is model-dependent [Hart and Davis, 1978; Korenaga and Kelemen, 2000].

Figure 10.

Plots of MgO versus Ni and FeO. (a) Shown are the linear MgO-Ni relationship in the Theistareykir basalts and calculated olivine fractionation curves (adjustment of DNi [Kinzler et al., 1990] after each 0.1% step of olivine crystallization, starting from sample 9390 (MgO about 20 wt%), and sample 9356 (MgO about 14 wt%), respectively). Curves show the range of compositions inferred for “primary” mantle melts according to Korenaga and Kelemen [2000] (labels olNi = 2500–3500 indicate melts with equilibrium olivine with 2500–3500 pmm Ni). Following the model of Hart and Davis [1978] and Kinzler et al. [1990], a linear relationship between MgO and Ni, such as that at Theistareykir (see Figure 10a) is evidence for olivine accumulation and the calculated MgO content of the “primary” Theistareykir lavas is approximately 10–10.5 wt%, indicating that all Theistareykir samples with MgO > 10 wt% would have accumulated olivine. However, for melt compositions with MgO > 9 wt%, olivine fractionation also results in a nearly linear MgO-Ni relationship (Figure 10a). Therefore, the linear MgO-Ni relationship of the Theistareykir basalts cannot be taken as definitive evidence for olivine accumulation (see also Korenaga and Kelemen [2000]), and MgO contents of about 10–10.5 wt% are not a reliable estimate for the primary MgO content of the Theistareykir lavas. Korenaga and Kelemen [2000] suggest that equilibrium olivine in primary melts have Ni contents within the range of Ni contents in olivine from peridotites (∼2500–3500 ppm) and that lower Ni contents in equilibrium olivines are caused by olivine fractionation (Figure 10a). Most Ni contents in equilibrium olivines calculated for the Theistareykir melts are lower than those in peridotites, and would therefore suggest that olivine fractionation, rather than accumulation as predicted by the Hart and Davis model [Hart and Davis, 1978], is the more important process (Figure 10a). Estimating “primary” melt compositions following the approach of Korenaga and Kelemen [2000] requires that olivine is added until the Ni content in equilibrium olivine reaches 2500–3500 ppm (the range of Ni contents in olivines in peridotites). For a Ni concentration in the equilibrium olivine of 3000 ppm, resulting “primary” melt compositions have MgO contents clustering between 13 and 15 wt%, although the total range is between 11.5 and 18 wt%. (b) MgO-FeO systematics in the Theistareykir basalts. Lines labeled Fo94-Fo90 indicate position of melts in equilibrium with olivine with fosterite contents 94–90. Points labeled ol94-ol86 indicate composition of olivine with fosterite contents 94–86. Olivine accumulation trends form linear trends between the composition of the accumulated olivine (e.g., ol90) and the line for melts in equilibrium with that olivine composition (e.g., Fo90; see e.g., Albarède [1997]). Except for the samples with MgO > 18 wt% such trends are not readily apparent and the MgO-FeO relationship in the Theistareykir basalts provides no clear evidence supportive of olivine accumulation. See text for further details. Bold line between Fo88 and Fo92 represents the range of ol-phenocryst composition of the Theistareykir basalts [Slater, 1996; Slater et al., 2001]. Symbols for the Theistareykir basalts as in Figure 2.

[16] Following the model of Hart and Davis [1978] and Kinzler et al. [1990], the calculated MgO content of the “primary” Theistareykir lavas is approximately 10–10.5 wt%. However, because it is difficult to discriminate between ol accumulation or ol crystallization processes based on the MgO-Ni relationship alone, the 10–10.5 wt% estimate is ambiguous (see Figure 10a for details). Following the approach of Korenaga and Kelemen [2000] (see Figure 10 for details) “primary” melt compositions for the Theistareykir lavas have MgO contents clustering between 13 and 15 wt%, and ol fractionation, rather than accumulation as predicted by the Hart and Davis model [Hart and Davis, 1978], appears to be the more important process for lavas with MgO < 15 wt% (Figure 10a). Therefore, owing to the high MgO content of the Theistareykir magmas and the inferred MgO of ∼14 wt% in the primary magmas, the amount of ol accumulation and fractionation is minor even for the most MgO–rich and MgO-poor lavas (< about 20%, see Figure 2).

[17] Positive correlations between CaO/Al2O3 and Al2O3/TiO2 and MgO (Figure 2) clearly show that ol fractionation/accumulation processes alone cannot explain all of the major element variability at Theistareykir, which is confirmed by the large range in trace element concentrations and ratios (Figures 35). In this context, the correlations between major element parameters and radiogenic isotope ratios (Figure 8) are a striking feature that bears on the origin of the major element variations. Especially the correlation between Mg# (MgO) and radiogenic isotopes (see Table 3) poses the question whether the amount of differentiation is somehow linked to source variations, or if inferred source signatures (e.g., Sr, Nd, Hf, Pb isotopes) are affected by crustal processes driven by crystal fractionation (e.g., assimilation fractional crystallization processes (AFC)) as has recently been proposed [Eiler et al., 2000]. Numerous observations confirm that AFC-type processes do not play a defining role for the chemistry and isotopic composition of the Theistareykir melts [e.g., Elliott et al., 1991], see detailed discussion in appendix A2). Is it likely then that magmas originating from different sources (as reflected in different isotopic composition) have a systematically different fractionation history? Or, alternatively, is the major element chemistry not disturbed to such an extent by fractionation processes that it still provides reliable information about partial melting and/or source heterogeneity?

[18] Correlations between parameters such as CaO/Al2O3, Al2O3/TiO2, K2O/TiO2, and also FeO and radiogenic isotope ratios are not influenced by ol fractionation or accumulation, regardless of the exact primary magma composition (in detail, some scatter in the correlations between Al2O3/TiO2 or Na2O/TiO2 and the various isotope ratios could be caused by plag accumulation or incipient plag fractionation; see Figures 4 and 8). Furthermore, correcting the major element compositions back to “primary” melt compositions using the approaches of either Hart and Davis [1978] or Korenaga and Kelemen [2000] shows that the sense and quality of the correlations of “corrected” major element abundances and ratios with isotope ratios is unaffected by these exercises (Table 3) with the exception of MgO (back-correcting all lavas to a certain MgO content eliminates all correlations with MgO). This suggests that the major element systematics are not substantially obscured by high-level fractionation after the magmas leave the mantle and that the correlations between major elements and isotope ratios are a near-primary feature of partial melting and source heterogeneity. Combined major and trace element systematics confirm this notion (e.g., CaO/Al2O3 versus V and Al2O3/TiO2 versus Sr; Figure 9).

3.2. Melting Processes

[19] In the preceding discussion, we have established that the fundamental major and trace element variations observed at Theistareykir are primary or near-primary signatures resulting from partial melting of the mantle - and we know, based on the isotopic variations in the Theistareykir and Icelandic rocks in general, that the mantle source is heterogeneous. In principal, partial melting and source heterogeneity can affect the major and trace element chemistry in two ways: (1) melting of different sources produces melts with distinct isotopic composition and different major and trace element chemistry or (2) the major and trace element chemistry is largely a function of pressure (P) and degree (F) of melting, implying that isotopically different parts of the source are sampled systematically as a function of pressure and degree of melting.

[20] The sub-Icelandic mantle is anomalously hot compared to other portions of the Mid-Atlantic ridge (MAR; ΔT about 200°C, [Sleep, 1990; Schilling, 1991]) and melting beneath Iceland is expected to start deeper than is typical for submerged portions of the MAR. Furthermore, because of the thinner lithosphere, melting beneath Iceland continues to shallower depths than at most other ocean islands. It is reasonable to expect, therefore, that variations in the pressure (P) and extent of melting (F) will have a comparatively important influence on Icelandic melt compositions (it is important to note that, in most melting models, P and F are physically coupled such that low-F melts at a single locality derive from higher average P than high-F melts). What are the basic features that originate from different P and F of melting and are these features compatible with the major and trace element chemistry of the Theistareykir suite?

[21] On the basis of melting experiments [e.g., Falloon et al., 1988; Kinzler and Grove, 1992a, 1992b; Langmuir et al., 1992; Hirose and Kushiro, 1993; Baker and Stolper, 1994; Baker et al., 1995; Kushiro, 1996; Lesher and Baker, 1996; Longhi and Bertka 1996; Longhi and O'Connell, 1996; Kinzler, 1997; Walter, 1998] and thermodynamic calculations with MELTS [Ghiorso, 1994; Ghiorso and Sack, 1995], the effects of pressure and degree of melting are such that high-P, low-F melts have higher TiO2, K2O, FeO, and lower CaO/Al2O3, CaO/Na2O, Al2O3/TiO2 than low-P, high-F melts produced by melting a similar source. High-P, low-F melts also have higher incompatible element concentrations and higher ratios of more to less incompatible trace elements ratios (e.g., higher (La/Sm)N, (Dy/Yb)N, δ(Lu/Hf) ratios [e.g., Salters and Hart, 1989; Salters, 1996]) and lower compatible trace element concentrations than low-P, high-F melts from a similar source. The major (possibly excluding MgO) and incompatible trace element variations in the Theistareykir basalts (Figures 5 and 8) are therefore both qualitatively and quantitatively [Slater et al., 2001] consistent with those expected to be produced by continuous melting of a homogeneous peridotite source with the picrites representing the low-P, high-F melts and the incompatible element enriched tholeiites representing the high-P, low-F melts.

[22] We know, however, that the source is not homogeneous and that source effects on melt composition must also be considered, especially in light of the good correlations observed between radiogenic isotope ratios and major and trace elements in the Theistareykir basalts (Figure 8). Calculations using MELTS show that melts from more fertile peridotites have higher TiO2 and Na2O, and lower MgO, CaO, CaO/Al2O3, CaO/Na2O and Al2O3/TiO2 ratios, than melts of less fertile peridotite at similar F (we used three peridotite compositions, in order of increasing “fertility”: DMM [Wasylenki et al., 1996; Hirschmann et al., 1998, 1999], MM3 [Baker and Stolper, 1994], and HK66 [Hirose and Kushiro, 1993]; see Asimow et al. [1995, 1997] and Hirschmann et al. [1998, 1999] for further details of modeling melting with MELTS). Furthermore, in the context of a single melting column, melt from fertile peridotite will be substantially more enriched in incompatible trace elements than melts produced from the less fertile peridotite higher in the column (at a given local F). Unfortunately, then, the effects of source heterogeneity are very similar to those resulting from variation in the pressure and degree of melting; to a large extent, the effects are additive, and, in general, it is difficult to disentangle the respective roles of source heterogeneity and partial melting on the major and trace element systematics. The depleted character of the Theistareykir melts, however, and the combination of major and trace element and isotopic data allow a more detailed assessment of the respective roles of melting and source composition.

[23] The overall depleted character of Icelandic tholeiites and their similarity to N-MORB in terms of both their isotopic and major and trace element composition suggests a volumetrically dominant role for a depleted, and roughly MORB-like (see discussion in section 3.3) source material. Note, for example, that even in case of melting of a depleted mantle source, very high degrees of melting (> 15–20%, depending on the preferred melting model and partition coefficients, see Figure 4) are required in order to produce the low absolute trace element abundances of the Theistareykir melts. In addition, only experimental melts from the most depleted starting compositions (e.g., Tinaquillo lherzolithe [Falloon et al., 1988]) have similarly low TiO2, Na2O and high CaO, and CaO/Na2O, CaO/Al2O3, similar to the results using MELTS and the Theistareykir major element variations are parallel to melting trends (experimental or using MELTS) from single sources. Because of the apparent dominance of the depleted material, it appears likely therefore that the major elements are substantially more sensitive to the pressure and degree of melting than they are to the presence of enriched source material and thus are best explained in the context of polybaric or dynamic melting (effectively, the extraction of melts that integrate over different pressure intervals within the melting regime [see, also, Wood et al., 1979; Wood, 1979, 1981; Elliott et al., 1991; Slater et al., 2001]. Klein and Langmuir [Klein and Langmuir, 1987, 1989; Langmuir et al., 1992] also concluded that “local” variations of Na8 and Fe8 with ridge depth at several localities along the MAR (including Iceland [Klein and Langmuir, 1987, 1989; Langmuir et al., 1992]) are not caused by source heterogeneity and are instead a function of degree and depth of partial melting.

[24] Slater et al. [2001] have shown that the trace element variations in both lavas and melt inclusions in the Theistareykir basalts are quantitatively consistent with incomplete mixing of instantaneous fractional melts from a homogeneous source. This, however, requires a source that is slightly more enriched than the depleted MORB source, contrary to other evidence for a dominant depleted source (see above and section 3.3). Moreover, trace element concentrations and ratios are more sensitive to the influence of source heterogeneity than to variations in the F and P of melting compared to the major elements. The variation in isotopic composition of Sr, Nd, Pb, and Hf must reflect source heterogeneity, and the correlations between trace elements and radiogenic isotopes in the Theistareykir basalts therefore show that a significant part of the trace element variation must be due to source heterogeneity. Melting and mixing calculations show that only small amounts of enriched source material (< about 20% in case of PUM [McDonough and Sun, 1995]; see also Slater et al. [2001]) is necessary to produce the variations in the Theistareykir basalts. In detail, it is difficult to model the process due to the lack of knowledge of appropriate partition coefficients for heterogeneous sources and the lack of knowledge of the number and composition of enriched components and the mixing process. However, because of the depleted nature of the Theistareykir melts, it appears safe to conclude that melts from the enriched portions of the source contribute only a small fraction relative to melts from the depleted source material.

[25] In contrast to the trace elements, major element variations reflect mainly different P and F of melting. Correlations between major elements and isotopes suggest that different sources are sampled systematically as a function of P and F of melting. This is most easily explained in the case where the enriched material has a lower solidus T. Often, isotopically enriched mantle source material is thought to be produced by intramantle migration of fluids or melts, or subduction of basalt, sediment, or other crustal material. Such material would have a more mafic composition than depleted peridotite and is indeed likely to have a lower solidus temperature, and may also have a narrower melting interval than “less enriched” or “depleted” upper mantle peridotite (i.e., MORB source material [see, e.g., Sleep, 1984; Zindler et al., 1984; Prinzhofer et al., 1989; Hirschmann and Stolper, 1996; Stracke et al., 1999]). Therefore an enriched source with a mafic mineralogy is likely to melt to a larger degree and be consumed in a relatively shorter time and depth interval following the onset of melting than is depleted peridotite. This causes the influence of the isotopically enriched component to be more pronounced in melts which derive from higher pressures. A similar effect, however, would be expected from melting more enriched peridotite. Compared to depleted peridotite, melting of enriched peridotite is also expected to start at higher pressure but is likely to continue melting to similar pressures as depleted peridotite. Therefore, unless partial melting of mafic compositions results in characteristic and identifiable chemical compositions of these melts, the effects of melting mafic and enriched peridotite is expected to be difficult to distinguish, and it is premature to make a definitive conclusion whether the enriched source material consists of enriched peridotite or more mafic compositions.

[26] In summary, melting beneath Theistareykir integrates over a large range of pressures, is characterized by extraction and incomplete aggregation of instantaneous melts with aggregated melts representing low to very high degrees of melting, and is dominated by melting of depleted peridotite (similar to N-MORB source material). An enriched source component is required but appears to be volumetrically minor. Despite the large influence of P and F of melting, trace element concentrations are substantially affected by source heterogeneity. In contrast, because of the dominance of the depleted component, major element variations are little influenced by source heterogeneity and are mainly controlled by dynamic melting. Therefore correlations between radiogenic isotope ratios and major and trace elements indicate that sampling of isotopically distinct components of the Icelandic source is not a random process but occurs systematically as a function of both the pressure and extent of melting. The most enriched melts in terms of their isotopic and trace element compositions are derived from the highest mean pressure of melting, and represent the smallest degrees of partial melting (Figures 5 and 8). This is most readily explained by a lower solidus temperature of the enriched component(s).

[27] This model is also consistent with predicted and observed glacial-interglacial variations in basalt chemistry, suggesting that the observed volumes of melts decrease and the abundance of incompatible elements increase during glacial epochs [Jull and McKenzie, 1996; Gee et al., 1998a; Slater et al., 1998] owing to a decrease of decompression melting in the mantle [Hardarson and Fitton, 1991; Jull and McKenzie, 1996].

[28] Furthermore, at other ocean islands, as at Iceland, an isotopic distinction is commonly observed between tholeiites, alkali basalts and silica-undersaturated magmas (e.g., Hawaii and Samoa). Even at mid-ocean ridges, the observed degree of isotopic variability decreases as the scale of melting, or rate of processing of mantle material (as measured by spreading rate), increases (Figure 11). These observations suggest that the scale on which enriched and depleted materials are intermingled in the mantle is small compared to the maximum dimension over which melts are produced and mixed beneath ridges.

Figure 11.

Isotopic variability in global MORB as a function of spreading rate. The isotopic variability in global MORB increases with decreasing spreading rate, while the average isotopic composition for each ocean basin is independent of the spreading rate. The MORB data include combined Sr, Nd, and Pb isotope data for about 490 samples from the Atlantic (MAR), 325 for Pacific (Pacific ridges) and 130 samples from the Indian ocean spreading centers (SEIR, South East Indian ridge, SWIR, South West Indian ridge). Data have been compiled from the Lamont Doherty petrological database (LDEO petrological database at (a) Nd and Pb isotopic variability versus spreading rate. The isotopic variability is simply the 2σ scaled to the analytical error, which is assumed to be ±0.00002 for 143Nd/144Nd, and 0.015 and 0.035 for 206Pb/204Pb, and 208Pb/204Pb, respectively. (b) Average 143Nd/144Nd, 206Pb/204Pb, and 208Pb/204Pb for each spreading ridge plus 2σ as shown in Figure 11a.

3.3. Isotopic Heterogeneity in the Icelandic Mantle

3.3.1. Depleted Component

[29] Several recent studies of Icelandic volcanism [Thirlwall et al., 1994, 1995; Hards et al., 1995; Kerr, 1995; Kerr et al., 1995; Fitton et al., 1997; Hardarson and Fitton, 1997; Hardarson et al., 1997; Nowell et al., 1998; Chauvel and Hemond, 1999] suggested that the depleted portion of the Icelandic mantle is geochemically distinct from MORB mantle. A non-MORB-type depleted component for Iceland will have important consequences for the traditional picture of simple plume-ridge interaction [Hart et al., 1973; Schilling, 1973]. The apparent dominance of a depleted mantle component on melting beneath Theistareykir (section 3.2) makes this the ideal location to reexamine this suggestion.

[30] Hf and Nd isotope data reported by Nowell et al. [1998] form a trend steeper than the Hf-Nd “mantle array” and appear to point away from the MORB field. In contrast, our Hf and Nd isotope data, as well as those recently reported by Kempton et al. [2000], show that the Theistareykir/Iceland trend is indistinguishable in slope from the Hf-Nd mantle array (Figure 7a). Additionally, the depleted end of the Theistareykir data trends falls within the MORB field in any n-dimensional isotopic construct involving Sr, Pb, Nd, and Hf isotopes (e.g., Figure 7).

[31] The 207Pb/204Pb ratios in Theistareykir basalts reported by Elliott et al. [1991], however, are lower than in even the least radiogenic Atlantic MORB and have been used to argue for a difference between Atlantic MORB and depleted Icelandic basalts [e.g., Hards et al., 1995; Thirlwall, 1994, 1995]. While our Pb isotope data and the re-analyses of the Theistareykir samples of Elliott et al. [1991] by Hanan and Schilling [1997] and Thirlwall [2000] substantially decrease the magnitude of this apparent distinction, the unradiogenic 207Pb/204Pb values in the Theistareykir picrites are indeed slightly less radiogenic than those in Atlantic MORB, including the Kolbeinsey Ridge (Figures 6d and 12). The picrites, however, form the intersection between the Kolbeinsey and the Theistareykir trend in a 207Pb/204Pb versus 208Pb/204Pb diagram (Figure 12), suggesting that the difference between the Kolbeinsey and Theistareykir basalts may be due to mixing with different enriched components (or different proportions of the same enriched components [e.g., Hanan and Schilling, 1997; Hanan et al., 2000]) rather than the lack of a common depleted component [e.g., Hards et al., 1995; Thirlwall, 1994, 1995], which is similar in Pb isotopic composition to the Theistareykir picrites. Therefore, there is no clear isotopic difference between the Theistareykir basalts, and the MORB from the adjacent MAR and the isotopic composition of the Theistareykir basalts are certainly consistent with the suggestion that ambient north Atlantic subridge mantle is involved in Icelandic volcanism.

Figure 12.

Plot of 207Pb/204Pb versus 208Pb/204Pb in the Theistareykir and Kolbeinsey ridge basalts (references as in given in caption of Figure 6). The regression lines for the two suites intersect in the Theistareykir picrites, suggesting that the difference in 207Pb/204Pb between the Theistareykir and Kolbeinsey ridge basalts is due to admixing of different enriched components, rather than the lack of a common depleted component [e.g., Thirlwall et al., 1994; Hards et al., 1995; Thirlwall, 1995] which is similar to the Theisatreykir picrites.

[32] In this context, it is important to note that there is a systematic change in the isotopic character of ambient MORB mantle that, fortuitously or otherwise, coincides with the position of Iceland: the Kolbeinsey Ridge to the north of Iceland is distinct from the Reykjanes Ridge to the south (and most of the rest of the MAR) in terms of Pb isotopes (and, to a lesser extent, Nd and Sr isotopes; Figures 6e and 6f). These differences (Reykjanes Ridge basalts have higher 206Pb/204Pb than the Kolbeinsey Ridge basalts) are clearly reflected in the most depleted Icelandic lavas from Theistareykir and the Reykjanes Peninsula, and argue strongly for the involvement of the local depleted upper mantle in the melting beneath Iceland (Figures 6e and 6f). Hanan et al. [2000] attributed this northward increase of 206Pb/204Pb to an increase of a so-called “e” component (a component similar to EMI [Zindler and Hart, 1986]) in the MORB source of the Reykjanes compared to the Kolbeinsey Ridge. Whether this increase is continuous or stepwise was left open. Alternatively, one could attribute the change in character of the subridge mantle from the Reykjanes Ridge to the Kolbeinsey Ridge (perhaps across the Tjörnes fracture zone) to be an inherent source characteristic which is not related to the increase of a component similar to “e.”

[33] Some trace element abundance characteristics of Icelandic basalts (i.e., apparent enrichments of Sr, Ba, Nb, Ta, and depletion in Zr and Hf relative to similarly incompatible counterparts; see section 2.2), have also been used to argue for a depleted source distinct from the ambient MORB mantle [e.g., Fitton et al., 1997; Chauvel and Hemond, 1999; Kempton et al., 2000]. Rather than apparent enrichments in some incompatible trace elements (see section 2.2), however, it is the REE depletion compared to VICE in the Theistareykir basalts that leads to a difference in some trace element ratios between primitive MORB and the Theistareykir suite. For example, the high Sr/Nd ratios in the Theistareykir basalts compared to MORB are caused by a Nd depletion, rather than a Sr enrichment [e.g., Chauvel and Hemond, 1999], since Sr concentrations in the Theistareykir basalts are within the range of those in MORB but Nd concentrations are clearly lower (Figure 4a). Similarly, high Nb/Zr (Nb/Y) ratios in the Theistareykir basalts appear to be caused primarily by a Zr and Y depletion compared to MORB, and not by a modest Nb (Ta) enrichment [e.g., Fitton et al., 1997; Kempton et al., 2000] (see also discussion of Nb-Zr-Y systematics in appendix 4). High Sm/Hf (Sm/Zr) ratios in the Theistareykir basalts compared to MORB are also caused by a Hf and Zr depletion (Figure 4a). Ce/Pb and Nb/U ratios in Theistareykir and Icelandic basalts in general [Chauvel and Hemond, 1999] deviate from the postulated narrow range in Ce/Pb and Nb/U [Hofmann et al., 1986; Newsom et al., 1986] in MORB and OIB. However, primitive MORB show a similar range and also a systematic change in Ce/Pb and Nb/U as a function of Pb and U concentrations (Figure 13), which suggests that Ce/Pb and Nb/U ratios in Theistareykir basalts outside the narrow range of MORB and OIB [Hofmann et al., 1986; Newsom et al., 1986] may not be an unusual or even unique feature of Icelandic basalts (in detail, however, comparing the U, Th, Pb characteristics of primitive MORB and Theistareykir is impaired by the few available high quality data in primitive MORB). In summary, the most distinguishing feature between the Theisatreykir basalts and primitive MORB is their REE and Zr, and Hf depletion, their steeper VICE (Cs-La) pattern, and their distinct Ba enrichment compared to primitive MORB. However, are these features caused by a difference between the depleted component and ambient MORB mantle, or by admixing of enriched material?

Figure 13.

Plot of (a) Nb/U versus U and (b) Ce/Pb versus Pb in primitive MORB and the Theistareykir basalts (the Draugarhraun and Asbyrgi basalts are omitted). Ce/Pb and Nb/U ratios in primitive MORB (> 8wt% MgO, (La/Sm)N < 0.7; compiled from the Lamont Doherty petrological database at decrease with increasing Pb and U concentrations and show a larger range than commonly assumed [e.g., Hofmann et al., 1986; Newsom et al., 1986] similar to the range observed in the Theistareykir basalts. See text for further discussion.

[34] In Figures 4c and 4d, the green crosses and red circles represent high-degree melts (24 and 22%, respectively) of depleted peridotite. To a first order, these patterns are very similar to the Theistareykir picrites and are relatively independent of the preferred melting model and partition coefficients (see Figure 4 and Table 4). Note that some elements (e.g., Cs, Rb, Ba, Sr) could not be modeled with the polybaric melting model due to the lack of appropriate partition coefficients. The calculated high-F melts, including the Th, U, Nb, Hf, Zr, and Sr concentrations, show fairly good agreement with the picrites, but the high Ba concentrations are difficult to model. We therefore suggest that the differences between the “high-F MORB” and the picrites document the enhanced abundance of enriched material in the sub-Icelandic mantle compared to ambient DM, and not a difference in the depleted component itself, and that the picrites represent an aggregate of melt extracted from the entire melting column, not just the uppermost part. The elements that show the “anomalous” behavior are controlled by the enriched component, even in the picritic melt, while others are more effectively “buffered” by contributions from the depleted material.

Table 4. Partition Coefficients Used for the Trace Element Modelinga
 OlivineopxLow Ca cpxHigh Ca cpxGarnetSpinel

[35] In summary, virtually all available isotopic and chemical evidence shows that the depleted Icelandic source component is similar to the ambient MORB source material in the north Atlantic. Numerous earlier studies have reached similar conclusions [Hart et al., 1973; O'Nions and Pankhurst, 1973; Schilling, 1973; O'Nions et al., 1976; Langmuir et al., 1978; Zindler et al., 1979; Elliott et al., 1991; Hemond et al., 1993; Hanan et al., 2000], and any other conclusion is physically difficult to reconcile with Iceland being the subaerially exposed portion of the actively spreading Mid-Atlantic ridge. Finally, while we argue strongly for an important, even dominant, role for “normal” subridge mantle in the melting beneath Iceland, it is important to remember that this component will be manifest in Icelandic lavas as one which is variable in composition, or, in other words, as two depleted components, owing to the change in character of the subridge mantle from the Kolbeinsey Ridge in the north to the Reykjanes Ridge to the south of Iceland.

[36] Alternative models that do not invoke melting of the ambient depleted mantle beneath Iceland recently suggested melting of a plume consisting of a recycled slab with a harzburgitic component (ancient oceanic lithosphere) admixed with gabbroic and basaltic portions of recycled oceanic crust [Chauvel and Hemond, 1999]. However, characteristically low Th/U and Rb/Sr ratios in gabbros [e.g., Zimmer et al., 1995; Hart et al., 1999] lead to 87Sr/86Sr and 206–208Pb/204Pb ratios that are lower than in any oceanic basalt, especially for the old ages proposed for the recycled material [Chauvel and Hemond, 1999]. Moreover, temperatures required to extensively melt harzburgite are extreme, essentially higher than those ever reached beneath normal ridges (or even Iceland with its elevated mantle temperature). The melting temperature and interval of basaltic and gabbroic portions of the proposed plume source is similar, but compared to lherzolite or, especially, harzburgite, such mafic lithologies have a lower melting temperature and smaller melting interval. Therefore it appears difficult to envision the proposed scenario where harzburgite is being partially melted at all levels but where the basaltic portions melt partially at significantly greater depth than the gabbros, and where partial melts from the basalts escape unmodified to produce the alkali basalts before the gabbros eventually melt extensively at shallow levels to produce the picrites [Chauvel and Hemond, 1999].

3.3.2. Enrichment in the Icelandic Mantle

[37] Owing to the dominance of the depleted component, melts with the largest contribution of enriched material are still effectively diluted by melts from the depleted component. Consequently, the isotopically most enriched Icelandic melts are markedly less enriched than most other ocean islands and most likely do not represent the chemical or isotopic composition of the enriched source material. Also, melts derived from the enriched source material(s) are averaged due to preeruptive mixing, so that evidence for the existence of different types of enriched source materials is difficult to extract from the isotopic composition of the erupted melts. Thus the origin and character of enriched source material is better studied at other ocean islands, where the enriched signals in the erupted melts should more closely reflect those of the source itself.

[38] On the basis of the Pb isotope systematics of Icelandic basalts and basalts from the adjacent Mid Atlantic ridge, Hanan and Schilling [1997] and Hanan et al. [2000] (hereafter referred to as Hanan et al.) suggested that two isotopically enriched components are involved in the melting beneath Iceland. While we agree with the conclusion of Hanan et al., we note that the interpretation of the Pb isotope data is subject to the grouping of subpopulations (in detail grouping the Eastern Tertiary basalts into a 2.7–7 m.y. and a 7.2–13.9 m.y. instead of a 2.7–7.5 m.y. and a 7.5–13.9 m.y. groups eliminates differences in the slopes of the regression lines between the two groups). The subtle differences in the Pb isotopes between different basalts suites from Iceland (e.g., lower 208Pb/204Pb for a given 206Pb/204Pb in young Eastern Tertiary basalts compared to older Eastern Tertiary basalts) and Icelandic basalts and the adjacent Mid-Atlantic ridge are difficult to interpret in light of the analytical error of the Pb isotope data. Therefore we consider evidence from combined Sr-Pb and Nd-Pb isotope systematics (Figures 6e and 6f) for more than one enriched component (or a single, variably enriched component in addition to the previously discussed variability at the depleted end) to be even more convincing than evidence from the Pb isotope data alone.

[39] In order to evaluate the character of the enriched components in the Icelandic source, we need to consider possible models for the global isotopic variability of MORB and OIB (Figure 7). Is the isotopic variability in MORB and OIB due to specific isotopically different materials with different origins for each plume or locality? Or, are there a certain number of components in the mantle which reflect reproducible processes resulting in a restricted range in composition for each component due to variations in the process itself and its timing? The systematic trends in the global isotopic systematics of MORB and OIB have long been taken to suggest the latter [e.g., White, 1985; Zindler and Hart, 1986], and based on a comparison of isotope data for ocean islands (Figure 7), we suggest that MORB and OIB may be placed into two broad groups: (1) basalts from islands (i.e., island groups or chains, individual islands, or even individual volcanoes) which show a predominance of mixing between “normal” MORB mantle (DMM) and a HIMU-type component (Figures 7c and 7f), which may have its origins in the recycling of oceanic crust [e.g., Hofmann and White, 1982; Chauvel et al., 1992; Hauri and Hart, 1993; Rehkämper and Hofmann, 1997; Salters and White, 1998], and (2) basalts from islands which appear to mix away from the DMM-HIMU array toward isotopically “enriched” compositions like EM I and EM II (Figures 7d and 7g), which have been associated with the recycling of continental crust and sedimentary components into the mantle [e.g., Zindler and Hart, 1986; Wright and White, 1987; Hart, 1988; Woodhead and McCulloch, 1989; Barling and Goldstein, 1990; Le Roex et al., 1990; Weaver, 1991; Chauvel et al., 1992; Hauri and Hart, 1993; Weis et al., 1993; Woodhead and Devey, 1993; Hemond et al., 1994; Roy-Barman and Allegre, 1995; White and Duncan, 1995; Hauri et al., 1996; Hofmann, 1997; Rehkämper and Hofmann, 1997; Blichert-Toft et al., 1999; Gasperini et al., 2000]. Although various authors have proposed that islands and groups, particularly of the second type, show a convergence of mixing trends toward a point or restricted region along the DMM-HIMU mixing array (variously called PREMA [Wörner et al., 1986; Zindler and Hart, 1986], FOZO [Hart et al., 1992; Hauri et al., 1994b], C [Hanan and Graham, 1996], PHEM [Farley et al., 1992], “p” [Hanan and Schilling, 1997; Hanan et al., 2000]), we find little support for this contention in the data compilation shown in Figures 7b7g. Rather, it appears that DMM–HIMU mixing is virtually a ubiquitous feature of both MORB and OIB mantle and that further isotopic variations imposed by the addition of enriched materials (EM-type) to these mixtures occurs in many, but not all, plumes. We note that subtle deviations from a straight mixing line between MORB and HIMU in, for example, a 206Pb/204Pb versus 208Pb/204Pb diagram (or Pb-Pb isotope plots in general) toward higher 208Pb/204Pb may be caused by admixing of a third, EM-like component.

[40] Iceland, along with other near-ridge ocean islands like Galapagos, belongs to the first group suggesting that the enriched material might be mainly HIMU-like but only appears in diluted form in the erupted melts due to mixing with ambient depleted mantle or melts thereof. “HIMU-like” is defined as an enriched component similar to the enriched components in the source of HIMU islands like the Cook-Australs or St. Helena; that is, the HIMU-like component itself must not necessarily be identical to the isotopically most enriched basalts from those islands but can have a range of compositions extending to compositions that are more or less extreme than those observed at St. Helena or the Cook-Austral Islands. Others, most recently Hanan et al. have suggested that the dominant enriched component in the Icelandic mantle is similar in composition to the enriched end of the Iceland array and to a component intermediate between DMM and HIMU (the “p” component of Hanan et al. which is similar to “C” etc.). However, the melting characteristics beneath Theistareykir suggest preferential melting of the enriched components and mixing with the depleted component or melts thereof during the melting and melt extraction process. Pure melts from the enriched components are unlikely to erupt, suggesting that the enriched component is isotopically more extreme than the most enriched Icelandic lavas. Hanan et al.'s suggestion of the composition of “p” was mostly based on the observations that lavas erupted during the phase of maximum production (7–8 m.y.) are most similar to “p” and that trends of the Tertiary and Neovolcanic Icelandic basalts and the Reykjanes ridge and Reykjanes Peninsula basalts appear to converge in a common point in a 206Pb/204Pb versus 208Pb/204Pb diagram. This common point lies close to the inferred composition of “C” and the most radiogenic Icelandic lavas. However, Figures 7b and 7e shows that the existence of “C” has to be regarded as being contentious. Moreover, considering a larger amount of Pb isotope data on Icelandic volcanics and Reykjanes ridge basalts, the apparent convergence of the Pb isotope data observed by Hanan et al. is not obvious (Figure 14). Therefore whether the dominant enriched component is HIMU-like or closer to the enriched end of the Iceland isotopic arrays is difficult to evaluate. For example, using a least squaress approach similar to Hanan et al. to model the isotopic variability in the Icelandic source by three components but replacing the “p” component of Hanan et al. by a HIMU component can equally well account for the isotopic variations in Icelandic lavas. Pb isotope systematics (see Hanan et al.) as well as the displacement of the Iceland array toward higher 87Sr/86Sr and lower 143Nd/144Nd for a given 206Pb/204Pb (Figures 7b and 7e), suggest the involvement of a third EMI-like component, which Hanan et al. suggested to represent the ubiquitous presence of recycled lithosphere or lower continental crust in the North Atlantic mantle. We note, however, that some of the trace element characteristics of crustal material (e.g., Th, U, Pb enrichment, Nb, Ta depletion) do not appear to characterize the enriched Icelandic lavas.

Figure 14.

Plot of 206Pb/204Pb versus 208Pb/204Pb in the Icelandic Neovolcanic basalts including Theistareykir, Icelandic Tertiary basalts and basalts from the Reykjanes Ridge (for references see Figure 6). Based on a smaller data-set as shown here, Hanan et al. [Hanan and Schilling, 1997; Hanan et al., 2000] suggested that these suites converge into a common component with 206Pb/204Pb of about 19.5 (see Figure 4 [Hanan and Schilling, 1997] and Figure 2 [Hanan et al., 2000]) representing the composition of the Iceland plume, which is not obvious from the data compilation and the regressions of the data used here. The gray line is a regression line through the Tertiary Icelandic basalts, the stippled line is a regression line through the Reykjanes Ridge basalts and the black line is a regression line through the Neovolcanic Icelandic basalts. For references see caption of Figure 6.

[41] Explaining the high 3He/4He ratios in Icelandic lavas by an enriched component similar to HIMU which originates from recycling oceanic crust, however, is problematic, since recycled material is expected to have low 3He/4He ratios [Kurz et al., 1982] (note, however, that the Theistareykir lavas have MORB-like, low 3He/4He ratios [Kurz et al., 1985; Breddam et al., 2000]). Reconciling the apparent association of high 3He/4He ratios and enriched isotope ratios with the idea that the enrichment is somehow related to recycling, at Iceland or any other OIB locality, appears to require that He, for whatever reason, behaves more compatibly during melting and recycling than commonly thought. A better understanding of the relationship between rare gas and radiogenic isotope systematics is therefore a central problem in understanding rare gas systematics in MORB and OIB.

[42] In summary, the enriched material in the Icelandic source is of relatively minor abundance, most likely characterized by a lower solidus temperature than the ambient MORB mantle, and is composed of two different types of enriched material. Both of the enriched source components are observed only in diluted form in the erupted melts due to preeruptive mixing with ambient MORB mantle or melts thereof. Therefore it appears likely that even the most enriched Icelandic lavas do not represent the composition of the enriched source material and evaluation of both the character and origin of the enriched component is not a straightforward task. Comparison of the isotopic arrays of Icelandic basalts with those of global OIB suggest that the dominant enriched component may also have a HIMU affinity rather than representing a component similar to the enriched end of the Iceland isotopic arrays and suggests the additional involvement of a minor enriched component similar to EMI type OIB sources (see also Hanan et al.).

3.3.3. Plume-Ridge Interaction

[43] In both the model of Hanan et al. and the model presented here, the depleted upper mantle is the dominant component in the melting beneath Iceland. In the model presented here, the enriched components are thought to be of small abundance relative to the depleted component. In the Hanan et al. model, the enriched component is carried by a geochemically enriched mantle plume, which is the classic picture of plume-ridge interaction [Hart et al., 1973; Schilling, 1973; Sun et al., 1975]. Relating the geochemical enrichment to an enriched plume but at the same time proposing that the depleted component is dominant appears problematic when considering the scale of the geophysical anomaly at Iceland. As there is no a priori reason why the geophysical anomaly at Iceland (or any other hot spot) needs to be related to a geochemically distinct plume, Iceland could also be perceived as a geophysical anomaly or a thermal plume only, and such a scenario has long been suggested as a possible model for Iceland [Zindler et al., 1984].

4. Conclusions

[44] New geochemical and isotopic data confirm that the Theistareykir basalts show no evidence of significant interaction with the Icelandic crust and that some of these basalts represent a rare example of a suite of near-primary magmas. Variations in isotope ratios require source heterogeneity. Correlations between isotope ratios and major and trace element abundances and ratios indicate that melting samples an isotopically and compositionally heterogeneous Icelandic mantle. This behavior is most readily explained if the enriched portions of the Icelandic source have a lower solidus temperature than the associated depleted material.

[45] Melting of geochemically and isotopically depleted source mantle, similar to that at mid-ocean ridges, dominates the compositions of the Theistareykir melts, indicating that the isotopically enriched materials in the Icelandic mantle are relatively minor in terms of mass or volume. Evidence from Iceland, MORBs, and other ocean islands suggests that the mantle in general is best described as consisting of variably enriched HIMU-type fragments in a depleted matrix. However, the enriched components in the Icelandic source mantle are inferred to be composed of two different types of enriched material. Comparison of the isotopic arrays of Icelandic basalts with those of global OIB indicates that the dominant enriched component may have a HIMU affinity, rather than representing a component similar to the enriched end of the Iceland isotopic arrays, and suggests the additional involvement of a minor enriched component similar to EMI type OIB sources.

Appendix A

A1. Fractionation Processes in High MgO Suites Compared to More Evolved Icelandic Rocks Suites

[46] The high-MgO suites like Theistareykir are characterized by fractionation/accumulation of olivine only. The major and trace element compositions of more evolved (and more abundant) quartz-tholeiites like the samples from Draugarhraun (which are characterized by low CaO and Al2O3, negative Sr and Eu anomalies; Table 1, Figures 2 and 3) indicate that combined olivine-plagioclase-clinopyroxene (ol-plag-cpx) fractionation must be important [Nicholson et al., 1991; Maclennan et al., 2001] as is typical for MORB in general [Albarède, 1992]. The Draugarhraun basalts and other Icelandic quartz-tholeiites, however, do not lie on a common liquid line of descent with the high-MgO Theistareykir basalts (> 8 wt% MgO; Figure 2). Rather, parental magmas for the quartz-tholeiites have to be fairly evolved themselves, similar to the parental melt composition of evolved Icelandic lavas suggested by Ghiorso and Carmichael [1985], which in turn is also difficult to produce by fractionation from the high MgO lavas. This can be shown on the example of the Asbyrgi basalts which have a very similar major element composition as the parental magma of Ghiorso and Carmichael [1985]. Although the samples from Asbyrgi appear to lie on a common liquid line of descent with the high-MgO Theistareykir lavas (MgO ∼ 14 wt%; Figure 2), a closer look reveals that different amounts of crystallization of the various phases are required for each oxide, and that they are, in fact, difficult to model by fractional crystallization (ol-plag-cpx) from the high-MgO Theistareykir magmas. This is confirmed by the distinct trace element composition of the Asbyrgi basalts, which cannot be explained by simple ol-plag-cpx fractionation of the high-MgO Theistareykir magmas (Figure 3). Therefore the high-MgO suites and the quartz-tholeiites do not share a common parent magma, and do not follow a common fractionation path (Figure 2); nor is there a continuous evolution from high-MgO to evolved quartz-tholeiites and more felsic rocks [e.g., Eiler et al., 2000]. Consequently, primitive rocks suites such as Theistareykir and evolved rocks suites like, for example, Krafla, have to be interpreted in a different evolutionary context.

A2. Assessing the Importance of AFC

[47] Eiler et al. [2000] (hereafter simply “Eiler et al.”) have recently advocated, based on oxygen isotope analysis of a subset of the samples studied here, that all of the geochemical and isotopic variations at Theistareykir can result from large amounts of assimilation of altered crust by a single parent magma in the context of combined assimilation-fractional crystallization (AFC). We strongly disagree with their conclusion, for many of the reasons that have been outlined by others [e.g., Muehlenbachs et al., 1974; Elliott et al., 1991; Gautason and Muehlenbachs, 1998] and are given below.

[48] The reason why AFC has always received a considerable amount of attention in the evaluation of the chemical evolution of Icelandic magmas is because of the relatively large thickness of altered crust that must be traversed by each erupted lava, and the fact that Icelandic lavas have long been known to display a range in δ18O from about 6 to less than 0 [e.g., Muehlenbachs et al., 1974; Hattori and Muehlenbachs, 1982; Condomines et al., 1983; Hemond et al., 1988; Nicholson et al., 1991; Sigmarsson et al., 1992a, 1992b; Hemond et al., 1993; Gautason and Muehlenbachs, 1998; Gee et al., 1998b; Eiler et al., 2000]. Basalts comprising the more “primitive” end of the Icelandic spectrum, however, like those from Theistareykir, have generally been thought to preserve their mantle-derived geochemical characteristics, due to their petrologic character, the variation in radiogenic isotope ratios (known, ultimately, to originate in the mantle), and to the fact that their δ18O values closely approach those of MORBs [Hemond et al., 1993; Gautason and Muehlenbachs, 1998]. However, even although some studies have invoked AFC to explain the low δ18O in evolved Icelandic rock suites [e.g., Nicholson et al., 1991; Sigmarsson et al., 1991a], others have argued that the low δ18O values in the assimilant (−8 to −10‰) required by these models are too low to be representative of any significant portion of the Icelandic crust [Jonasson, 1994; Muehlenbachs, 1998]. Furthermore, there is mounting evidence that the dacites and rhyolites, the most evolved members of these suites with the lowest δ18O values, are products of crustal anatexis [Oskarsson et al., 1982, 1985; Thy et al., 1990; Sigmarsson et al., 1991b, 1992a; Jonasson, 1994; Gautason and Muehlenbachs, 1998]. Thus to argue for AFC control at Theistareykir, drawing on comparisons to the more evolved suites where AFC remains contentious, and furthermore, considering that those evolved rock suites cannot be related to the Theistareykir basalts by factional crystallization (see appendix A), appears to be conjectural.

[49] Large amounts of combined plagioclase (plag) and clinopyroxene (cpx) fractionation (up to 60%) form the backbone of the Eiler et al. AFC model, as it is only by such large amounts of crystal fractionation that the inferred amounts of assimilation can be accomplished (assuming ratios of fractional crystallization to assimilation of 4:1 [Taylor, 1980; Taylor and Sheppard, 1986; Eiler et al., 2000]). As discussed in section 3, large amounts of cpx and plag fractionation are not supported by the major element systematics nor the phenocryst populations of the Theistareykir basalts. It is particularly important to note in this context that, unlike other locales where AFC has been most successfully applied [e.g., Taylor, 1980; Taylor and Sheppard, 1986], there is no physical or textural evidence in the Theistareykir volcanics for assimilation, contrary to expectations for a crystallizing magma with an assimilant of broadly similar composition (and, therefore, melting temperature).

[50] In order to quantitatively evaluate the case for AFC at Theistareykir, we have chosen a quartz-tholeiite with an “average” Icelandic isotopic composition (87Sr/86Sr = 0.7033, 143Nd/144Nd = 0.5130, 206Pb/204Pb = 18.5) as an assimilant (NAL71, [Hemond et al., 1993; Chauvel and Hemond, 1999]). Hemond [1993] and Gautason and Muehlenbachs [1998] have argued that quartz-tholeiite is a reasonable approximation of the average Icelandic crust, consistent with the distribution of Icelandic rock types as given, for example by Jakobsson [1972]. A ubiquitous feature of evolved Icelandic quartz-tholeiites is their low Sr/Nd [Hemond et al., 1993; Chauvel and Hemond, 1999], resulting in negative Sr anomalies on multielement variation diagrams (e.g., Figures 3 and 4). AFC calculations for the Theistareykir basalts show that even small amounts of assimilation, with crystallization of a plag-free assemblage, will result in negative Sr anomalies in the derivative magma, in contrast to the positive Sr anomalies observed at Theistareykir (Figures 3 and Figure A1). HREE concentrations increase uniformly during both fractional crystallization and assimilation. The observed range of HREE concentrations at Theistareykir, therefore, place an upper limit on the amount of AFC at an F of ∼ 0.8 (where F is the amount of magma remaining during AFC), in good agreement with the value inferred from major elements in section 3. However, LREE and VICE enrichments in the more evolved “normal” tholeiites require an F of ∼0.6, and the most evolved rocks from Asbyrgi and Draugarhraun require F to be even lower (see Figure A1a). The problem for the isotopes is even worse. In order to reproduce the Sr and Nd isotopic variability with F ≤ 0.8, the assimilant must have very high Nd and Sr concentrations (Nd ≥ 100 ppm and Sr ≥ 2000 ppm) and/or Sr and Nd isotope ratios (87Sr/86Sr ≥ 0.704 and 143Nd/144Nd ≤ 0.5129), well outside the range of those found on a significant geographic scale in even the most extreme or altered Icelandic volcanics (Figures A1c and A1d).

Figure A1.

Shown are the results of modeling the effects of assimilation fractional crystallization processes (AFC; following DePaolo [1981]) on the trace element and isotopic composition of the Theistareykir basalts. (a) Trace element variations in response to combined assimilation of a typical Icelandic quartz-tholeiite (NAL 71 [Hemond et al., 1993; Chauvel and Hemond, 1999]) and fractional crystallization of olivine (the parent magma is sample 9390, the ratio of crystallization to assimilation is 4:1; Dol are from Kennedy et al. [1993]). Heavy rare earth element (HREE) concentrations constrain the maximum amount of AFC to F ≤ 0.8 (F = fraction of magma remaining). Within this range, AFC processes fail to reproduce the range in LREE and VICE in the Theistareykir basalts and impose a negative Sr anomaly, contrary to the pronounced positive Sr anomaly observed in the Theistareykir basalts. (b) Assimilation of Icelandic andesite (H4T [Hemond et al., 1993]) combined with fractional crystallization of cpx, plag and ol (0.475cpx:0.475plg:0.05ol) as suggested by Eiler et al. [2000]. The low Sr/Nd in Icelandic andesites combined with plag fractionation lead to strong negative Sr anomalies even at F < 0.9, while at the same time, for any given F, the LREE and VICE are substantially too enriched for any given HREE content compared to the Theistareykir samples. Also shown is the composition of the “andesitic” assimilant chosen by Eiler et al., which compared to H4T, has about two times lower La and Nd concentrations. The ratio of crystallization to assimilation is 4:1; partition coefficients as given in Figure 9. (c, d) Variations in the Sr, Nd, and Pb isotope composition of the Theistareykir basalts as a result of AFC (the isotopic composition of the assimilant is assumed to be similar to those of the average Icelandic crust: 87Sr/86Sr = 0.7033, 143Nd/144Nd = 0.5130, 206Pb/204Pb = 18.5; other parameters as above). AFC cannot reproduce the Sr, Nd, and Pb isotope variation in the Theistareykir basalts even if F is run to zero. Note the discrepancy between F ≤ 0.8 inferred from the HREE (Figure 11a) and inferred amounts of F based on the isotopic variation. The combined results of AFC modeling indicate that the variation in trace element and isotopic composition of the Theistareykir basalts cannot be caused by AFC processes. Symbols for the Theistareykir basalts as in Figure 2.

Figure A1.


[51] Given the apparent failure of the AFC model at Theistareykir using quartz-tholeiite as the assimilant representative of average Icelandic crust, might it be that some other material present in the crust below Theistareykir could account for the observed chemical characteristics and variations? More extreme assimilants might include rhyolite or alkali basalt, although these rock types are by far less abundant than quartz-tholeiite in the Icelandic crust [Jakobsson, 1972]. Typical Icelandic rhyolites have lower Sr/Nd and higher absolute trace element concentrations than quartz-tholeiites [Hemond et al., 1993; Chauvel and Hemond, 1999] and will therefore impart negative Sr anomalies to the evolving magma even more readily than the quartz-tholeiite. Icelandic alkali basalts generally have sloping HREE patterns and are slightly more enriched in LREEs and VICEs than the quartz-tholeiites [Hemond et al., 1993; Chauvel and Hemond, 1999], so that like the quartz-tholeiites, the range of LREE and VICE enrichments in the derivative melts will be too large for the observed range of HREE abundances. The case gets even worse with an andesitic assimilant and a crystallizing assemblage of cpx and plag as inferred by Eiler et al. (Figure A1b). As a result of both the low Sr/Nd in Icelandic andesites and the crystallization of plag, strong negative Sr anomalies result even at F of 0.9, and, again, crossing multielement patterns between AFC trends and Theistareykir samples require grossly different F for HREEs, and LREEs and VICEs (Figure A1b).

[52] Also, previous studies provide no support for andesite being representative of the crust at Theistareykir, or greater Iceland [Jakobsson, 1972; Hemond et al., 1993]. We further note that compared to actual Icelandic andesites as used in the calculation in Figure A1b (e.g., H4T and SNS17 from Hemond [1993]), the “andesite” used by Eiler et al. (no reference is given to an actual sample) has unusually low Na2O, K2O, La, and Nd (Na2O = 3.0 wt% and K2O = 0.75 wt% versus Na2O ≈ 4.4 wt% and K2O = 1.4–2.6 wt%; La and Nd are about a factor of 2 too low; Figure A1b), and is substantially overenriched in FeO (FeO = 12 wt% versus FeO = 7.3–8.3 wt%).

[53] Therefore, although it is impossible to categorically exclude a role for AFC in modifying the compositions of some of the Theistareykir lavas in a minor way, it clearly does not play a defining role. Both (radiogenic) isotopic variations and the fundamental chemical characteristics of the Theistareykir magmas must originate in the mantle and be sampled during the melting process.

A3. Oxygen Isotope Variation at Theistareykir

[54] Taken at face value, the Theistareykir basalts show variation in δ18O from about 4.2 to 4.7‰ (olivine (ol) phenocryst values equivalent to magmatic values of ∼4.7 to 5.2‰ [Anderson, 1971; Kyser et al., 1981; Muehlenbachs and Byerly, 1982; Eiler et al., 2000]). Although the variation is small (2σ for the suite at ∼0.25‰ is less than twice the estimated reproducibility of a single homogeneous sample of ∼0.14‰ 2σ), the lowest values occur in the most evolved samples and the highest values occur in the picrites. However, the two lowest δ18O values were measured in plag phenocrysts and comparison with the ol values involves an additional 0.2‰ uncertainty [Eiler et al., 2000], so that these two samples are effectively indistinguishable from the bulk of the samples with δ18O from ∼4.3 to 4.6‰. Moreover, several of the samples show considerable variation in δ18O. In particular, repeated analyses (n = 5; Eiler, personal communication, 2000) of sample 9435 from Borgarhraun showed a range of ±0.54‰, and the entire range of the individual analyses from Borgarhraun is 0.7‰, and if interpreted as intraflow heterogeneity (as opposed to analytical difficulty), documents more variation in a single flow than in the entire suite.

[55] Figure A2 shows a comparison between 87Sr/86Sr versus 143Nd/144Nd (panel b) and δ18O versus 143Nd/144Nd (panel a) for Theistareykir (note that the correlation of δ18O with 143Nd/144Nd is among the very best, when correlations of δ18O with all other geochemical parameters are considered). In this figure, all data values have been normalized to the mean and effective uncertainty for each of the analytical methods (using the ±0.14‰ for δ18O, taking into account that this might be an underestimate for many of the samples; see caption for details). Note that when the data are plotted this way, the analytical uncertainties have the same magnitude for each parameter; that is, the error bar shown in the lower part of the diagram is square, and applies to the upper panel as well as the lower. The predictive quality of the data in panel a, however, specifically the δ18O, is poor, where the “predictive quality” is essentially proportional to the slope of the data trend in this representation; the steeper the slope, the better the “predictive quality”.

Figure A2.

Comparison of the co-variations of δ18O and 143Nd/144Nd (a) and 87Sr/86Sr and 143Nd/144Nd (b). The data have been normalized by subtracting the mean of the suite of analyses, and dividing by an estimate of the analytical uncertainty for each of the methods (0.00003 for 87Sr/86Sr, 0.00001 for 143Nd/144Nd, 0.14‰ for δ18O; the value used for δ18O does not include to additional known sources of uncertainty, such as the intrasample or intraflow heterogeneity and the plag-ol fractionation factor, and therefore represents a best-case scenario for the oxygen). With the data thus normalized, direct comparison between the two arrays can be made without consideration of disparities between the uncertainties of the three parameters; that is, the error bars for each of the axes are identical. The slope of the two arrays is essentially a measure of their relative significance. The Nd and Sr data strongly suggest mixing between two components with substantially different Sr and Nd isotope ratios. While the oxygen data are consistent with such an interpretation, the resolution of the oxygen measurements, in the context of the total range of observed values, is not sufficient to construe the δ18O as strongly supporting this or any other model.

[56] Given the intrinsic limitations of the small spread in values, δ18O does “correlate” with other geochemical and isotopic parameters at Theistareykir, and invites itself to be included in the analysis of the rest of the data. Without discussing the details of such correlations, then, it would seem that either the Theistareykir mantle source is heterogeneous in δ18O [e.g., Muehlenbachs et al., 1974; Gautason and Muehlenbachs, 1998], and characterized by systematic differences in δ18O between the enriched and depleted portions, or that the more evolved magmas are preferentially inclined to pick up some crustal oxygen, via interaction with wall rock or fluid, en route to the surface [Muehlenbachs et al., 1974; Gautason and Muehlenbachs, 1998].

[57] Thus, while the oxygen data are consistent with such interpretations that rely on systematic variation with (some) chemical and isotopic parameters in the Theistareykir suite, the systematics are just barely above the noise level, even in the best cases, and while the data are permissive of these interpretations, they are not exactly required. Said in another way, the variations in δ18O are not sufficiently well resolved to permit the development of unique or meaningful models describing the origin of that variation.

A4. Nb-Y-Zr Systematics in Icelandic Basalts

[58] On the basis of Nb-Zr-Y systematics in Icelandic basalts and MORB, Fitton et al. [1997] argued that there is a clear difference between MORB and Icelandic basalts. On the basis of this apparent distinction, ΔNb, which is a function of the Nb/Y and Zr/Y ratios, was defined so that positive ΔNb values indicate compositions similar to Icelandic basalts and negative ΔNb values compositions similar to MORB. Assuming that ΔNb is insensitive to variable extents of melting, Fitton et al. [1997] further suggested that ΔNb can be used as a direct tracer of mantle source composition in much the same way as radiogenic isotope ratios. Considering a more comprehensive set of Nb-Zr-Y data for MORB, the apparent difference in ΔNb between MORB and Icelandic basalts suggested by Fitton et al. [1997] becomes less clear, casting some doubt on the usefulness of ΔNb as a parameter to discriminate between MORB and Icelandic basalts (Figure A3a). Furthermore, ΔNb proves to be affected significantly by variable extents and pressure of melting and therefore cannot be considered to be a process-independent indicator of source composition (Figure A3c).

Figure A3.

(a) Nb-Zr-Y systematics in Icelandic and mid ocean ridge basalts (MORB). Most Icelandic basalts, including the Theistareykir basalts, fall in the stippled field for Icelandic basalts defined by Fitton et al. [1997]. However, based on a larger data compilation than used by Fitton et al. [1997], the lower bound of the Icelandic field does not appear to be discriminative between Icelandic and mid ocean ridge basalts. Therefore, Nb-Zr-Y systematics or ΔNb as defined by the vertical deviation from the lower bound of the stippled array in Figures A3a does not unambiguously distinguish between Icelandic basalts and MORB as has been proposed based on a more limited number of MORB data [Fitton et al., 1997]. The Nb-Zr-Y concentrations in Mid Atlantic ridge (MAR) and East Pacific Rise (EPR) basalts are taken from the Lamont Doherty petrological database (LDEO petrological database at The Iceland literature data are from Chauvel and Hemond [1999], Furman et al. [1991], Hanan et al. [2000], Hardason et al. [1997], Hards et al. [1995], and Kempton et al. [2000]. See text for further discussion. (b) Changes in Nb/Y and Zr/Y ratios in response to variable degrees of melting. Shown are melting trajectories resulting from melting of depleted mantle (DMM: Nb = 0.112ppm, Zr = 6.2ppm, Y = 3.7ppm; GERM-model,, note that melting of a more Zr depleted composition is needed to account for the most depleted Icelandic compositions) and melting primitive upper mantle (PUM, [McDonough and Sun, 1995]). Polybaric melting is used with depth dependent partition coefficients and melt reactions (see caption of Figure 4 and Table 4 for the details of the melting calculations). Ticks indicate the degree of melting in %. Melting of a depleted mantle source is required in order to explain the most depleted Nb/Y and Zr/Y (Nb/Zr) ratios in Icelandic basalts because fertile mantle sources (e.g. PUM) are too enriched whereas previously melted enriched mantle sources (residues) are too depleted in Nb/Y (Nb/Zr, Zr/Y). Therefore, Nb-Zr-Y systematics argue in favor, rather than against [Fitton et al., 1997; Hardarson and Fitton, 1997; Hardarson et al., 1997] the involvement of the depleted MORB mantle. The stippled array in Figure A3a and A3b is the field for Icelandic basalts defined by Fitton et al. [1997]. (c) Variation of ΔNb with the degree of melting (F). ΔNb as defined by Fitton et al. [1997] is the vertical deviation from the lower bound of the stippled array in Figures A3a and A3b.  ΔNb varies irregularly as a function of the degree of melting, i.e. it decreases within the first 10% of melting and then increases for F > 10% and its absolute value is also sensitive to the amount of melting in the garnet stability field. Therefore, ΔNb does not directly reflect the ΔNb of the mantle source.

[59] Fitton et al. [1997] and others [e.g., Hardarson and Fitton, 1997; Hardarson et al., 1997; Kempton et al., 2000] also used Nb-Zr-Y systematics to argue that there is a difference between ambient depleted MORB mantle and the depleted component in the Icelandic source mantle. However, the most depleted Nb/Y and Zr/Y ratios in Icelandic basalts require melting of a depleted mantle source similar to the ambient MORB mantle. Remelting of an originally more fertile residue as suggested by Fitton et al. [1997] results in compositions that are too low in Nb/Y (Nb/Zr, Zr/Y) (Figure A3b). Since melting trends using any composition within or close to the Iceland array are shallower than the Iceland array, the Nb-Zr-Y systematics in Icelandic basalts appear to reflect mainly mixing between depleted and enriched sources (see section 3.2).

[60] Therefore Nb-Zr-Y systematics can neither be used to unambiguously infer source characteristics nor to distinguish between the ambient depleted MORB mantle and the depleted Icelandic source component (Figure A3; see also Hanan et al. [2000]).

A5. Analytical Techniques

A5.1. Sample Preparation

[61] Most of the sample powders analyzed here were prepared by Slater [1996] at Cambridge University. These rock samples were cut with a diamond-impregnated steel blade and the blocks were washed in deionized water. About 60 g of sample were crushed, using a hardened steel mortar and pestle, and powdered in an agate swing mill [see Slater, 1996, appendix B2]. Powders from samples 9306, 9309, 9311, 9313, 9330, 9333, 9356, 9359, 9390, 9391, 9394, 9416, 9476, 9499, 94102, 94112, and TH 29 were prepared at the NHMFL from ∼10 g pieces of rock. Sawed surfaces were polished with quartz sandpaper, washed with deionized water, and crushed with a rock hammer into 5 to 8 mm-sized pieces. Rock powders were then prepared in an agate mill.

A5.2. Chemical Separation and Mass Spectrometric Techniques

[62] Sr, Nd, and Pb separation was carried out at the NHMFL using standard techniques as described by Hart and Brooks [1977], Richard et al. [1976], Zindler et al. [1979], Zindler [1980], and Manhès et al. [1978], respectively. The chemical procedures for the separation of Hf were carried out by A.S. at the Ecole Normale Supérieure in Lyon and at the NHMFL, and follow the techniques described by Blichert-Toft et al. [1997].

[63] Sr and Nd isotope ratios were measured by thermal ionization mass spectrometry (TIMS) using a multidynamic routine on a Finnigan MAT 262 RPQ mass spectrometer at the NHMFL. 87Sr/86Sr and 143Nd/144Nd ratios are normalized, using a power law mass fractionation model [Russell et al., 1978; Hart and Zindler, 1989] to 88Sr/86Sr = 0.1194, and 146Nd/144Nd = 0.7219, respectively. During the analysis period, repeated measurements of 87Sr/86Sr in the E&A SrCO3 standard (n = 16) and 143Nd/144Nd in the La Jolla Nd standard (n = 32) averaged 0.708009 ± 17 (2σ) and 0.511848 ± 11 (2σ), respectively. These values agree well with the longer-term lab averages for these standards (0.708007 ± 19 (2σ) for the E&A standard and 0.511846 ± 12 (2σ) for the La Jolla Nd standard). All replicate sample analyses of Sr and Nd isotope measurements agree to better than the standard-derived estimates for the external precision, and most (9 out of 10 for Sr and 19 out of 22 for Nd isotopes) reproduced within the limits of the in-run precision of less than ±0.000008 (2σm) (Table 2).

[64] Leached and unleached sample powders were analyzed to investigate possible effects of seawater alteration on the 87Sr/86Sr ratios. Sample powders were leached in 6N HCl at 100°C for ∼4 hours. In most cases, differences in 87Sr/86Sr ratios between leached and unleached powders are within the analytical uncertainty (Table 2). Nd isotope ratios were determined partly on leached and partly on unleached samples as our experiments have shown that the leaching has no effect on the Nd isotope compositions. Pb isotope analyses were performed on rock chips (300 mg) that were leached in 6N HCl at 100°C for ∼4 hours.

[65] Hf isotope ratios were obtained by plasma ionization multicollector mass spectrometry using the VG Instruments Plasma 54 in Lyon, according to the techniques described in Blichert-Toft et al. [1997]. 176Hf/177Hf ratios are reported relative to 176Hf/177Hf = 0.282160 for the JMC-475 Hf standard and are normalized to 179Hf/177Hf = 0.7325. As reported by Vervoort and Blichert-Toft [1999], measurements of a 500 ppb solution of the JMC-475 Hf standard yield a long-term average of 176Hf/177Hf = 0.282161 ± 14 (n = 37). Repeated measurements of the JMC-475 Hf standard during the analysis period for the Theistareykir samples averaged 0.282154 ± 28 (2σ, n = 23). The average in-run precision is generally better than ±0.000012 (2σm). All but one of the replicate analyses agreed to within the limits of the external reproducibility given by the JMC-475 results obtained during the time our analyses were made (Table 2).

[66] Pb isotope ratios, likewise, were obtained by plasma ionization multicollector mass spectrometry on the VG Instruments Plasma 54 in Lyon. Analytical procedures are described in detail by White et al. [2000]. Repeated measurements of the NBS 981 Pb standard (n = 33) during the course of performing our analyses yielded 206Pb/204Pb = 16.9385 ± 38 (2σ), 207Pb/204Pb = 15.4909 ± 45 (2σ), and 208Pb/204Pb = 36.6982 ± 13 (2σ), in excellent agreement with the values given by Todt et al. [1996] (206Pb/204Pb = 16.9356, 207Pb/204Pb = 15.4891, and 208Pb/204Pb = 36.7006). Reproducibility for sample analyses is slightly worse than for the standard (see Table 2), as has also been observed by White et al. [2000]. This is mostly attributed to the low amount of Pb analyzed in the samples (< 60 ng), resulting in a larger error in the measurement of 204Pb compared to 206Pb, 207Pb, and 208Pb. This is supported by the observation that reproducibility for the sample with the lowest Pb content (9381; difference between duplicate and original is 0.06%, 0.1% and 0.1% for 206Pb/204Pb, 207Pb/204Pb, and 208Pb/204Pb, respectively) is worse than that for samples with higher Pb content (9222, 9332, 9354, 9435; difference between duplicate and original ranges from 0.02–0.04%, 0.03–0.05% and 0.05–0.07% for 206Pb/204Pb, 207Pb/204Pb, and 208Pb/204Pb, respectively; Table 2).

A5.3. Trace Element Analyses

A5.3.1. ICP-MS analyses

[67] The 100 mg quantities of sample powder were dissolved in 30 mL Savillex beakers with a 3:1 HF:HNO3 mixture for about 48 hours. After evaporating to dryness at about 100°C, samples were redissolved in 7N HNO3 and left in an ultrasonic bath for 30–45 min. At this stage, most samples were completely dissolved. If complete dissolution was not achieved at this point, samples were heated to 175°–200°C for several hours and were repeatedly placed in an ultrasonic bath for ∼10 min intervals during this procedure. This assured complete dissolution of all samples. Samples were then transferred into clean HDPE bottles (pre-cleaned in 7N HNO3 for several hours), diluted to ∼1% HNO3 and a total dissolved solid content of 250 ppm.

[68] The concentrations of 32 trace elements were determined by inductively coupled plasma mass spectrometry (ICP-MS) using a Finnigan MAT ELEMENT high-resolution ICP-MS at the NHMFL. The use of a CD1E interface (“guard electrode” or platinum shield between the load coil and torch) provided a sensitivity of ∼700,000 cps/ppb 115In at sample and auxiliary gas flow rates of ∼1 l/min. An Elemental Scientific Teflon nebulizer with a flow rate of 100 μL/min was used in self-aspiration mode in combination with a Teflon spray chamber and Teflon tubing. The use of Teflon for the inlet system improved washout characteristics significantly compared to a conventional setup consisting of a combination of Tygon tubing and glass nebulizer and spray chamber, and allowed the use of a mixture of 2% HNO3-0.03N HF for cleaning between samples. The use of HF in the cleaning solution especially improved washout characteristics for “sticky” elements such as Nb, Zr, Hf, Pb, Th, and U (by a factor of 20–200), so that blank levels for most elements were < 10–100 cps. A 90 s sample uptake time was chosen in combination with a washout time of 120 s between samples.

[69] To reduce the effects of signal drift, the analysis time was kept short by measuring samples in small sequences consisting of only six unknowns (five samples and one blank) and three standards, so that each group of three unknowns is bracketed by measurements of a rock standard (in this case BIR-1). Thus signal drift was kept small and is usually < 5% between the first and last sample of each sequence. Correction for signal drift is based on an internal standardization to In (all samples and standards are spiked with In to a concentration of 1 ppb) and a linear interpolation between the external rock standards for each isotope. The reference concentrations of BIR-1 are given in Table 1. Repeated measurements were made of two unknowns (the most trace element depleted sample 9390 and the most enriched sample 9396), and of the rock standard BHVO-I to check for accuracy and precision. On the basis of repeated measurements of sample 9390 (n = 8), 9396 (n = 10), and of BHVO-I (n = 7), precision for most elements is about 2% (1σ, Tables 1 and 2). Reproducibility for the rare earth elements (REE) is generally better than 2%, reproducibility for Nb, Ta, Zr, Y, Hf, Rb, Sr, and Th, is generally better than 3%, and Sc, Ti, V, Cr, Co, Ni and U reproduced to better than 5–6% (Tables 1 and 2). The average concentration of BHVO-I agrees very well with the values given by Eggins et al. [1997] (Tables 1 and 2), most elements agree within the limits of the analytical uncertainty with the values reported by Eggins et al. [1997].

[70] Although the use of the CD1E interface improves sensitivity about five times and allows the use of a lower total dissolved solid content in the solution, which was found to improve signal drift considerably (∼4 times less signal drift with a 250 ppm solution compared to 500–1000 ppm solutions over the courses of a two hour measurement), it has the tendency to increase oxide formation up to a factor of 2. Potential significant oxide interferences occur on some of the rare earth elements (REE) such as Eu, Gd and Tb (135BaO on 151Eu, 141PrO on 157Gd, 143NdO on 159Tb). However, in light rare earth (LREE) depleted samples, these oxide interferences are usually insignificant up to oxide formation levels of several percent. Moreover, oxide formation was observed to be approximately constant between samples with similar matrices, so that with the use of a matrix-matched external standard with similar Ba, Pr and Nd concentrations compared to the analyzed samples (as in case of BIR-1 and depleted Icelandic basalts), the error due to oxide formation is to a large degree canceled out by applying the external rock standard correction. However, especially where samples are more enriched in light REE and very incompatible elements than the external rock standard, the effects of oxide formation can be significant, as, for example, between BIR-1 and a LREE enriched rock such as BHVO-I. Comparing our results for the BHVO-1 standard with those measured by Eggins et al. [1997] (Tables 1 and 2) indeed shows that our measured Gd concentrations in BHVO-I are significantly higher than those given by Eggins et al. [1997] (Table 1), whereas most other elements agreed to within the analytical uncertainty of the values of Eggins et al. [1997] (Tables 1 and 2).

A5.3.2. Pb Concentrations by Isotope Dilution (ID)

[71] Extremely variable Pb concentrations measured by ICP-MS, high absolute Pb concentrations compared to other Icelandic samples [Chauvel and Hemond, 1999] as well as differences between Pb isotope values measured on leached chips and unleached sample powders all indicate that the samples were contaminated with Pb during the powdering process and/or are contaminated by some other non-magmatic Pb component (see discussions by McDonough and Chauvel [1991], Chauvel and Hemond [1999], Abouchami et al. [2000]). The excellent reproducibility of Pb isotope analyses of strongly leached sample chips indicates the reliability and success of the leaching procedure for Pb isotope measurements [see also Abouchami et al., 2000].

[72] The effects of leaching on the Pb concentrations were investigated by a series of sequentially more aggressive leaching steps. Pb concentrations were measured in unleached rock chips, chips washed in 2.5N HCl for 20 min at room temperature, for 1 hour in 2.5N HCl at 100°C, and for 1 hour in 6N HCl at 100°C. A slight decrease in the Pb concentrations between unleached chips and chips leached in cold 2.5 N HCl is ascribed to removal of surface contamination. This conclusion appears to be supported by the fact that more aggressive leaching in 2.5N HCl (hot for 1 hour) does not lead to a further decrease in the Pb concentrations, whereas strong leaching in 6N HCl can lead to substantial Pb loss (Figure A4). Replicate measurements in each step confirm good reproducibility for the applied leaching techniques, with the exception of strong leaching with 6N HCl.

Figure A4.

Variability of the Pb concentrations in sample 9381 and 9366 as a function of different leaching techniques. Mild leaching in 2.5N HCl leads to a slight decrease in the Pb concentrations, however, with no difference between brief, cold and long, hot leaching. Thus, the slight decrease from unleached to mildly leached samples is interpreted to remove surface contamination. Note that strong leaching in 6N HCl can lead to substantial Pb loss.

[73] All reported Pb concentrations are determined on chips washed in 2.5N HCl for 20 min. The samples were spiked with a 208Pb spike according to the optimum spike criterion: 208Pb/206Pbmeasured = equation image [e.g., Albarède, 1995] and dissolved in HF and HNO3. Redissolution in HCl assured complete dissolution and spike sample equilibration. Pb separation was carried out using standard techniques as described by Manhès et al. [1978]. Samples were loaded on Re filaments with silica gel and H3PO4 and measured on Finnigan MAT 262 RPQ in static collection mode. On the basis of replicate analyses, reproducibility of the Pb concentrations is better than 2–3% (9330: 0.241 versus 0.246 ppm, 9381: 0.113 versus 0.114 ppm, 9370: 0.159 versus 0.161 ppm, 9366: 1.004–1.013–1.022–1.029 ppm). Pb blanks range from 200–500 pg and are negligible compared to the amount of analyzed sample.


[74] Especially Michael Bizimis but also Barry Hanan, Al Hofmann, Chip Lesher, and Paul Ragland are thanked for constructive discussions. Philippe Telouk and Ted Zateslo are thanked for keeping the Plasma 54 in Lyon and the Finningan MAT 262 at the NHMFL, respectively, in excellent condition. Tim Elliott and Barry Hanan are thanked for their thoughtful and constructive reviews, and Dave Graham and Bill White are thanked for their comments and the editorial handling of the manuscript. A.S. was in part supported by a HSP-III doctoral fellowship from the German Academic Exchange Service (DAAD). D. McKenzie acknowledges support from the NERC and the Royal Society and J. Blichert-Toft and F. Albarède from the Institut National des Sciences de l'Univers. This work was supported by the U.S. National Science Foundation (EAR 1323-608-22 to AZ).