Compositional trends of Icelandic basalts: Implications...

Article

Volume 12, Number 11

18 November 2011

Q11008, doi:10.1029/2011GC003748

ISSN: 1525‐2027

Compositional trends of Icelandic basalts: Implicationsfor short–length scale lithological heterogeneityin mantle plumes

O. Shorttle and J. MaclennanDepartment of Earth Sciences, University of Cambridge, Downing Street, Cambridge CB2 3EQ, UK([email protected])

[1] Lithological variations in the mantle source regions under mid‐ocean ridges and ocean islands havebeen proposed to play a key role in controlling melt generation and basalt composition. Here we com-bine compositional observations from Icelandic basalts and modeling of melting of a bilithologicperidotite‐pyroxenite mantle to demonstrate that, while short–length scale major element variation ispresent in the mantle under Iceland, source heterogeneity does not make an important contribution toexcess melt production. By identifying the major element characteristics of end‐member Icelandic melts,we find enriched melts to be characterized by low SiO2 and CaO, but high FeO. We quantitatively com-pare end‐member compositions to experimental partial melts generated from a range of lithologies, pres-sures and melt fractions. This comparison indicates that a single source composition cannot account forall the major element variation; depleted Icelandic melts can be produced by depleted peridotite melting,but the major element composition of enriched melts is best matched by melting of mantle sources thathave been refertilized by the addition of up to 40% mid‐ocean ridge basalt. The enriched source beneathIceland is more fusible than the source of depleted melts, and as such will be overrepresented in accu-mulated melts compared with its abundance in the source. Modeling of peridotite‐pyroxenite melting,combined with our observational constraints on the composition of the Icelandic mantle, indicates thatcrustal thickness variations in the North Atlantic must be primarily due to mantle temperature and flowfield variations.

Components: 20,000 words, 14 figures, 7 tables.

Keywords: Iceland; major elements; mantle heterogeneity; mantle plumes; melting model.

Index Terms: 1009 Geochemistry: Geochemical modeling (3610, 8410); 1025 Geochemistry: Composition of the mantle;1030 Geochemistry: Geochemical cycles (0330).

Received 14 June 2011; Revised 11 October 2011; Accepted 12 October 2011; Published 18 November 2011.

Shorttle, O., and J. Maclennan (2011), Compositional trends of Icelandic basalts: Implications for short–length scalelithological heterogeneity in mantle plumes, Geochem. Geophys. Geosyst., 12, Q11008, doi:10.1029/2011GC003748.

1. Introduction

[2] Subduction has long been identified as a gen-erator of mantle isotopic and trace element het-erogeneity [Hofmann and White, 1982; White andHofmann, 1982]. However, it is only during the

last 15 years that a significant role for recycledoceanic crustal material in oceanic island and mid‐ocean ridge basalt (OIB and MORB) genesishas become widely proposed [e.g., Hauri, 1996;Hirschmann and Stolper, 1996; Niu et al., 2008;Sobolev et al., 2005; Herzberg, 2006; Prytulak and

Copyright 2011 by the American Geophysical Union 1 of 32

http://dx.doi.org/10.1029/2011GC003748

Elliot, 2007; Sobolev et al., 2007; Dasgupta et al.,2010; Herzberg, 2011]. In particular, these authorsidentify the recycled basaltic component to bepresent as discrete pyroxenitic or eclogitic litholo-gies. Regions in the mantle of distinct mineralogyhave important implications not only for modelsof melt generation, but also for mantle dynamicsand interpretation of geophysical data [e.g., Ita andStixrude, 1992; Van der Hilst and Kárason, 1999;Ritsema et al., 2009a, 2009b].

[3] The identification of clear geochemical proxiesfor the involvement of varying lithologies in basaltgenesis remains a significant challenge. Part of thedifficulty lies in the interpretation of basalt majorelement chemistry, which is both strongly affectedby low‐pressure crystallization, and the depthand extent of melting [Grove et al., 1992; Klein andLangmuir, 1987; Langmuir et al., 1992]. As a result,major element signatures inherited from the miner-alogy of the source are difficult to uniquely dis-tinguish. Therefore, while the trace element andisotopic heterogeneity of the mantle has longbeen recognized [Hart, 1971; Schilling, 1973], thisdiversity was for a long time incorporated into theclassic peridotite model of Ringwood [1962].

[4] Recently, Herzberg [2006, 2011] has inferredthe majority of the shield building phase on Hawaiito be composed of 100% pyroxenite melts, thebasis for this interpretation being the lower CaOcontent of Hawaiian basalts than predicted frompartial melting of peridotite [Walter, 1998]. Simi-larly, the high Ni content of olivines found asphenocrysts in Hawaiian and Icelandic basalts hasbeen taken by Sobolev et al. [2005, 2007, 2008] toindicate the role of olivine‐free lithologies in theirmantle source regions. Focusing on Iceland, pre-vious workers have inferred the presence of recy-cled basalt in the source from incompatible traceelement, radiogenic isotope and stable isotopecompositions of Icelandic basalts [Fitton et al.,1997; Chauvel and Hémond, 2000; Thirlwall et al.,2004; McKenzie et al., 2004; Kokfelt et al., 2006;Peate et al., 2010]. The question remains, however,to what extent these geochemical observationsrequire a lithologically distinct basalt component inthe mantle, as opposed to the material havingreacted with and refertilized depleted peridotites.Equally important is the scale of this heterogeneity;trace elements and isotopes have been used toestablish the presence of short–length scale het-erogeneity in the Icelandic mantle [Zindler et al.,1979; Slater et al., 2001; Stracke et al., 2003;Maclennan, 2008a], however it remains to be

demonstrated that small‐scale source variability isalso present in the major elements.

[5] The aim of this work is to investigate the natureof lithological heterogeneity in the mantle underIceland, and to constrain how such heterogeneitycan influence melt flux in mid‐ocean ridge set-tings. In order to achieve this target, the study isdivided into two parts. In the first part, we system-atically assess major, trace element and isotopicdata from Icelandic basalts, to place constraints onthe lithology of the sources producing the enrichedand depleted melts. We find that while major ele-ment variability of the source beneath Icelandis required, none of the geochemical observationsnecessitate the role of an olivine‐free lithologydirectly contributing to erupted melts. Instead,variably enriched peridotites, refertilized by up to40% MORB material, are capable of generating allthe geochemical patterns observed, including theNi content of olivines from four eruptions repre-sentative of the range of Icelandic melt composi-tions. Enriched and depleted major element signalsexist within similar aged eruptions from a singlevolcanic system, indicating the major elementvariability in the Icelandic mantle source mustbe present on a short length scale. The nature ofthis source enrichment means the enriched lithol-ogy will be more fusible than the depleted source,aggregate melts will therefore be biased towardmelts from the deep melting, productive, enrichedlithology. In the second part of this paper wecombine the constraints on mantle heterogeneityfrom the first part, with the development of a newmodel of melting a bilithological mantle, in orderto investigate the contribution of mantle composi-tional heterogeneity to excess melt generationaround Iceland. Even allowing for melting of thehighly productive G2 pyroxenite [Pertermann andHirschmann, 2003b], oceanic crustal thicknessvariations observed cannot be produced by sourceenrichment alone. Mantle potential temperature (Tp)variations of at least +100°C are required, coupledwith a component of plume‐driven upwelling, togenerate the thickest oceanic crust observed inthe North Atlantic.

[6] For 400 km of its ∼20,000 km length the Mid‐Atlantic Ridge becomes subaerial on Iceland. Thisexposure provides a unique opportunity to gathera high‐resolution spatial and temporal record ofmantle melting at a divergent plate boundary. Wehave compiled a comprehensive major, trace andisotopic data set from both onland and adjacentsubmarine ridges, to characterize the geochemi-

GeochemistryGeophysicsGeosystems G3G3 SHORTTLE AND MACLENNAN: COMPOSITIONAL TREND OF ICELANDIC BASALTS 10.1029/2011GC003748

2 of 32

cal variations associated with the plume andexplore their implications for lithological variationin the North Atlantic mantle. The list of sourcesused in the Icelandic geochemical data set can befound in Appendix A, and in Figure 1 the sampledistribution is mapped.

[7] In Figure 2, the first‐order major element trendsalong the Mid‐Atlantic Ridge are displayed. Ice-land lies at the extreme end of the array of globalMORB compositions, having relatively low Na8,Si8 and high Fe8 contents (Figures 2a, 2c, and 2e,correction described in the figure caption). How-ever, there is an offset in Iceland’s position withrespect to the array defined bymost AtlanticMORB;with the onland average of Icelandic basalts lyingto slightly higher Na and lower Si than would beexpected given the elevation of the ridge axis. Thisoffset hints at the combined role that meltingdynamics and source composition play in a plume‐influenced section of mid‐ocean ridge, in additionto Tp variations. The along‐ridge trends of Figures 2b,2d, and 2f show the specifics of geochemical var-iation on and away from Iceland; basalt Si contentshows especially clear variation, dropping sharplyin the volume‐averaged composition of onlandbasalts. These major element patterns raise thequestion of whether lithological heterogeneity isrequired in the mantle source beneath Iceland. Weaddress this possibility, first by identifying themajor element characteristics of the enriched and

depleted melts erupting on Iceland, and determiningif nonperidotitic lithologies are necessary to matchthe end‐member compositions.

2. Characterizing the IcelandicEnd‐Member Melt Compositions

2.1. The Signal of Basalt Enrichmentand Depletion in Iceland

[8] Trace element and isotopic studies on Icelandicbasalts have established the presence of distinctdepleted and enriched compositions [e.g., Elliott et al.,1991; Hémond et al., 1993; Chauvel and Hémond,2000; Fitton et al., 2003; Thirlwall et al., 2004].Trends of enrichment in Icelandic basalts show bothincompatible trace element and isotopic enrichmentcorrelating positively [e.g., Hémond et al., 1993;McKenzie et al., 2004; Stracke et al., 2003]. Acorollary of this is that trace element ratios canbe used as a proxy for mantle source enrichment.This is demonstrated by Stracke et al. [2003], wherepositive correlations are found between incompati-ble trace element and isotopic compositions ofbasalts from Theistareykir. These relationships alsoextend to the major elements, which covary withincompatible trace element ratios. Both extent ofmelting and the fusibility of the different lithologiesare convolved in this pattern of joint isotopic andtrace element enrichment, but despite this the iso-

Figure 1. Map of Iceland showing the geographical distribution of data used in this study, with the neovolcaniczones marked in dark grey. The eruptions from which new olivine data are presented are marked: Thp, Theistareykirpicrite; Gae, Gaesafjöll; Hal, Háleyjabunga; Stp, Stapafell. Inset shows Iceland in relation to the North Atlantic andthe along‐ridge distribution of compiled literature data. Data sources listed in Table 6.


3 of 32

topic data of Stracke et al. [2003] show the trend tobe consistent between different eruptions.

[9] The covariation of trace element and isotopicenrichment in Icelandic basalts means we can useNb/Zr as a proxy for enrichment. The choice of thisratio was partly pragmatic: while similar major andtrace element relationships are obtained by takingREE ratios, the availability of published major andtrace element data on the same samples is greatestfor those elements accurately analyzed by XRF.

Equally, Nb/Zr is resistant to change during frac-tional crystallization; neither this process, nor mix-ing with partial melts of altered crust, dominatethe Nb/Zr of moderately evolved basalts (MgO >5wt%), their compositions instead reflect an aver-age of the melts derived from the mantle. There-fore, at high MgO Nb/Zr preserves a signal ofmantle source enrichment. In this section we seekto identify the major element variation associatedwith high and low Nb/Zr of Icelandic basalts.

Figure 2. (a, c, and e) Icelandic major element chemistry in context with Atlantic MORB. Only samples with com-positions between 7.5 and 8.5 wt% MgO have been selected so as to diminish the effects of fractional crystallizationand provide an equivalent to the Klein and Langmuir [1987] global array. In calculating the onland Icelandic mean,samples have been weighted by the volume of the eruption they relate to. (b, d, and f) The major element chemistryalong the plume‐influenced portion of Mid‐Atlantic Ridge, plotted as a function of radial distance from the plumecenter [Shorttle et al., 2010]. As before, only 7.5 to 8.5 wt% MgO samples have been selected, and onland data hasbeen volume averaged every 50 km of ridge; vertical bars record the one sigma variation about the mean. Data sourceslisted in Table 6.


4 of 32

2.2. Identifying Geochemical VariabilityFrom the Mantle

[10] One method that has proved effective in cap-turing a broad spectrum of geochemical variability,is the analysis of olivine‐hosted melt inclusions[Sobolev and Shimizu, 1993; Slater et al., 2001;Maclennan, 2008a]. However, while these inclu-sions can faithfully record the isotopic and traceelement characteristics of the basalts in which theyform, their preservation of major element signals isstrongly affected by post entrapment crystallizationand diffusion [Gaetani and Watson, 2000]. There-fore, major element compositions of melt inclusionsmay not give reliable constraints on variations inmantle source lithology. Instead, we take whole‐rock geochemical data from multiple eruptionswithin the neovolcanic zones to capture the geo-chemical diversity of melts leaving the mantle.Although a given eruption may only have limitedwhole‐rock variability [Maclennan et al., 2003a], byexamining analyses from many eruptions, reflectingmelts mixed and fractionated to differing degrees, itis possible to recover a larger proportion of geo-chemical diversity.

[11] In constraining end‐member melt composi-tions it is necessary to be able to identify the role ofmixing and crystallization processes, and isolatethose eruptions with more primitive chemistries.We therefore construct plots of whole rock Nb/Zragainst MgO, and color the data points by a secondmajor element in order to identify major elementvariation that is caused by mantle rather thancrustal processing. We apply this type of plot totrack the effects of concurrent crystallization andmixing processes, and to select melts reflecting theenriched and depleted end‐member compositions;enriched melts will have high Nb/Zr and depletedmelts low Nb/Zr, with this initial variabilitydecreasing as the melts evolve. The color of thedata points then demonstrates how the major ele-ment systematics relate to the enriched and depletedmantle melts.

2.3. Major Element Composition of End‐Member Neovolcanic Zone Melts

[12] Separate plots of major and trace element datahave been constructed for the Northern and West-ern Volcanic Zones (NVZ & WVZ) and are shownin Figures 3 and 4. In each zone, basalts showevidence for concurrent mixing and crystallization,with basalts of >9.5 wt% MgO displaying themaximum diversity of Nb/Zr. Below 9.5 wt%MgO, variability of Nb/Zr decreases monotonically

as storage, mixing and crystallization average outthe primary geochemical signals exiting the mantle.By 5 wt% MgO, mixing has almost completelyhomogenized the basalts to a weighted mean ofthe compositions input to the base of the crust[Maclennan, 2008b].

[13] The relation of the trace element chemistryto the major elements is highlighted by the colorof data points in Figures 3 and 4. As expected,there is a positive correlation between incompati-ble element concentrations (Ti, Na, K) and Nb/Zrat fixed MgO. There is also a systematic relation-ship between Nb/Zr and the SiO2, FeO and CaOcontent of primitive melts. The trend in both theNVZ and WVZ is for more enriched basalts tohave lower SiO2 and CaO contents, but with higherFeO, than the depleted melts. This pattern of cou-pled FeO and trace element enrichment has beendescribed in Iceland by Langmuir et al. [1980],Elliott et al. [1991], and Nicholson and Latin[1992], with correlations attributed to extent anddepth of melting. However, our finding of low CaOin the most enriched melts is contrary to the find-ings of previous authors [McKenzie et al., 2004].The difference between our finding and that ofMcKenzie et al. [2004] can be attributed to con-current mixing and crystallization; if a basalt ini-tially has low CaO and Nb/Zr (i.e., is depleted),olivine crystallization will both lower MgO andraise CaO, while mixing with more enriched meltswill raise Nb/Zr (and to a lesser extent lower CaO).The net effect is a trend of increasing CaO and Nb/Zr, purely a product of crustal processes. Presentingthe data as in Figures 3 and 4 controls for theseprocesses, and allows the mantle derived majorelement signal to be identified.

[14] Figures 3 and 4 display depleted and enrichedend‐member compositional characteristics qualita-tively, but in order to make a systematic compari-son with the results of experimental petrology weneed to estimate compositions for each end‐member.We do this by taking an average of the ten percent ofbasalts with the highest and lowest Nb/Zr. The pointsselected in this way are circled in black in Figures 3aand 4a, and their means presented in Figures 3band 4b and given in Table 1. While there is someoverlap between end‐member compositions, SiO2,FeO, CaO, Na2O, K2O and TiO2 from both zonesshow clear differences between the two populationsof basalts selected (Table 1).

[15] A possible complication of using whole‐rockdata to compile these plots is that crystal accumu-lation can shift whole‐rock compositions to higher


5 of 32

or lower MgO, while leaving Nb/Zr unaffected.The sensitivity of the major element patternsdescribed in Figures 3 and 4 to crystal accumula-tion is tested in Appendix B. The result is that,while some of the variability in major elementcompositions of high or low Nb/Zr basalts will bedue to accumulated crystals, the first‐order trendscannot be explained by the addition of olivine ±clinopyroxene ± plagioclase. A further test ofwhether the enriched melt signal is being correctlyidentified in the neovolcanic zone basalts is bycomparison to flank zone melts. In the absence ofwell‐developed extensional tectonics and with athicker lithosphere, flank zone melts are on averagethought to represent smaller degrees of melting, at

higher pressures, than on‐axis melts [Kokfelt et al.,2006; Peate et al., 2010]. Their aggregate compo-sitions are thus expected to be weighted towardenriched fusible lithologies. It can be seen inFigures 3b and 4b that the enriched flank zonemelts show the same systematics as the enrichedneovolcanic zone melts. We can thus conclude, thatenriched basalts in Iceland are consistently char-acterized by lower SiO2 and CaO, and higher FeO,TiO2, Na2O and K2O than depleted melts.

[16] For the remainder of this contribution weconsider only end‐member compositions of theNVZ basalts when comparing to other data sets. Asalready discussed, the major element and trace

Figure 3. Plots relating the major element chemistry of NVZ basalts to Nb/Zr. (a) Each plot shows the same datacolored by a different major element. Plotting the data this way allows the processes controlling major element chem-istry to be constrained. Magmas with more than 9.5 wt% MgO have experienced primarily olivine fractionation, andso their major and trace element chemistry remains largely controlled by source composition and melting processes.The distribution of grey data points, below 9.5 wt% MgO, shows predominant control by polyphase fractional crys-tallization and melt mixing. Circled points were selected to represent end‐member melt compositions. (b) Mean end‐member melt compositions, with boxes marking the one sigma variation. The enriched flank zone melt compositionwas selected by taking the mean of the 10% of flank lavas with the highest Nb/Zr and MgO > 8.5 wt% and is shownfor comparison.


6 of 32

element systematics between the zones are verysimilar (Table 1), and so there is no significantdifference in taking either the WVZ, NVZ or anaverage of the two. The choice of the NVZ end‐member compositions was motivated by the greaterabundance of trace element data available fromNVZ basalts. The major element composition ofthe NVZ end‐member melt is given in Table 1.

3. Identifying the Lithologies ProducingEnd‐Member Icelandic Melts

[17] With the major element characteristics of Ice-landic depleted and enriched end‐member meltsidentified, the next step is to test whether this rangeof major element compositions is consistent withmelting of a single peridotitic source composi-tion [Langmuir et al., 1980; Elliott et al., 1991;Nicholson and Latin, 1992]. If varying depthsand degrees of melting of a single source cannot

match the range of major element compositionsobserved between the end‐member melts, then thisis evidence for the presence of lithological varia-tion in the mantle beneath Iceland. We then needto determine the nature of the lithological hetero-geneity this corresponds to, such as the presenceof olivine‐free pyroxenite or clinopyroxene richperidotites.

[18] There has recently been discussion as to whetherparticular major element characteristics requirethe presence of pyroxenite in the source [Herzberg,2011; Putirka et al., 2011]. In particular, the lowCaO content of some Hawaiian tholeiites and oli-vines has been interpreted as their primary meltsbeing partly, or wholly, derived from pyroxenite[Sobolev et al., 2007; Herzberg, 2006, 2011]. Asfound in the previous section, enriched Icelandicbasalts are similarly distinguished by their lowerCaO content than the depleted basalts. This raisesthe question of whether pyroxenite melting beneath

Figure 4. Plots relating the whole‐rock major element chemistry of WVZ and REP basalts to Nb/Zr. Details as inFigure 3.


7 of 32

Iceland should be invoked to explain the majorelement systematics of the enriched basalts.

[19] For a pyroxenite lithology to be a plausiblemantle source, its melts ought to provide a betterfit to the observed basalt chemistry than thatoffered by a lherzolite partial melt. Importantly, thepyroxenite partial melt must improve the total fitacross all of the major elements, and not just CaO;given the along‐ridge patterns of major elements inFigure 2, and the differences between the enrichedand depleted end‐members in Figure 3, it is evidentthat suitability of source lithology needs to at leastbe assessed in the CaO‐FeO‐MgO‐Al2O3‐SiO2

(CFMAS) system. The approach we take is there-fore to systematically and quantitatively comparethe end‐member melt compositions with the resultsof a wide range of partial melting experiments, foreach of the major element oxides. Quantificationof the misfit between experimental partial meltsand the Icelandic end‐member compositions thenallows for selection of those lithologies withexperimental melts closest to the composition ofbasalts on Iceland.

[20] This type of analysis, comparing directly exper-imental and natural melts, requires melts producedfrom a given lithology for each melting scenario tobe compositionally alike, despite the difference inmelting style. If isobaric batch melting producesmelts radically different in composition from thepolybaric fractional case, then to see through thisprocess difference to major element heterogeneityin the source is hopeless. We address this issue inAppendix C, and conclude that for Iceland at least,given the lithologies that provide the best fittingmelt compositions, the difference between experi-ment and nature should not be influencing ouridentification of source heterogeneity.

3.1. Method of Calculating Misfit BetweenEnd‐Members and Experimental Melts

[21] We first compiled a data set of 377 partialmelting experiments with lherzolite, eclogite, pyrox-enite and harzburgite starting materials, over a widepressure range, and with hydrous and carbonatedexperiments incorporated. The full list of experi-mental data sources is presented in Appendix A,

Table 1. Estimated Compositions of Depleted and Enriched and Flank Zone End‐Member Meltsa

Element Depleted 1s Enriched 1s pdepleted–enrichedb pNVZ–WVZ

depleted pNVZ–WVZenriched

NVZSiO2 48.7 0.6 47.8 0.9 0.004 0.261 0.218TiO2 0.69 0.11 1.61 0.28 0.000 0.001 0.945Al2O3 14.9 0.9 14.8 0.6 0.531 0.025 0.379FeO 9.2 0.6 11.3 0.7 0.000 0.009 0.379MgO 11.8 1.2 10.7 1.2 0.011 0.888 0.925CaO 12.9 0.6 11.2 0.6 0.000 0.228 0.121Na2O 1.7 0.2 2.02 0.2 0.000 0.088 0.075K2O 0.04 0.01 0.23 0.09 0.000 0.000 0.002

WVZSiO2 48.3 0.5 47.8 0.5 0.087 ‐ ‐TiO2 0.56 0.13 1.63 0.14 0.000 ‐ ‐Al2O3 15.7 0.9 14.3 0.9 0.002 ‐ ‐FeO 8.8 0.6 11.2 0.2 0.000 ‐ ‐MgO 11.8 1.7 11.1 1.6 0.489 ‐ ‐CaO 13.1 0.6 11.4 0.5 0.000 ‐ ‐Na2O 1.5 0.2 1.9 0.1 0.000 ‐ ‐K2O 0.01 0.01 0.27 0.06 0.000 ‐ ‐

Flank ZoneSiO2 ‐ ‐ 47.5 0.7 ‐ ‐ ‐TiO2 ‐ ‐ 2.05 0.4 ‐ ‐ ‐Al2O3 ‐ ‐ 14.2 1.0 ‐ ‐ ‐FeO ‐ ‐ 10.3 0.6 ‐ ‐ ‐MgO ‐ ‐ 11.1 1.8 ‐ ‐ ‐CaO ‐ ‐ 11.4 0.3 ‐ ‐ ‐Na2O ‐ ‐ 2.1 0.2 ‐ ‐ ‐K2O ‐ ‐ 0.81 0.16 ‐ ‐ ‐

aFor both zones the p values indicate the significance of the compositional difference between the enriched and depleted end‐members.bThe p value significance estimate from a Kolmogorov‐Smirnov test, returning the probability that the two data sets being compared are drawn

from the same distribution [Press et al., 1992, 757–1152].


8 of 32

with a description of how the data were filtered.The misfit between melts in this data set and theIcelandic end‐member compositions is determinedfor CFMAS and CFNKTMAS (CaO‐FeO‐Na2O‐K2O‐TiO2‐MgO‐Al2O3‐SiO2) separately. Calcula-tions proceed according to the following steps.

[22] 1. An experimental partial melt is comparedoxide by oxide to the end‐member melt compo-sition. The absolute differences between oxides(in wt%) of the experimental partial melt (wx) andthe end‐member composition (w0) are summed,giving an initial total misfit (M) from equation (1),

M ¼X

i

jwxi � w0

i j; ð1Þ

where the subscript i refers to a particular oxide.

[23] 2. As it is likely that our estimated enrichedand depleted end‐member compositions haveexperienced some olivine fractionation or accumu-lation, the next stage tests both incremental frac-tional crystallization and accumulation of olivinefrom/into the experimental melt to determine ifthe misfit is improved. Removal/addition of olivineoccurs in 0.05wt% steps, calculating the equilibriumcomposition according to Herzberg and O’Hara[2002].

[24] 3. After each increment of olivine removal oraddition the misfit is recalculated as in step 1.

[25] 4. If, after a crystallisation/addition step, theexperimental melt composition (modified by extrac-tion of olivine) is closer to the end‐member meltcomposition (has a lower misfit), then crystallisation/addition of olivine continues. Crystallisation/additionof olivine is stopped when the misfit begins toincrease again, or the mass fraction removed/addedreaches 50% of the original mass. The final misfitof the experimental melt is taken as the lowestreturned from the crystallization and accumulationsteps. If instead, neither olivine removal nor addi-tion can improve the misfit, then the initial misfitcalculated in step 1 is used.

[26] With the above procedure, each experimentalmelt is assigned a misfit with respect to depleted,enriched and flank compositions. To assess thesignificance of the calculated misfit, we propagateboth the total error in the analyses of experimentalmelt compositions across the oxides considered andthe uncertainty of Icelandic end‐member compo-sitions, through the misfit calculation. An experi-mental melt having a summed misfit (M) less thanthe calculated 2s error is then considered a sig-nificant match to the end‐member composition.

3.2. Comparison of End‐MemberCompositions With Experimental PartialMelts

[27] The results of quantifying misfit between Ice-landic end‐member compositions and melts fromthe experimental database are recorded in Table 2.Only the melts with a summed misfit less than thetotal 2s propagated uncertainty are included. Thetable also provides the experimental conditions foreach melt composition. Results are divided up first,according to whether comparison is to the neo-volcanic zone enriched or depleted end‐member, orthe flank zone enriched end‐member composition;secondly, by whether the misfit calculated is forboth major and minor elements (CFNKTMAS) ormajor elements alone (CFMAS).

[28] The most important result is the dominance oflherzolite lithologies in producing melts whichmatch the Icelandic end‐member compositions(Table 2). Although melts from pyroxenite andeclogite partial melting experiments are included inthe database, none of these proved adequate fits tothe Icelandic depleted end‐member, enriched end‐member or flank zone compositions. In addition,apart from the one high degree melt of PERC (acarbonated lherzolite [Dasgupta et al., 2007]) fittingthe flank zone composition, all other matching meltsare from anhydrous peridotite lithologies.

[29] There are however important differences betweenthe starting materials melting to generate the depletedand enriched compositions. The depleted end‐memberis consistently matched by the 3 GPa melts fromexperiments on KR4003 by Walter [1998]. Thisperidotite represents a pyrolitic composition, close tothat of KLB‐1, with an Mg# of 89.2 [Walter, 1998](Table 3). In contrast, the enriched end‐members arematched by melts from KG1/KG2 [Kogiso et al.,1998]; with these peridotites contributing the onlypartial melts within uncertainty of the neovolcaniczone enriched composition. The starting materials forKG1 and KG2 are homogeneous mixtures of KLB‐1and average MORB in ratios of 1:1 and 2:1, respec-tively [Kogiso et al., 1998]. They therefore representhighly fertile peridotites, with a lower bulk Mg#(81–85) and higher FeO, CaO, Na2O and TiO2

concentrations than are present in KR4003 (Table 3).The consequence of these major element differencesis for melts of KG1 and KG2 to have higher FeO,and lower SiO2 and CaO than melts of KR4003 atequivalent pressures and melt fractions.

[30] The effect on solidus temperatures is poorlyconstrained by the limited pressure overlap between


9 of 32

the experiments of Kogiso et al. [1998] and Walter[1998]. However, in the 2–3 GPa range wherethere are data available, KG1 and KG2 exhibit sol-idus temperatures lower by up to several tens ofdegrees. We can extrapolate from this, given theimportance of incompatible elements such as K

in controlling the onset of melting [Hirschmann,2000], and Na in determining modal clinopyr-oxene content, that KG1 and KG2 are likely to havelower solidus temperatures and higher productiv-ities than KR4003 over a broad pressure interval

Table 2. The Results of Determining the Misfit Between Experimental Partial Melts and Icelandic End‐Member Compositionsa

Misfit (wt%) Material Lithology a/h/cb P (GPa) T (°) Fc (%) Mineralogyd Reference

Depleted, Major Elements1.18 KR4003 lrze a 3.0 1540 24.40 ol:opxf Walter [1998]1.48 MPY 87 lrz a 1.0 1328 23.00 ol:opx Falloon et al. [2001]1.99 KR4003 lrz a 3.0 1530 18.50 ol:opx:cpx Walter [1998]2.46 KR4003 lrz a 4.0 1660 38.80 ol:opx Walter [1998]2.52 TQ 40 lrz a 1.0 1495 36.40 ol:opx Falloon et al. [2001]

Depleted, Major and Minor Elements2.31 KR4003 lrz a 3.0 1540 24.40 ol:opx Walter [1998]3.13 KR4003 lrz a 3.0 1530 18.50 ol:opx:cpx Walter [1998]

Enriched, Major Elements1.76 KG2 lrz a 3.0 1525 22.00 ol:cpx:gt Kogiso et al. [1998]1.90 KG1 lrz a 3.0 1525 32.00 ol:cpx:gt Kogiso et al. [1998]

Enriched, Major and Minor Elements3.66 KG1 lrz a 3.0 1525 32.00 ol:cpx:gt Kogiso et al. [1998]3.80 KG2 lrz a 3.0 1525 22.00 ol:cpx:gt Kogiso et al. [1998]

Flank Zone, Major Elements1.11 KG2 lrz a 3.0 1525 22.00 ol:cpx:gt Kogiso et al. [1998]2.05 KG1 lrz a 3.0 1525 32.00 ol:cpx:gt Kogiso et al. [1998]2.47 KR4003 lrz a 4.0 1610 12.90 ol:cpx:gt Walter [1998]2.54 KR4003 lrz a 4.0 1660 38.80 ol:opx Walter [1998]2.93 KR4003 lrz a 4.5 1650 37.20 ol:opx:gt Walter [1998]3.48 PERC lrz c 3.0 1550 40.30 ol:opx Dasgupta et al. [2007]

Flank Zone, Major and Minor Elements2.68 KG2 lrz a 3.0 1525 22.00 ol:cpx:gt Kogiso et al. [1998]3.46 KG1 lrz a 3.0 1525 32.00 ol:cpx:gt Kogiso et al. [1998]3.64 KR4003 lrz a 4.0 1610 12.90 ol:cpx:gt Walter [1998]

aThe details of the partial melts lying within a 2s uncertainty of the end‐member compositions are shown here. Calculations were performedseparately, comparing major elements only (CaO, FeO, MgO and SiO2), and major and minor elements together (also Na2O, K2O and TiO2) foreach of the end‐member compositions considered. Details of the calculations can be found in the text.

bAbbreviations: a, anhydrous; h, hydrous; c, carbonated experiment.cExtent of melting estimated from experimental charge.dResidual mineralogy after melting.eAbbreviations: lrz, lherzolite; px, pyroxenite.fAbbreviations: ol, olivine; opx, orthopyroxene; cpx, clinopyroxene; sp, spinel; gt, garnet.

Table 3. Experimental Starting Materials Producing the Best Fitting Melts to Icelandic End‐Member Basalt Compositions

Material SiO2 TiO2 Al2O3 FeOa MnO MgO CaO Na2O K2O Cr2O3 Mg#b

KR4003c 44.9 0.16 4.26 8.02 0.13 37.3 3.45 0.22 0.09 0.41 89.2KG2d 46.22 0.57 7.69 9.22 ‐ 28.83 6.05 1.11 0.09 0.22 84.8KG1d 46.97 0.78 9.75 9.77 ‐ 23.57 7.35 1.52 0.12 0.17 81.1

aFeO represents total Fe as FeO.bMg number calculated on molar (Mg/(Mg + Fe)) × 100.cWalter [1998].dKogiso et al. [1998].


10 of 32

[Hirschmann, 2000;Pickering‐Witter and Johnston,2000; Schwab and Johnston, 2001].

[31] The ability of experimental melts to match thespecific oxide compositions of the Icelandic end‐members is presented graphically in Figure 5, withthe best fitting melts from Table 2 highlighted.Considering only CFMAS, the matching experi-mental melts fit the end‐member compositionswith little systematic misfit. However, introducingthe minor elements reveals consistent differencesbetween the experimental compositions and theIcelandic end‐members. Figure 5b shows that thematching experimental melts have systematicallylower Na2O than the depleted Icelandic end‐member, and almost every melt from the experi-mental database to have a higher K2O content.Similarly, the melts of KG1 and KG2 both over-estimate the TiO2, Na2O and K2O concentrations ofthe enriched end‐member (Figure 5d). To an extentthese misfits can be understood as reflecting thedifferent nature of melting in an experimentalcharge, isobaric batch melting, to that which pro-duced the melts erupted on Iceland, polybaricfractional melting. This difference will have thebiggest impact on the incompatible elements, whichwill have been almost totally stripped from thesource by the point at which the depleted meltsform and separate. Hence, unless the startingmaterial experiences prior depletion, batch meltingexperiments are not going to be able to reproducethe low K2O contents of natural high degree melts.The higher concentration of incompatibles in KG1and KG2 melts compared with the enriched end‐member, possibly reflects mixing of the enrichedIcelandic signature with more depleted melts dur-ing transport, a process exacerbated by the highdegrees of melting likely beneath Iceland Strackeet al. [2003]. That this same pattern of incom-patible element overenrichment in KG1 and KG2is not observed with respect to the flank zonemelts (which are thought to derive from smallerextents of melting [Kokfelt et al., 2006; Peate et al.,2010], and therefore are presumably less likely toundergo mixing with depleted melts) supports thishypothesis (Figure 5).

[32] Very few of the experiments considered con-tain melt fractions less than ∼10%. This leavesthe possibility of a small‐degree partial melt of aperidotite in fact being able to match the enrichedIcelandic composition, without the need for sourceenrichment. However, we know from the 3 GPaexperiments ofDavis et al. [2009, 2011] on a KLB‐1‐like starting composition, that this is unlikely to bethe case. Although the SiO2 and CaO content

of small‐degree melts of peridotite are similar tothe enriched Icelandic end‐member (45 wt% and10.5 wt%, respectively [Davis et al., 2009, 2011]),the FeO content (9 wt%) remains too low. Thisinability of small‐fraction melts of KLB‐1‐likecompositions to match the FeO content of Icelandicbasalts is a key indication of source major elementheterogeneity.

[33] Allowing for discrepancies resulting frombatch melting versus fractional melting, the peri-dotites KG1/KG2 and KR4003 provide adequatefits to the major element systematics of Icelandicbasalts. Specifically, the FeO enrichments and lowSiO2 and CaO contents of enriched basalts can begenerated by the high‐pressure melting (≥3 GPa) ofperidotite lithologies, made more fertile through theaddition of MORB‐like material. Direct melting ofpyroxenitic lithologies however, proves unable tomatch the systematics of Icelandic end‐members.Despite both depleted and enriched lithologiesbeing matched by peridotite melting, it is importantto note that the highly fertile nature of the KG1/KG2 like enriched lithology beneath Iceland, makesit likely to have a lower solidus temperature andincreased productivity compared with the depletedlithology.

3.3. Comparison of End‐MemberCompositions With Melts Mixed From TwoSource Lithologies

[34] In the previous section we found that anhy-drous peridotite lithologies of varying fertility canmatch the major element patterns of depleted andenriched Icelandic basalts. It has, however, beensuggested that melts from pyroxenite sources con-tribute 20–40% by mass of the enriched meltserupting on Iceland, with the remaining fractionformed of peridotite partial melts [Sobolev et al.,2007, 2008]. As the misfit calculations presentedabove looked only at the compositions of directmelts from single lithologies, there is the possi-bility that the highly fertile KG1/KG2 melts,found to match the enriched end‐member, couldbe replicated by more refractory peridotite meltsmixed with pyroxenite melts. To assess whether aperidotite‐pyroxenite melt blend is consistent withthe major element patterns, we rerun the misfit cal-culations allowing for mixing between peridotitelithologies and the Px‐1 pyroxenite melts of Sobolevet al. [2007].

[35] The procedure followed for the misfit calcu-lations is the same as in the previous section,


11 of 32

Figure 5. The quality of fit of experimental melts to Icelandic end‐member melt compositions. Experimental meltcompositions have been normalized to the depleted or enriched end‐member melt. The fits to major elements only,and fits to both the major and minor elements are shown separately. Drawn as black lines are the experimental meltsfrom Table 2 that match, within uncertainty, Icelandic end‐member melt compositions. The average fits of partialmelts from other lithologies are shown by the colored lines. (a and b) Experimental melts normalized to the depletedneovolcanic zone end‐member melt compositions, and (c and d) experimental melts normalized to neovolcanic zoneenriched end‐member. (e and f) Experimental melts normalized to the enriched flank zone melt composition.


12 of 32

simply iterated to allow for varying proportions ofperidotite‐pyroxenite melt mixtures (equation (2)),

Mn ¼X

i

wni � w0

i

�� ; ð2Þ

where Mn is the misfit for the particular mixingproportion n of peridotite and pyroxenite experi-mental melts, and wi

n the oxide composition ofthat mixed melt. The returned misfit for a particularmelt combination is that of the best fitting from allthe different peridotite:pyroxenite melt proportionscycled through (i.e., the smallest Mn). The resultsof these calculations in matching the enrichedIcelandic end‐member are shown in Table 4. Onlymixtures within uncertainty of the end‐membercomposition are displayed, but results have alsobeen filtered on the basis of physical plausibility:the pyroxenite melt must represent a higher degreeof melting and be formed at a higher temperatureand pressure than the peridotite melt for the mix-ture to be considered valid. Filtering the results likethis ensures that only melts which could haveformed from the same upwelling column of mantleare combined; the lower solidus temperatures andgreater fusibility of pyroxenite lithologies meanthat they should always have melted to a higherdegree and at an average higher pressure than anysurrounding peridotite.

[36] Calculations allowed mixing between Px‐1melts and melts from a wide range of peridotitelithologies. Table 4 shows that, KG1/KG2 are thepreferred peridotite melt sources in models thatmatch the enriched end‐member major elementcomposition. The addition of melts from the garnetpyroxenite Px‐1 has improved the fit of the melt

mixes to the end‐member composition, but thisimprovement is slight (compare with pure KG1/KG2 melts Table 2). The mass fraction of addedpyroxenite melt is also small: at most 10 wt% hasbeen added to the melts of KG1, and 5 wt% toKG2, to optimize their match to the enriched end‐member. A breakdown of the improvement infit for individual oxides is shown graphically inFigure 6. Compared to Figures 5a and 5b, there islittle change in the pattern of misfit in CFMAS orCFNKTMAS. The offsets that do remain betweenKG1/KG2 melts and the Icelandic enriched end‐member are therefore not resolved by mixing witha direct pyroxenitic melt, and probably result fromdifferences in the style of melting between exper-iment and nature.

[37] These observations indicate that mixing betweenperidotite‐ and pyroxenite‐derived melts is notnecessary to match the major and minor elementsof the Icelandic enriched end‐member. The clearestsign of this is that the peridotite component thatthe pyroxenite melts are mixing with, is itselfcapable of melting to produce, within uncertainty,the Icelandic end‐member composition. Px‐1 meltscannot by themselves match the enriched Icelandiccomposition, and any addition of pyroxenite‐derived melt in the quantities suggested by Sobolevet al. [2007, 2008] generates a poorer fit than onlyadding small amounts.

[38] To summarize the main results of this section:The major and minor element systematics ofenriched Icelandic basalts are explained by themelting of a single fertile peridotite source, similarin composition to KG1/KG2. The depleted end‐member melts can be derived from melting of

Table 4. The Results of Misfit Calculations Between Peridotite Partial Melts and the Pyroxenite Melts of Px‐1 From Sobolev et al.[2007], to Determine if Mixing of Melts From a Peridotite and Pyroxenite Can Explain the Icelandic Enriched End‐Member MeltComposition

Misfit Materials Fpxa P (GPa) T (°C) F (%) Residual Mineralogy

Enriched, Major Elements1.59 KG1 + Px‐1b 0.10 3.0/3.5 1525/1540 32.0/72.0 ol:cpx:gt/opx:cpx:gt1.60 KG1 + Px‐1 0.09 3.0/3.5 1525/1575 32.0/79.0 ol:cpx:gt/opx:cpx:gt1.61 KG1 + Px‐1 0.09 3.0/3.5 1525/1550 32.0/76.0 ol:cpx:gt/opx:cpx:gt1.61 KG2 + Px‐1 0.05 3.0/3.5 1525/1575 22.0/79.0 ol:opx:gt/opx:cpx:gt1.61 KG2 + Px‐1 0.05 3.0/3.5 1525/1550 22.0/76.0 ol:opx:gt/opx:cpx:gt

Enriched, Major and Minor Elements3.05 KG1 + Px‐1 0.10 3.0/3.5 1525/1540 32.0/72.0 ol:cpx:gt/opx:cpx:gt3.09 KG1 + Px‐1 0.09 3.0/3.5 1525/1575 32.0/79.0 ol:cpx:gt/opx:cpx:gt3.10 KG1 + Px‐1 0.09 3.0/3.5 1525/1550 32.0/76.0 ol:cpx:gt/opx:cpx:gt3.49 KG2 + Px‐1 0.05 3.0/3.5 1525/1575 22.0/79.0 ol:opx:gt/opx:cpx:gt3.49 KG2 + Px‐1 0.05 3.0/3.5 1525/1550 22.0/76.0 ol:opx:gt/opx:cpx:gt

aMass fraction of pyroxenite derived melt in the melt mixture.bKG1 and KG2 partial melts from Kogiso et al. [1998].


13 of 32

KLB‐1‐like peridotites, but melts of a MORB‐like pyroxenite cannot mix with these peridotitemelts to produce the enriched end‐member.

4. Trace Element Constraints on SourceLithology

[39] The trace element contents of basalts and ofthe olivines they crystallize have been used both tosupport and refute the presence of pyroxeniticlithologies in the source of oceanic island basalts[Hirschmann and Stolper, 1996; Stracke et al.,1999; Kokfelt et al., 2006; Sobolev et al., 2007;Herzberg, 2011]. In this section we first demon-strate that the compatible and incompatible traceelement characteristics of Iceland’s end‐membercompositions, from neovolcanic zone depletedtholeiites to the highly enriched flank zone melts,are adequately modeled by peridotite melting. Sec-ondly, presenting new olivine analyses from fourIcelandic eruptions, which cover extremely depletedand enriched compositions, we demonstrate that thelithological variation encompassed by KR4003and KG2 source lithologies is enough to reproducethe measured range of olivine Ni contents; directmelting of an olivine‐free lithology is not required.

4.1. Trace Element Characteristicsof End‐Member Melts

[40] The same basalts selected from Figure 3 todefine the end‐member major element character-istics, were used here to constrain the trace elementcomposition of the end‐member melts. Overall, thetrace element patterns of the neovolcanic zone

enriched end‐member and flank zone end‐memberare very similar (Figure 7). Both zones show strongREE and incompatible element enrichment, andvery similar compatible element abundances. Theprimary difference between the two end‐members isin their incompatible element concentrations, withthe flank zone melts systematically offset to higherabundances. That flank and neovolcanic zoneenriched basalts exhibit very similar trace elementpatterns, but with different absolute concentrations,indicates that degree of melting is the most impor-tant control on the chemical variation betweenbasalts from the two settings. However Pb, forwhich the concentration in the flank zone meltsoverlaps with the concentrations in the neovolcaniczone enriched melts, indicates the role of a sec-ondary process [e.g., Hofmann, 2003]. The REEsmake evident the importance of residual garnetduring melting, generating the steep REE profiles ofthe enriched and flank end‐member. While thelikelihood of there being residual garnet in a purelyperidotitic MORB source has been questioned[Hirschmann and Stolper, 1996], both higher Tp

beneath Iceland and the trace amounts of water in thesource [Nichols et al., 2002; Asimow and Langmuir,2003], should allow even a depleted mantle tobegin melting by 3 GPa.

4.2. Inversion of Trace ElementDistributions for Peridotite Melting

[41] To quantitatively model the trace elementobservations we applied the inversion code ofMcKenzie and O’Nions [1991], modified by Whiteet al. [1992] and Gibson and Geist [2010]. Thefailure or success of this model in matching the

Figure 6. Spider diagrams recording the fit of mixes of Px‐1 melts and experimental peridotite melt compositions tothe enriched Icelandic end‐member. The five best fitting compositions from Table 4 are shown as black broken lines,with other permutations of Px‐1; peridotite melt mixes in grey for comparison. The mixed melt compositions havebeen normalized to the enriched end‐member melt.


14 of 32

trace element patterns, should indicate whether thetrace elements require the involvement of a sourceother than peridotite. The inversion runs a forwardmodel, predicting the REE concentrations of anaggregated polybaric fractional peridotite melt,calculates the root mean square misfit betweenpredicted and observed REE concentrations, andthen uses Powell’s algorithm in adjusting the meltfraction with depth curve to minimize this misfit[McKenzie and O’Nions, 1991; Press et al., 1992,pp. 757–1152]. As well as minimizing misfit withrespect to the REEs, model output also includes thenon‐REE trace elements, allowing for the model’ssuccess in reproducing the data to be tested across a

broad range of compatible‐incompatible elements.Initial source compositions were selected to beprimitive mantle (bulk earth equivalent) for mod-eling the enriched neovolcanic zone and flank end‐members, and a depleted mantle composition formodeling of the depleted end‐member [McKenzieand O’Nions, 1991].

[42] The results of the inversion are shown inFigure 7. The model is successful in matching theconcentrations of all the REE and also many of theother incompatible and compatible elements con-sidered. Even where the model result deviates fromexact matches to the concentrations, it generallyremains close to the pattern of elemental abundance.

Figure 7. Elemental abundances in the mean Icelandic enriched neovolcanic zone and flank zone end‐member meltcompositions. Concentrations have been normalized to the composition of the depleted end‐member melt, which isthe horizontal line at unity in each diagram. One sigma variations about the mean for enriched and flank melts areindicated by vertical bars, and by a grey field for the depleted end‐member composition. Rare earth elements fromthe neovolcanic and flank zones have been separately inverted for, using the peridotite melting and inversion codeof McKenzie and O’Nions [1991], with the model’s predictions of incompatible and compatible element abun-dances also shown. The success of the model in matching the data illustrates that the trace element patterns ofIcelandic basalts do not require pyroxenite in the source. Percent uncertainties in the source compositions used in theinversion are given above each plot.


15 of 32

Given the combined uncertainty in the end‐membercompositions, the starting source compositions andin the partition coefficients, the model result is inmost instances within uncertainty of the data.

[43] The only point of significant deviation betweenthe inverse model and the data is in Pb con-centrations. While model predictions of Pb contentare within uncertainty of the enriched end‐membercomposition, they are predicted to be seven timeshigher in the flank zone end‐member than isobserved, Figure 7. This is perhaps evidence forthe nature of the process generating the enrichedend‐member. The KG1/KG2 sources were referti-lized from a KLB‐1 starting composition by addi-tion of average MORB. As noted by previousauthors [e.g., Zindler et al., 1982; Vidal, 1992;Kokfelt et al., 2006], recycled MORB material islikely to have an increased U/Pb ratio due to pref-erential Pb loss during the dehydration reactionsexperienced on subduction. It is probable therefore,that the relatively low Pb concentration of flankzone melts reflects the addition of low Pb MORBmaterial to the mantle beneath Iceland. This mate-rial is now assimilated into the peridotite producinga fertile peridotitic lithology, enriched in all but themost fluid mobile incompatible elements. Such aninterpretation might also explain the slight, butconsistent, offset of other fluid mobile elements inFigure 7 (e.g., Cs, Rb, K and Ba) to lower con-centrations than that predicted by the model.

[44] Despite the slight differences between themodel result and the observed trace element pat-terns, the overriding feature of Figure 7 is thequality of fit a peridotite melting model providesto the trace element data. While this cannot provethe presence of an exclusively peridotitic sourcebeneath Iceland, it at least indicates its viability asa source lithology for the range of melt composi-tions observed.

4.3. Ni Content of Olivine as a Proxyfor Source

[45] To understand what constraints the trace ele-ment content of olivine places on the sourcelithologies beneath Iceland, we analyzed olivinesfrom four eruptions, the locations of which aremarked in Figure 1. The complete data set ofelectron microprobe analyses of the olivines anddetails of the analytical conditions can be foundin the auxiliary material.1 Two of the eruptions,

Háleyjabunga and the Theistareykir picrite, arerepresentative of the depleted Icelandic end‐member, both having very depleted trace elementratios (Nb/Zr = 0.036 and 0.032, respectively), iso-topic values and overlapping olivine compositions(Figure 8). The remaining two eruptions, Stapafelland Gaesafjöll, more closely resemble the com-position of enriched Icelandic end‐member, eachhaving enriched incompatible element (Nb/Zr =0.157 and 0.114, respectively) and isotopic sig-natures. The main difference between olivinesfrom Gaesafjöll and Stapafell eruptions is in for-sterite(Fo) content, with Gaesafjöll olivines offsetfrom the Stapafell olivines to lower Fo values(Fo = 83–84 for Gaesafjöll compared with mostlyFo = 84–87 for Stapafell). At a given Fo contentolivines from Gaesafjöll therefore have a higherNi content than olivines from Stapafell, by up to500 ppm.

[46] Using the melts from KR4003 and KG2 asstarting compositions, closest to the depleted andenriched end‐member, respectively (Table 2), wenext model the Fo and Ni contents of olivinesproduced during fractional crystallization. Thedetails of the calculations can be found in the captionto Figure 8. The path of olivine compositions crys-tallizing from the KR4003 melt (under olivine onlyfractionation), marked as a thick line in Figure 8,generates an array through the Theistareykir picriteand Háleyjabunga olivine compositions. To modelthe olivines forming from enriched parental melts,we take a partial melt from KG2 as an initialcomposition and model both olivine only andolivine + plagioclase + clinopyroxene fractionation(thin and dashed lines, respectively, Figure 8). TheNi contents of Gaesafjöll olivines can mostly beaccounted for by olivine crystallization from a KG2parental melt. However, allowing the removal of agabbroic crystal assemblage after the liquid hasreached 9 wt% MgO explains the slight spreadof data to higher Ni at a given Fo than is generatedby olivine only crystallization. The Stapafell oliv-ine compositions lie at intermediate Ni contentsbetween the array of olivines from a pure KG2 meltcrystallizing olivine only, and the KR4003 meltolivine array.

[47] It is worth emphasizing that there are largeuncertainties in both source Ni compositions, andwith the partition coefficients assumed during ini-tial melting and low‐pressure fractionation steps.Therefore, while the role of gabbro fractionation ingenerating higher Ni contents of cocrystallizingolivines is examined, the Gaesafjöll data fall within

1Auxiliary materials are available at ftp://ftp.agu.org/apend/gc/2011gc003748.


16 of 32

the combined uncertainty from the errors on all ofthe model parameters.

[48] From the results discussed above and pre-sented in Figure 8, it seems likely that peridotiticmantle sources can produce all the variability in Nicontent observed in Icelandic olivines. This inter-pretation differs from that of Sobolev et al. [2007],which requires olivine‐free lithologies to be presentin order to match the observations. However, ourmodel still places an importance on MORB in thesource region, as peridotites can only producemelts crystallizing high‐Ni olivines if they havebeen refertilized by MORB addition. Olivine Nicontents therefore do require major element het-erogeneity in the source. Whether this occurs in themelt region beneath Iceland by melting and reac-tion of MORB pyroxenites with surrounding peri-

dotites, or takes place over long timescales in thesolid state, we cannot constrain.

5. Isotopic Constraints on SourceLithology

[49] In Figures 9c–9f whole‐rock Nd isotopiccomposition is plotted against either CaO, SiO2,FeO or Al2O3 for high MgO neovolcanic and flankzone basalts. The relationship between enrichmentand major element composition is as identifiedfrom the incompatible trace element relationshipsof Figure 3: enriched (low �Nd) basalts have higherFeO, but lower CaO and SiO2 than the depletedbasalts. For FeO in particular (Figure 9e), a mixtureof material lying between HIMU‐ and MORB‐likecompositions explains the first order distribution of

Figure 8. Ni content of olivines from four Icelandic eruptions. The lines are modelled olivine compositions pro-duced by fractionally crystallizing KG2 or KR4003 parental melts. The melts chosen for modeling were those withthe best fit to Icelandic end‐member melt compositions from Table 2. Primary Ni contents of partial melts of KR4003and KG2 were calculated following the same approach as Sobolev et al. [2007], in which the bulk solid composition,phase proportions and partition coefficients are used to solve for melt Ni content. The starting Ni composition ofKR4003 was taken from Walter [1998]. KG2 Ni content was calculated knowing the 2:1 proportion in which KLB‐1and average MORB were mixed, using the Ni content of KLB‐1 from Hirose and Kushiro [1993] and assumingaverage MORB contains 0.025 wt% NiO. Melts then had equilibrium olivine fractionally extracted from themaccording to the Fe‐Mg and Ni partitioning model of Beattie [1993]. A second model was run allowing olivine,clinopyroxene and plagioclase to crystallize in gabbroic proportions once the melt had reached 9 wt% MgO, the majorelement compositions of plagioclase and clinopyroxene were held fixed and taken as average Theistareykir pheno-cryst values. The paths followed by equilibrium olivine compositions crystallizing from these melts are drawn inblack, with boxes indicating the assumed bulk Ni composition of the solid mantle source producing these primarymelts. A representative 3s error bar on the olivine analyses is included. Data from Sobolev et al. [2007] are shown forcomparison.


17 of 32

Figure 9. (a and b) The Pb‐Nd‐Sr isotopic systematics of Icelandic basalts. (c–f) The major element compositions ofhigh MgO basalts (MgO > 9.5 wt% for the neovolcanic zones, >8.5 wt% for flank zone melts) plotted with their cor-responding whole‐rock Nd isotopes. Samples were selected to include only those with joint major and isotopic anal-yses. Vectors point to the global isotopic major element end‐member compositions identified by Jackson andDasgupta [2008], of fractionation uncorrected basalts between 8 and 16 wt% MgO. The mixing of no two end‐member components can fully account for the distribution of Icelandic basalts.


18 of 32

the data. However, considering also CaO, SiO2

and Al2O3, a more complicated mix of HIMU andEM‐like end‐members would be required toreproduce the enriched end of the array. Thedepleted end‐member in contrast, at least in termsof major elements and Nd isotopes, is consistentlymatched by a source of similar lithology and iso-topic composition to that producing MORB.

[50] The high FeO and low SiO2 contents of Ice-landic enriched basalts are broadly consistent withwhat Jackson and Dasgupta [2008] identify asHIMU‐like characteristics. However, unlike theHIMU of Jackson and Dasgupta [2008], enrichedIcelandic basalts tend toward low CaO. As the nearsolidus CaO contents of peridotite melts have beenlinked to CO2 content in the source [e.g., Brey,1978; Hirose, 1997; Dasgupta et al., 2007], lowCaO perhaps indicates that the enriched Icelandiccomponent is not especially volatile rich. Thisconclusion would be consistent with our analysisof experimental partial melts, in which volatile‐present conditions generally produced melts withpoor fits to the enriched end‐member composi-tion. Similarly, the absence of CO2 enrichments inthe source could account for why SiO2, althoughdecreasing with enrichment (Figure 9b), does notshow the extreme lows of the HIMU end‐member.An additional possibility is that the enriched end‐member consists of a mix of EM and HIMU‐likecomponents. Certainly, from the isotope systemat-ics alone, HIMU is a poor choice of enriched end‐member. Despite this, the high FeO content ofenriched Icelandic basalts would discount the EM‐like components [Jackson and Dasgupta, 2008]being a large proportion by mass of the source,even if they are able to contribute significantly totrace element and isotopic variation.

[51] For Iceland, none of the end‐member basaltcompositions identified by Jackson and Dasgupta[2008] fully account for the isotopic and majorelement variation observed. This probably asmuch reflects the nonuniqueness of the globalend‐members, as it does demand a complex mixbetween them to account for the distribution ofmajor element isotopic systematics. While the hightime‐integrated U/Pb of Icelandic basalts (as recor-ded by Pb isotopic compositions, Figures 9a and 9b)indicates a role for recycled oceanic crustal material,nothing from the relationships in Figure 9 necessi-tates that component to still be present as a distinctpyroxenitic lithology. In fact, the association ofenrichment with low SiO2 and CaO, but high FeO,points toward that precursor MORB material nowhaving homogenized with adjacent peridotite, pro-

ducing a fertile peridotite lithology much like theKG2 starting material of Kogiso et al. [1998].

6. The Effect of a Fusible Lithologyon the Volume of Melt Generation

[52] In the previous section we demonstrated thatboth enriched and depleted Icelandic basalt can begenerated by melting of peridotite, albeit, in thecase of enriched basalt, a peridotite of highly fertilecomposition (KG2). A number of authors havesuggested that melting anomalies around Iceland,and large igneous provinces more generally [Sobolevet al., 2011], are controlled by compositional varia-tion in the mantle [Foulger and Anderson, 2005].Here we develop a model of bilithological meltingthat allows for assessment of the role of sourceheterogeneity in controlling overall melt produc-tion. In combination with our observational con-straints on the fraction of recycled material, thismodel allows us to evaluate a likely upper boundon excess melt production associated with compo-sitional heterogeneity under the North Atlantic.

[53] The significance of source enrichment for meltproduction depends on how it changes: (1) the soli-dus temperature, and (2) the productivity (dF/dP), ofthe bulk matrix‐enriched streak mixture. As dis-cussed previously, it is likely that enriched litholo-gies will have both lower solidus temperatures andgreater productivities than depleted peridotites, andthus their presence in a source could tend to increasemelt production. However, mantle temperature alsohas a large effect on the volume of melt generatedand its average chemistry [McKenzie and Bickle,1988; Langmuir et al., 1992]. Given the progres-sive rise in Tp (from <1400°C–1500°C) as the centerof the Iceland plume is approached [White et al.,1995], the effect of Tp on melt production will alsoneed to be considered by the melting model.

[54] We develop a two‐lithology melting modelconsisting of a peridotite‐pyroxenite dual source,and vary both the proportion of pyroxenite (�) andthe Tp to predict crustal thicknesses over a range ofconditions. This is coupled with a simple parame-terization of melt SiO2 content for both peridotiteand pyroxenite, so that the effect of increasingpyroxenite fraction on melt composition can becompared with along‐ridge SiO2 trends.

6.1. Two‐Lithology Melting Model

[55] We base our model on the thermodynamic treat-ment of adiabatic decompression fractional melting


19 of 32

by Phipps Morgan [2001]. When coupled withparameterizations of the solidus and liquidus sur-faces for each lithology, as well as a functiondescribing how productivity varies in the meltinginterval, the equations of Phipps Morgan [2001]calculate dT/dP and dF/dP for the two‐lithologymixture. Each increment of melting is calculated byfollowing a reversible adiabatic path, with completemelt extraction. There are in effect four separateregimes that can be passed through by an upwellingparcel of two‐lithology mantle, depending uponthe temperature of the mantle T, with respect to thepyroxenite (superscript px) and peridotite (super-script pd) liquidus and solidus surfaces:

[56] 1. T < Tsoliduspx , T < Tsolidus

pd : Neither lithology ismelting. dT/dP reduces to the equation for the solidmantle adiabat, but with the mean thermal expan-sivity and density of the two‐lithology mixture.

[57] 2. T > Tsoliduspx , T < Tliquidus

px , T < Tsoliduspd :

Pyroxenite only melting regime. dF/dP is the pro-ductivity of the pyroxenite, and heat flows from theperidotite into the pyroxenite to maintain thermalequilibrium. The effect of this is to refrigerate theperidotite, suppressing the onset of its melting, butenhancing the productivity of the pyroxenite.

[58] 3. T > Tsoliduspx , T < Tliquidus

px , T > Tsoliduspd :

Pyroxenite and peridotite are melting. Productivityof the pyroxenite drops as the peridotite ceases tothermally buffer the pyroxenite melting. dF/dP iscontrolled by both the peridotite and pyroxenitemelting reactions.

[59] 4. T > Tliquiduspx , T > Tsolidus

pd : Peridotite‐onlymelting, pyroxenite has been exhausted fromsource. dF/dP and dT/dP reduce to the singlelithology melting equations for the peridotiteparameterization. This scenario occurs when Tp ishigh, or the mass fraction of pyroxenite in theperidotite:pyroxenite mixture is low. As can beseen from Figure 10a, the pyroxenite liquidus isalmost intersected by 1 GPa if its mass fraction inthe source is infinitesimal.

[60] The key assumptions are that thermal equilib-rium is maintained between the two lithologies,and that sources and melts remain chemically iso-lated during and after melting. The former of theseassumptions is discussed by Phipps Morgan [2001];the resulting relations imply that for the rates ofupwelling generated by passive plate spreadingbeneath Iceland (half spreading rate ∼10 mm/yr[DeMets et al., 1994]), material <3 km across shouldbe in thermal equilibrium with its matrix. However,if plume‐driven upwelling is important, the length

scale of thermal equilibrium could be an order ofmagnitude smaller toward the base of the meltregion. The validity of the second assumption, ofchemical isolation between sources and their melts,will depend upon the melt extraction mechanism.Melts of the enriched lithology have potentiallyquite different major element compositions to meltsin equilibrium with the surrounding matrix, evenwhen both are peridotitic (e.g., melts of KG2 andKR4003 [Kogiso et al., 1998; Walter, 1998]). Ifthese enriched melts infiltrate surrounding perido-tite they may react, altering both melt and matrixcompositions [Yaxley and Green, 2000]. However,partial chemical isolation of enriched melts is aconsequence of the channelized melt extractionwhich is believed to occur under ridges [Kelemenet al., 1997]. It is therefore possible that onlysmall fractions of melt ever interact chemicallywith their surrounding matrix.

[61] The parameterizations used in modeling arethose of Katz et al. [2003], which represent a rel-atively high solidus temperature depleted perido-tite, and the G2 pyroxenite parameterization ofPertermann and Hirschmann [2003b], represent-ing a more fusible source. The solidus and liquiduscurves of the two lithologies are shown in Figure 10a.G2, a MORB‐like pyroxenite [Pertermann andHirschmann, 2003b], was chosen as an extremecase of enriched component fertility, having a lowersolidus temperature (∼150°C) and much narrowermelting interval (by >100°C) than the peridotite ofKatz et al. [2003] and of KG1/KG2 [Kogiso et al.,1998]. It therefore represents the upper bound ofwhat source enrichment can achieve in increasingmelt volume, and as such offers the most favorablescenario for those models invoking the presenceof recycled MORB in the mantle source to explainelevated crustal thicknesses. If the actual Icelandicsource contains KG1/KG2‐like enriched peridotite,the potential increase in crustal thicknesses arelikely to be much less [Kogiso et al., 1998]. Modelparameters are listed in Table 5, and the detailsof the parameterizations for calculating the SiO2

contents of melts are given in Appendix D.

6.2. Results of Melting Model

[62] The most important result from the pyroxenite‐peridotite melting calculations is Figure 10b, whichshows that even with a purely pyroxenitic source, atemperature anomaly of >100° is required to matchthe crustal thicknesses observed in central Iceland.While this result indicates that source enrichmentalone cannot provide the necessary melt volume to


20 of 32

match Icelandic crustal thicknesses, Figure 10balso shows that pure peridotite melting at a plau-sible Tp for Iceland of 1500°C [White et al., 1995],cannot produce crustal thicknesses in central Icelandeither. However, a peridotite source can reproduce

melt production under central Iceland, if there is a20 fold increase in upwelling rate at the base ofthe melting region (open diamond, Figure 10), con-sistent with the inferences of Ito et al. [1999],Maclennan et al. [2001b], and Kokfelt et al. [2003].

Figure 10. Modeling of two‐lithology melting to constrain the contributions to crustal thickness and major elementchemistry of having a more fusible component embedded in peridotite. (a) Decompression paths for the end‐membercases of pure peridotite melting, dark blue, parameterized from Katz et al. [2003], and pure pyroxenite melting, red,parameterized from Pertermann and Hirschmann [2003b], for a Tp of 1305°C. Solidus and liquidus locations aremarked: px‐s, pyroxenite solidus, px‐l, pyroxenite liquidus, pd‐s, peridotite solidus, pd‐l, peridotite liquidus.(b) Crustal thicknesses generated by melting a peridotite source containing varying pyroxenite mass fraction (�) at arange of Tp, with the top of the melt region set at 1 GPa. The open diamond represents the crustal thicknessproduced by melting pure peridotite, but incorporating 20× plume‐driven upwelling for the first 5.5% of peridotitemelting. (c) The relative contribution of pyroxenite melting (tc

px) to the overall crustal thickness (tctotal). The dotted

lines record the pyroxenite fraction in the source if it contributes 40% of the melt generation, for a given Tp.(d) Predicted SiO2 of pooled melts from mixed lithology melting, compared with the volume‐averaged SiO2 ofIcelandic rift zone basalt. The silica content of peridotite partial melts has been parameterized as a function of P and F,and G2 pyroxenite partial melts as a function of F.

Table 5. Parameters Used in Modeling Two‐Lithology Melting of a Peridotite From Katz et al. [2003], and the G2 PyroxeniteLithology of Pertermann and Hirschmann [2003b]

Parameter Peridotite Pyroxenite Units Referencea

Mass fraction cpx 0.2 ‐Heat capacity 1000 1000 Jkg−1K−1 1Thermal expansivity, solid 40 × 10−6 40 × 10−6 K−1 1Density, solid 3300 3550 kgm−3 1,2Entropy of fusion 300 300 JK−1 1

aReferences: 1, Katz et al. [2003], peridotite and pyroxenite assumed to have the same properties; 2, Pertermann and Hirschmann [2003b].


21 of 32

The high productivity of G2 pyroxenite and itslower solidus temperature, means that meltingpreferentially samples the enriched pyroxenitic partof the mantle source, biasing the average melt com-position with respect to bulk source (Figure 10c);only 5–12% of solid pyroxenite is required beforeits melts are dominant by mass.

[63] The effect on melt SiO2 composition fromincreasing pyroxenite fraction in the source isrecorded in Figure 10d. MORB‐like pyroxenitessuch as G2 melt to produce high SiO2 melts[Pertermann and Hirschmann, 2003b], with thesimple consequence that mean melt SiO2 increasesas the mass fraction of pyroxenite in the sourceincreases. As the peridotite fraction is reduced thereis the added effect of reducing the thermal insula-tion it provides to the pyroxenite, reducing thepyroxenite’s overall degree of melting and furtherincreasing the average SiO2 of the pyroxenitemelts. The primary effect of higher Tp is to increasethe average pressure and extent of melting, whichfor both the pyroxenite and peridotite source lowersthe mean SiO2 content of the melts they produce.Compared to the volume‐averaged SiO2 of Ice-landic basalts (horizontal grey bars, Figure 10d),adding almost any pyroxenite of G2 composition tothe source causes SiO2 to be higher than observed.Peridotite‐only melting, with and without plume‐driven upwelling at the base of the melt region, canmatch the SiO2 contents of Icelandic basalts;although the systematic difference between SiO2 ofthe SNVZ and WVZ basalts requires a differenceof either source, melting dynamics or crustal pro-cessing to be explained.

6.3. Implications of the Two‐Phase MeltingModel for Mantle Lithology

[64] Using the modeling results presented inFigure 10c it is possible to estimate the fraction ofpyroxenite present in a source, provided the frac-tional melt contribution of pyroxenite is known.Sobolev et al. [2008] estimated up to a ∼40%pyroxenite contribution to Icelandic melts. There-fore, taking the Sobolev et al. [2008] estimate as theupper limit of possible pyroxenite contribution toIcelandic melts, gives a maximum ∼10% mass frac-tion of pyroxenite in the Icelandic source (Figure 10c,dotted lines). This is consistent with our ownobservations: as KG2 (33:67 MORB:pyrolite mix-ture) matches the most enriched Icelandic majorelement compositions, no more than 33% recycledmaterial (albeit not as a lithologically distinctcomponent) is required in the source regions.

[65] The above analysis mapped onto Figure 10b(vertical dotted line), constrains the maximumcrustal contribution from pyroxenite derived melts tobe ∼4 km (compare with the crustal thickness (tc)estimates from central Iceland, where tc = 35–40 km[Darbyshire et al., 2000; Allen et al., 2002; Kabanet al., 2002; Li and Detrick, 2006]). The two‐lithology melting model therefore, indicates thatsource enrichment alone is very unlikely to be thecause of excess melt production at Iceland. Withouta temperature anomaly >100°C, a source muchmore fusible than even a pure MORB compositionwould be required to generate 35 km of crustalthickness. It is likely that any source more fertilethan G2 would have wide ranging implications forgeochemical and geophysical observables aroundIceland, of which we see no evidence. A tempera-ture anomaly beneath Iceland of >100°C, shouldtherefore be a requirement of any pyroxenite orperidotite source model. Given our finding of vari-ably enriched peridotite lithologies being able toexplain the geochemical systematics of Icelandicbasalts (i.e., a highly productive MORB‐like lithol-ogy in the source is not an observational require-ment), it is probable that both a temperature anomalyof >100°C and plume‐driven upwelling are requiredto match crustal thicknesses on Iceland.

[66] The along‐ridge compositional trend of Figure 2f,describes a decrease in the mean SiO2 content ofbasalts as Iceland is approached. The two‐lithologymodeling results show that this is the opposite ofwhat would be predicted if the contribution of directmelts of silica saturated (i.e., MORB‐like) pyroxe-nite was increasing toward Iceland (Figure 10d).The major elements alone therefore, refute thepresence of recycled MORB lithologies as a pri-mary source of melts in the Iceland plume(seesection 3.2). Our earlier result however, indicatedthat a mixed peridotite + MORB lithology providedthe closest fit to the Icelandic enriched end‐member,so MORB is likely to have been added to the Ice-landic mantle. Any addedMORBmust have reactedwith adjacent depleted peridotite (in the solid stateor during pyroxenite melting) to produce a fertile,but nonetheless peridotitic, KG2‐like lithology;this reaction permits later generation of melts withsufficiently low SiO2 (and low CaO and highFeO) for Icelandic enriched basalt compositions tobe matched.

7. Summary

[67] We have used a novel method to isolate theeffects of mantle and crustal processes on the major


22 of 32

element systematics of basalts. This has confirmedthe strong relationship between the major and traceelement isotopic compositions of mantle melts.Isotopically and trace element enriched melts onIceland are characterized by high FeO, but lowSiO2 and CaO.

[68] Comparing a large database of experimentalpartial melt compositions to the enriched anddepleted melts from Iceland indicates that a singlesource composition cannot match the range ofmajor element diversity observed. While KLB‐1‐like peridotite lithologies produce melts similar tothe compositions of depleted Icelandic melts,enriched melts are best matched by melting ofKG1/KG2 [Kogiso et al., 1998], a peridotite refer-tilized with MORB material. Major element varia-tion is therefore required in the Icelandic mantlesource.

[69] The enriched source beneath Iceland is morefusible than the source contributing the depletedmelts. Modeling shows that greater fusibility willbias the average melt composition toward that ofthe enriched lithology. The extent of this effect willdepend upon the solidus/liquidus slopes of thedepleted and enriched lithologies melting beneathIceland.

[70] The variation in the mantle source’s majorelement composition and fusibility is present onsmall length scales. Within single zones, eruptionsseparated by only several kilometers exhibit con-trasting major element and olivine Ni contents,characteristic of the depleted and enriched melts.

[71] Both the depleted and enriched sources can beolivine bearing peridotites, with the enrichedsource being derived by mixing ∼40% MORB withKLB‐1 to create a fertile peridotite. This litholog-ical range in the source is consistent with the Nicontent of olivines crystallizing from basalts onIceland. While the major element compositions ofenriched Icelandic melts do not support direct originfrom a pyroxenite, we cannot distinguish betweenpyroxenite melting having infiltrated and referti-lized peridotite, versus a solid state re‐equilibrationbetween MORB and surrounding peridotite.

[72] Oceanic crustal thickness variation in the NorthAtlantic cannot result from changing the sourcecomposition alone. Modeling results from themelting of a bilithologic peridotite‐pyroxenitemantle, combined with geochemical observations,indicate no more than 10% recycled MORB ispresent in the bulk mantle, probably much less.This 10% MORB could only account for a ∼4 km

increase in crustal thickness. Therefore, generationof the >35 km tc observed on Iceland requireselevated mantle potential temperature of at least100°C, and if the source is predominantly perido-titic, also needs plume‐driven upwelling in the baseof the melt region.

Appendix A: Data Sources

[73] For this study data from a large number ofsources was compiled to provide the best availablegeochemical constraint on the composition of theIcelandic mantle. In Table A1 the references fromour database of Icelandic geochemistry are repro-duced, with an indication of what type of data

Table A1. List of Icelandic Data Sources

Reference

Data Types

Majors Traces Isotopic

Brandon et al. [2007] 13 0 0Breddam [2002] 30 5 0Breddam et al. [2000] 0 0 4Chauvel and Hémond [2000] 1 0 3Devey et al. [1994] 24 0 0Eason and Sinton [2009] 10 0 0Elliott et al. [1991] 5 0 6Furman et al. [1991] 0 29 0Gee [1999] 194 0 0Haase et al. [2003] 25 40 0Hardarson and Fitton [1997] 109 10 0Hards et al. [2000] 0 235 0Hémond et al. [1993] 25 0 0Jakobsson et al. [2008] 332 0 0Jónasson [2005] 12 6 0Kempton et al. [2000] 0 0 8Kokfelt et al. [2006] 35 46 9Kuritani et al. [2011] 1 4 1Kurz et al. [1985] 2 0 0Lacasse et al. [2007] 0 24 0Maclennan et al. [2001a] 55 16 0Maclennan et al. [2003a] 108 49 0Macpherson et al. [2005] 5 11 0Mattsson and Oskarsson[2005]

0 10 0

Melson et al. [2002] 6 0 0Meyer et al. [1985] 1 0 0Murton et al. [2002] 176 98 0Nicholson et al. [1991] 30 3 0Peate et al. [2009] 23 26 4Schiellerup [1995] 21 0 0Schilling et al. [1983] 8 0 1Schilling et al. [1999] 18 0 2Sigurdsson et al. [1978] 16 0 0Sigurdsson [1981] 3 0 0Sinton et al. [2005] 258 0 0Skovgaard et al. [2001] 20 15 8Slater et al. [2001] 73 54 22Stracke et al. [2003] 9 11 7Sun et al. [1979] 2 0 0Thirlwall et al. [2004] 0 0 22


23 of 32

was obtained from each. Where possible, analysesof the same sample by different authors havebeen combined to provide a comprehensive set ofanalyses.

[74] Table A2 lists the references from whichexperimental partial melt compositions were usedin finding matches to the Icelandic end‐membercompositions. Experiments which have subse-quently been shown to be in disequilibrium are leftout of the data set. However, even with a broaderinclusion of experimental data, the overall conclu-sions of the paper are not altered. In addition, highdegree melts of basaltic starting compositions

were filtered from the data set. The logic behindthis was that as the end‐member Icelandic com-positions were basaltic, a good match betweenthem and the high degree melts of basalts wouldlikely follow. Yet such a match would not placeparticularly meaningful constraints on the lithologyof the source.

Appendix B: Can Crystal AccumulationGenerate “Enriched” MeltCharacteristics?

[75] To demonstrate that crystal accumulation isunlikely to be controlling the inferred end‐membermajor element compositions, we have modeled theeffect of crystal accumulation on a starting magmawith a Nb/Zr similar to the enriched end‐membercomposition. The results of adding different pro-portions of olivine, clinopyroxene and plagioclaseto the high Nb/Zr starting composition are pre-sented in Figure B1. None of the permutations ofan accumulated crystal cargo are able to move thestarting liquid toward the major element char-acteristics of the enriched (or depleted) melt. Fromthis result we infer the variation of major elementcompositions in Figures 3 and 4 to reflect mantlemelting and/or source processes, and not resultingfrom accumulation.

Appendix C: Relation of AccumulatedPolybaric Fractional Melts to IsobaricBatch Melts: Nature and Experiment

[76] The identification of lithological heterogene-ity, by systematic comparison of natural meltcompositions to an experimental database of partialmelts, raises the question of how the differencesin melting style between nature and experimentmay influence the compositional similarity of theirrespective melts. In nature, melt generation beneathridges and oceanic islands is envisaged to proceedby polybaric fractional melting, with melt extractedafter a small threshold porosity has been reached.These melts may then aggregate during transportout of the mantle and storage in the crust, to eruptas accumulated fractional melts. Experiments incontrast, almost exclusively study batchmelting,withany evidence of chemical disequilibrium betweencoexisting phases rendering the experimental runsuspect. This difference in melting style betweenthe experimental and natural melts we are com-paring to, ought to be considered as a possiblereason why melts of a single lithology cannot

Table A2. List of Experimental Data Used in This Study

Reference Lithology a/h/caPressureRange

Dasgupta et al. [2004] ecl c 3.0–4.1Dasgupta et al. [2005] ecl c 3.0Dasgupta et al. [2006] ecl c 1.0Dasgupta et al. [2007] lrz c 3.0Falloon et al. [1999] lrz a 1.0–1.5Falloon et al. [2001] lrz a 1.0Gerbode and Dasgupta [2010] ecl c 2.9Hammouda [2003] ecl ch 5.0–6.5Hirose [1997] lrz c 3.0Kawamoto and Holloway[1997]

lrz h 5.0

Keshav et al. [2004] px a 2.0–2.5Kinzler [1997] px a 1.5–2.3Kogiso et al. [1998] lrz a 1.5–3.0Kogiso and Hirschmann[2001]

px a 1.0

Kogiso et al. [2003] px a 1.0–5.0Kogiso and Hirschmann[2006]

ecl a 3.0–5.0

Laporte et al. [2004] hrz a 1.0Parman and Grove [2004] hrz a 1.5–2.0Pertermann and Hirschmann[2003a]

ecl a 2.0–3.0

Pickering‐Witter and Johnston[2000]

lrz a 1.0

Robinson et al. [1998] lrz a 1.5Schwab and Johnston [2001] lrz a 1.0Sobolev et al. [2007] px a 3.5Spandler et al. [2008] ecl a 3.0–5.0Takahashi [1986] lrz a 0.0–5.0Takahashi et al. [1993] lrz a 0.0–6.5Takahashi and Katsuji [2002] ecl a 2.0–2.7Tsuruta and Takahashi [1998] px a 0.0–6.0Villager et al. [2004] lrz a 1.0Walter [1998] lrz a 3.0–6.0Wang and Takahashi [1999] ecl a 5.0Wasylenki et al. [2003] lrz a 1.0Yasuda et al. [1994] ecl a 3.0–5.0Yaxley and Green [2000] ecl a 3.5Yaxley and Brey [2004] ecl c 2.5–5.0Yaxley and Sobolev [2007] ecl a 3.5–4.5

aAbbreviations: a, anhydrous; h, hydrous; c, carbonated.


24 of 32

match the range of melt compositions observed onIceland. Specifically, we need to reject the possi-bility that polybaric fractional melts from a singlelithology could be best matched by isobaric batchmelts of another lithology, simply because thedifference in melting style makes isobaric batchmelts of the same lithology a worse match.

[77] To test whether major element heterogeneity inthe source is going to be misidentified because ofthe difference between experimental and naturalmelting, we have used pMELTS [Ghiorso andSack, 1995; Asimow and Ghiorso, 1998; Ghiorsoet al., 2002] to calculate accumulated fractionalmelts, and produce a corresponding data set ofisobaric batch melts, for KR4003 and KG2 startingcompositions [Walter, 1998; Kogiso et al., 1998].These lithologies were chosen because they melt toproduce basaltic compositions similar to the Ice-landic enriched (KG2) and depleted (KR4003) end‐members. The melting path of KR4003 and KG2during decompression melting, and further detailsof the melting conditions are given in Figure C1.

The accumulated fractional melts are then com-pared over their pressure interval of melting to allthe isobaric batch melts calculated, to find the bestfitting batch melt to the accumulated melt compo-sition, and therefore an inferred source composi-tion. The misfit routine used is the same as insection 3.1. If the best fits to polybaric accumu-lated fractional melts of one lithology came fromisobaric batch melts of a different lithology, then itwould suggest that the difference between experi-mental and natural melting style is too great for adirect comparison between the two to be valid forconstraining source composition.

[78] The results of the misfit comparisons areshown in Figure C2, the quality of fit of the bestfitting isobaric batch melt at each pressure incre-ment of accumulated fractional melting is plotted.For KR4003, Figure C1a, its own isobaric batchmelts most often fit its accumulated fractional meltcomposition, apart from in a 0.1 GPa intervalbelow 1.5 GPa. The result is similar for fractionalmelts of KG2, which most frequently look like

Figure B1. A test of the less compatible/more compatible element versus MgO plots as a means of identifying end‐member melt compositions. Accumulation of olivine ± cpx ± plag in whole rock analyses of magmas will primarilymove the data points to the right on Figures 3 and 4. To demonstrate that the observed major element systematics athigh MgO are mostly controlled by source and melting process, the graphs above illustrate the effect of olivine only,olivine + clinopyroxene, and olivine + clinopyroxene + plagioclase addition to a starting melt composition chosen tohave a similar Nb/Zr to the enriched Icelandic end‐member. Typical Krafla clinopyroxene and plagioclase compo-sitions were used, and olivine compositions were calculated using the model of Herzberg and O’Hara [2002].Crystals were added in 0.1 wt% increments until the magma contained ∼12 wt% MgO, or the composition deviatedmarkedly from the enriched end‐member chemistry.


25 of 32

Figure C1. Melting paths for KR4003 and KG2 [Walter, 1998; Kogiso et al., 1998] calculated using pMELTS[Ghiorso and Sack, 1995; Asimow and Ghiorso, 1998; Ghiorso et al., 2002]. The solutions were initiated at 4 GPa and1500°C. The presence of Na2O and K2O in the bulk composition means that the stable solution always includes aliquid phase, producing the tail to high pressures. The higher alkali content of KG2 compared with KR4003 isresponsible for the former’s high‐pressure melt tail being at a relatively high melt fraction (∼7%).

Figure C2. Calculated isobaric batch melts of KR4003 and KG2 compared with calculated accumulated polybaricfractional melts of (a) KR4003 and (b) KG2. Accumulated fractional melts of KR4003/KG2 have been selected atdifferent pressures to the top of the melting column (Ptop) and compared to a synthetic data set, in which the samelithologies have been melted isobarically at a range of pressures and temperatures (0.5–4 GPa and 1300–1650°C).The misfit calculations were performed as described in section 3.1. The quality of fit (= [Mmax − Mi]/Mmax) of thebest fitting isobaric batch melt to each of the accumulated fractional melts is plotted as a horizontal bar, which rangesbetween 0 and 1; Mmax is the maximum misfit of all the records found to fit the accumulated melt compositions, andMi the misfit of the particular best fitting batch melt at a given pressure. The bars are colored according to whichlithology has produced the best fitting isobaric batch melt. All melting calculations were performed using pMELTS[Ghiorso and Sack, 1995; Asimow and Ghiorso, 1998; Ghiorso et al., 2002].


26 of 32

batch melts of KG2, rather than KR4003 batchmelts. The quality of fit to each of the individualoxides is recorded in Figure C3; in each plot thebest fitting isobaric batch melt of the specifiedlithology is compared with the composition of theaccumulated fractional melt it is trying to fit. Com-paring Figures C3a and C3b, shows that KR4003batch melts reliably fit KR4003 fractional meltsacross CFMAS, but that KG2 melts tend to show ahigher FeO content than KR4003 fractional melts.The equivalent comparison between Figures C3cand C3d for fractional melts of KG2, demonstratesthat in most instances KR4003 batch melts are apoor fit to the KG2 fractional melts; they system-atically over estimate the MgO content of the meltand to a lesser extent the Al2O3 content. KG2 batchmelts in contrast, provide an overall better fit, suf-fering less from the high MgO and Al2O3 of theKR4003 melts.

[79] We can conclude from Figures C2 and C3, thatwe are unlikely to falsely identify major element

heterogeneity in the source because of the differ-ences between experimental melting (isobaric batch),and natural melt generation (polybaric fractional).Both KR4003 and KG2 produce isobaric andfractional melts more similar to themselves than tomelts of the other lithology. However, this resultdoes not readily generalize to other lithologies;KR4003 and KG2 systematically provide a good fitto the depleted and enriched end‐member Icelandicmelts, respectively, and so demonstrating that batchmelts of one cannot look like the fractional meltsof the other, is the burden of proof required inIceland’s case for there to be convincing majorelement heterogeneity in the source. In other set-tings, with different lithologies appearing to pro-vide closer matches to the melt compositions, it isnot a given that isobaric and fractional melts oflithologies will appear so compositionally alike.

[80] This result is heavily model dependent, and asimproved thermodynamic models of mantle melt-ing are produced the distinction between isobaric

Figure C3. A comparison of the best fitting KR4003/KG2 isobaric batch melts to (a and b) the accumulated frac-tional melts of KR4003 and (c and d) the accumulated fractional melts of KG2. Cbatch is the oxide concentration in thebatch melt. For plotting Cbatch has been normalized to the corresponding oxide concentration in the accumulated frac-tional melt (Cfrac) it is being compared to. Lines are colored according to the mean pressure of melting (Pfrac) of theaccumulated fractional melt being compared to.


27 of 32

batch and polybaric fractional melting should berevisited, and thus the relation between experi-mental and natural melts made more certain.

Appendix D: Estimating SiO2 Contentof Peridotite and G2 Melts

[81] To estimate the SiO2 content of melts pro-duced from peridotite and pyroxenite melting, weparameterized the experimental results of fourperidotite melting experiments and the G2 meltingexperiment of Pertermann and Hirschmann [2003a].The pressure (in GPa), melt fraction (F) and SiO2

data of Walter [1998], Wasylenki et al. [2003],Falloon et al. [2001], and Robinson et al. [1998]were used to constrain a surface describing SiO2

concentrations in peridotite melts between 1 and7 GPa and F = 0–1: this gave the function,

SiO2½ �pd ¼ 51:13 �0:04ð Þ � 1:88 �0:77ð ÞPþ Fpd 2:769 �3:5ð Þ½� 0:14 �0:77ð ÞP� þ 0:15 �0:11ð ÞP2: ðD1Þ

SiO2 in peridotite melts therefore increases withincreasing melt fraction, and for most of the pressureinterval considered decreases with increasing pres-sure. 95% Confidence ranges for each of the fittedparameters are given in parentheses. For the G2pyroxenite, the limited pressure interval over whichmelting experiments were performed means aparameterization of melt composition as a functionof pressure is not possible. However, melts from G2do show a decrease in SiO2 with increasing meltfraction, and it is this relationship that we haveincluded to predict the SiO2 content of pyroxenitemelts. The equation for SiO2

px obtained from a linearregression of the G2 data is,

SiO2½ �px¼ 56:19 �1:4ð Þ � 7:31 �2:7ð ÞFpx: ðD2Þ

Unlike the peridotite melt systematics, the G2pyroxenite melts show a decrease in SiO2 as the meltfraction increases. Equations (D1) and (D2) werethen integrated over the melt region using equation14 of McKenzie and Bickle [1988]. All integrationswere stopped at 1 GPa, which is approximately thepressure at the base of the crust beneath centralIceland.

Acknowledgments

[82] The authors gratefully acknowledge the work of AnnetteJezierska in collecting the olivine analyses used in this paper.Constructive reviews from Godfrey Fitton and one anonymousreviewer are gratefully acknowledged, as are comments on anearly draft from Sally Gibson and David Neave. OS was sup-ported by NERC grant NE/H2449/4.

References

Allen, R. M., et al. (2002), Plume‐driven plumbing and crustalformation in Iceland, J. Geophys. Res., 107(B8), 2163,doi:10.1029/2001JB000584.

Asimow, P. D., and M. S. Ghiorso (1998), Algorithmic mod-ifications extending MELTS to calculate subsolidus phaserelations, Am. Mineral., 83, 1127–1131.

Asimow, P. D., and C. H. Langmuir (2003), The importance ofwater to oceanic mantle melting regimes,Nature, 421, 815–820.

Beattie, P. (1993), Olivine‐melt and orthopyroxene‐melt equi-libria, Contrib. Mineral. Petrol., 115, 103–111.

Brandon, A. D., D. W. Graham, T. Waight, and B. Gautason(2007), 186Os and 187Os enrichments and high 3He/4Hesources in the Earth’s mantle: Evidence from Icelandicpicrites, Geochim. Cosmochim. Acta, 71, 4570–4591.

Breddam, K. (2002), Kistufell: Primitive melt from the Icelandmantle plume, J. Petrol., 43(2), 345–373.

Breddam, K., M. D. Kurz, and M. Storey (2000), Mapping outthe conduit of the Iceland mantle plume with helium iso-topes, Earth Planet. Sci. Lett., 176, 45–55.

Brey, G. (1978), Origin of olivine melilitites: Chemical andexperimental constraints, J. Volcanol. Geotherm. Res., 3,61–88.

Chauvel, C., and C. Hémond (2000), Melting of a completesection of recycled oceanic crust: Trace element and Pb iso-topic evidence from Iceland,Geochem.Geophys.Geosyst., 1(2),1001, doi:10.1029/1999GC000002.

Darbyshire, F. A., R. S. White, and K. F. Priestley (2000),Structure of the crust and uppermost mantle of Iceland froma combined seismic and gravity study, Earth Planet. Sci.Lett., 181, 409–428.

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

Dasgupta, R., M. M. Hirschmann, and N. Dellas (2005), Theeffect of bulk composition on the solidus of carbonated eclo-gite from partial melting experiments at 3 GPa, Contrib.Mineral. Petrol., 149, 288–305, doi:10.1007/s00410-004-0649-0.

Dasgupta, R., M. M. Hirschmann, and K. Stalker (2006),Immiscible transition form carbonate‐rich to silicate‐richmelts in the 3 GPa melting interval of eclogite + CO2 andgenesis of silica‐undersaturated ocean island lavas, J. Petrol.,57(4), 647–671.

Dasgupta, R., M. M. Hirschmann, and N. D. Smith (2007),Partial melting experiments of peridotite + CO2 at 3 GPaand genesis of alkalic ocean island basalts, J. Petrol., 48(11),2093–2124, doi:10.1093/petrology/egm053.

Dasgupta, R., M. G. Jackson, and C.‐T. A. Lee (2010), Majorelement chemistry of ocean island basalts–Conditions ofmantle melt heterogeneity of mantle source, Earth Planet.Sci. Lett., 289, 377–392.

Davis, F. A., M. M. Hirschmann, and M. Humayun (2009),The composition of low‐degree partial melts of garnet peri-dotite at 3 GPa by modified iterative sandwich experiments(MISE), Eos Trans. AGU, 90(52), Fall Meet. Suppl., AbstractV31F‐04.

Davis, F. A., M. M. Hirschmann, and M. Humayun (2011),The composition of the incipient partial melt of garnet peri-dotite at 3 GPa and the origin of OIB, Earth Planet. Sci.Lett., 308, 380–390, doi:10.1016/j.epsl.2011.06.008.

DeMets, C., R. G. Gordon, D. F. Argus, and S. Stein (1994),Effect of recent revisions to the geomagnetic reversal time


28 of 32

scale on estimates of current plate motions, Geophys. Res.Lett., 21(20), 2191–2194.

Devey, C. W., C. D. Garbeshonberg, P. Stoffers, C. Chauvel,and D. F. Mertz (1994), Geochemical effects of dynamicmelting beneath ridges: Reconciling major and trace‐elementvariations in Kolbeinsey (and global) mid‐ocean ridge basalt,J. Geophys. Res., 99(B5), 9077–9095.

Eason, D. E., and J. M. Sinton (2009), Lava shields and fissureeruptions of the Western Volcanic Zone, Iceland: Evidencefor magma chambers and crustal interaction, J. Volcanol.Geotherm. Res., 186, 331–348.

Elliott, T. R., C. J. Hawkesworth, and K. Gronvöld (1991),Dynamicmelting of the Iceland plume,Nature, 351, 201–206.

Falloon, T. J., D. H. Green, L. V. Danyushevsky, and U. H.Faul (1999), Peridotite melting at 1.0 and 1.5 GPa: An exper-imental evaluation of techniques using diamond aggregatedand mineral mixes for determination of near‐solidus melts,J. Petrol., 40(9), 1343–1375.

Falloon, T. J., L. V. Danyushevsky, and D. H. Green (2001),Peridotite melting at 1 GPa: Reversal experiments on partialmelt compositions produced by peridotite‐basalt sandwichexperiments, J. Petrol., 42(12), 2363–2390.

Fitton, G. J., A. D. Saunders, M. J. Norry, B. S. Hardarson,and R. N. Taylor (1997), Thermal and chemical structureof the Iceland plume, Earth Planet. Sci. Lett., 153, 197–208.

Fitton, G. J., A. D. Saunders, P. D. Kempton, and B. S.Hardarson (2003), Does depleted mantle form an intrinsic partof the iceland plume?, Geochem. Geophys. Geosyst., 4(3),1032, doi:10.1029/2002GC000424.

Foulger, G. R., and D. L. Anderson (2005), A cool model forthe Iceland hotspot, J. Volcanol. Geotherm. Res., 141, 1–22.

Furman, P., F. Frey, and K. Park (1991), Chemical constraintson the petrogenesis of mildly alkaline lavas from vestman-naeyjar, Iceland: The eldfell (1973) and surtsey (1963–1967)eruptions, Contrib. Mineral. Petrol., 109, 19–37.

Gaetani, G. A., and E. B. Watson (2000), Open system behav-iour of olivine‐hosted melt inclusions, Earth Planet. Sci.Lett., 183, 27–41.

Gee, M. A. M. (1999), Volcanology and geochemistry of theReykjanes Peninsula: Plume‐mid‐ocean ridge interaction,Ph.D. thesis, Univ. of London, London.

Gerbode, C., and R. Dasgupta (2010), Carbonate‐fluxed melt-ing of MORB‐like pyroxenite at 2.9 GPa and genesis ofHIMU ocean island basalts, J. Petrol., 51(10), 2067–2088,doi:10.1093/petrology/egq049.

Ghiorso, M. S., and R. O. Sack (1995), Chemical mass trasferin magmatic processes IV. A revised and internally consis-tent thermodynamic model for the interpolation and extrapo-lation of liquid‐solid equilibria in magmatic systems atelevated temperatures and pressures, Contrib. Mineral. Petrol.,119, 197–212.

Ghiorso, M. S., M. M. Hirschmann, P. W. Reiners, and V. C.Kress III (2002), The pMELTS: A revision of MELTSfor improved calculation of phase relations and major elementpartitioning related to partial melting of the mantle to 3 GPa,Geochem. Geophys. Geosyst., 3(5), 1030, doi:10.1029/2001GC000217.

Gibson, S. A., and D. Geist (2010), Geochemical and geophys-ical estimates of lithospheric thickness variation beneathGalápagos, Earth Planet. Sci. Lett., 300 , 275–286,doi:10.1016/j.epsl.2010.10.002.

Grove, T. L., R. J. Kinzler, and W. B. Bryan (1992), Fraction-ation of mid‐ocean ridge basalt (MORB), in Mantle Flowand Melt Generation at Mid‐Ocean Ridges, Geophys.Monogr. Ser., vol. 71, edited by J. P.Morgan, D.K. Blackman,

and J. M. Sinton, pp. 281–310, AGU, Washington, D. C.,doi:10.1029/GM071p0281.

Haase, K. M., C. W. Devey, and M. Wieneke (2003), Mag-matic processes and mantle heterogeneity beneath theslow‐spreading northern Kolbeinsey ridge segment, northatlantic, Contrib. Mineral. Petrol., 144, 428–448.

Hammouda, T. (2003), High‐pressure melting of carbonatedeclogite and experimental constraints on carbon recyclingand storage in the mantle, Earth Planet. Sci. Lett., 214,357–368.

Hardarson, B. S., and J. G. Fitton (1997), Mechanisms ofcrustal accretion in Iceland, Geology, 25, 1043–1046.

Hards, V., P. Kempton, R. Thompson, and P. B. Greenwood(2000), The magmatic evolution of the Snæfell volcaniccentre; An example of volcanism during incipient riftingin Iceland, J. Volcanol. Geotherm. Res., 99, 97–121.

Hart, S. R. (1971), K, Rb, Cs, Sr and Ba contents and Srisotope ratios of ocean floor basalts, Philos. Trans. R. Soc.London A, 268(1192), 573–587.

Hauri, E. H. (1996), Major‐element variability in the Hawaiianmantle plume, Nature, 382, 415–419.

Hémond, C., N. Arndt, U. Litchenstein, A. Hofmann,N. Oskarsson, and S. Steinthorsson (1993), The heteroge-neous Iceland plume: Nd‐Sr‐O isotopes and trace elementconstraints, J. Geophys. Res., 98, 15,833–15,850.

Herzberg, C. (2006), Petrology and thermal structure of theHawaiian plume from Mauna Kea volcano, Nature, 444,605–609.

Herzberg, C. (2011), Identification of source lithology in theHawaiian and Canary Islands: Implications for origins,J. Petrol., 52(1), 113–146.

Herzberg, C., and M. J. O’Hara (2002), Plume‐associatedultramafic magmas of Phanerozoic age, J. Petrol., 43(10),1857–1883.

Hirose, K. (1997), Partial melt compositions of carbonatedperidotite at 3 GPa and role of CO2 in alkali‐basalt magmageneration, Geophys. Res. Lett., 24(22), 2837–2840.

Hirose, K., and I. Kushiro (1993), Partial melting of dry peri-dotites at high pressures: Determination of compositions ofmelts segregated from peridotite using aggregates of dia-mond, Earth Planet. Sci. Lett., 114, 477–489.

Hirschmann, M. M. (2000), Mantle solidus: Experimental con-straints and effects of peridotite composition, Geochem.Geophys. Geosyst., 1, 1042, doi:10.1029/2000GC000070.

Hirschmann, M. M., and E. M. Stolper (1996), A possible rolefor garnet pyroxenite in the origin of the “garnet signature”in MORB, Contrib. Mineral. Petrol., 124, 185–208.

Hofmann, A. W. (2003), Sampling mantle heterogeneitythrough oceanic basalts: Isotopes and trace elements, inTreatise on Geochemistry: The Mantle and Core, edited byR. W. Carlson, H. D. Holland, and K. K. Turekian, chap.2.03, pp. 61–101, Elsevier, New York.

Hofmann, A. W., and W. M. White (1982), Mantle plumesfrom ancient oceanic crust, Earth Planet. Sci. Lett., 57,421–436.

Ita, J., and L. Stixrude (1992), Petrology, elasticity, and com-position of the mantle transition zone, J. Geophys. Res.,97(B5), 6849–6866.

Ito, G., Y. Shen, G. Hirth, and C. J. Wolfe (1999), Mantleflow, melting, and dehydration of the Iceland mantle plume,Earth Planet. Sci. Lett., 165, 81–96.

Jackson, M. G., and R. Dasgupta (2008), Compositions ofHIMU, EM1 and EM2 from global trends between radio-genic isotopes and major elements in ocean island basalts,Earth Planet. Sci. Lett., 276, 175–186.


29 of 32

Jakobsson, S., K. Jónasson, and I. Sigurdsson (2008), Thethree igneous rock series of Iceland, Jökull, 58, 117–138.

Jónasson, K. (2005), Magmatic evolution of the Heiðarsporðurridge, NE‐Iceland, J. Volcanol. Geotherm. Res., 147,109–124.

Kaban, M. K., O. G. Flóvenz, and G. Pálmason (2002), Natureof the crust‐mantle transition zone and the thermal state ofthe upper mantle beneath Iceland from gravity modelling,Geophys. J. Int., 149, 281–299.

Katz, R. F., M. Spiegelman, and C. H. Langmuir (2003), Anew parametrization of hydrous mantle melting, Geochem.Geophys. Geosyst., 4(9), 1073, doi:10.1029/2002GC000433.

Kawamoto, T., and J. R. Holloway (1997), Melting tempera-ture and partial melt chemistry of H2O‐saturated mantle peri-dotite to 11 GPa, Science, 276, 241–243.

Kelemen, P. B., G. Hirth, N. Shimizu, M. Spiegelman, andH. J. B. Dick (1997), A review of melt migration processesin the adiabatically upwelling mantle beneath oceanicspreading ridges, Philos. Trans. R. Soc. London A, 355,283–318.

Kempton, P. D., J. G. Fitton, A. D. Saunders, G. M. Nowell,R. N. Taylor, B. S. Hardarson, and G. Pearson (2000), TheIceland plume in space and time: A Sr‐Nd‐Pb‐Hf study ofthe North Atlantic rifted margin, Earth Planet. Sci. Lett.,177, 255–271.

Keshav, S., G. H. Gudfinnsson, G. Sen, and Y. Fei (2004),High‐pressure melting experiments on garnet clinopyroxe-nite and the alkalic to tholeiitic transition in ocean‐islandbasalts, Earth Planet. Sci. Lett., 223, 365–379.

Kinzler, R. J. (1997), Melting of mantle peridotite at pressuresapproaching the spinel to garnet transition: Application tomid‐ocean ridge basalt petrogenesis, J. Geophys. Res., 102(B1),853–874.

Klein, E. M., and C. H. Langmuir (1987), Global correlationsof ocean ridge basalt chemistry with axial depth and crustalthickness, J. Geophys. Res., 92(B8), 8089–8115.

Kogiso, T., and M. M. Hirschmann (2001), Experimentalstudy of clinopyroxenite partial melting and the origin ofultra‐calcic melt inclusions, Contrib. Mineral. Petrol., 142,347–360, doi:10.1007/s004100100295.

Kogiso, T., and M. M. Hirschmann (2006), Partial meltingexperiments of bimineralic eclogite and the role of recycledmafic oceanic crust in the genesis of ocean island basalts,Earth Planet. Sci. Lett., 249, 188–199, doi:10.1016/j.epsl.2006.07.016.

Kogiso, T., K. Hirose, and E. Takahashi (1998), Meltingexperiments on homogeneous mixtures of peridotite andbasalt: Application to the genesis of ocean island basalts,Earth Planet. Sci. Lett., 162, 45–61.

Kogiso, T., M. M. Hirschmann, and D. J. Frost (2003), High‐pressure partial melting of garnet pyroxenite: Possible maficlithologies in the source of ocean island basalts, EarthPlanet. Sci. Lett., 216, 603–617, doi:10.1016/S0012-821X(03)00538-7.

Kokfelt, T. F., K. Hoernle, and F. Hauff (2003), Upwelling andmelting of the Iceland plume from radial variation of238U‐230Th disequilibria in postglacial volcanic rocks, EarthPlanet. Sci. Lett., 214, 167–186.

Kokfelt, T., K. Hoernle, F. Hauff, J. Fiebig, R. Werner, andD. Garbe‐Schönberg (2006), Combined trace element andPb‐Nd‐Sr‐O isotope evidence for recycled oceanic crust(upper and lower) in the Iceland mantle plume, J. Petrol.,47(9), 1705–1749.

Kuritani, T., T. Yokoyama, H. Kitagawa, K. Kobayashi, andE. Nakamura (2011), Geochemical evolution of historical

lavas from Askja volcano, Iceland: Implications for mechan-isms and timescales of magmatic differentiation, Geochim.at Cosmochim. Acta, 75, 570–587.

Kurz, M. D., P. S. Meyer, and H. Sigurdsson (1985), Heliumisotopic systematics within the neovolcanic zones of Iceland,Earth Planet. Sci. Lett., 74, 291–305.

Lacasse, C., H. Sigurdsson, S. N. Carey, H. Jóhannesson, L. E.Thomas, and N. W. Rogers (2007), Bimodal volcanism atthe Katla subglacial caldera, Iceland: Insight into the geo-chemistry and petrogenesis of rhyoliticmagmas,Bull. Volcanol.,69, 373–399, doi:10.1007/s00445-006-0082-5.

Langmuir, C. H., G. N. Hanson, and M. J. O’Hara (1980), Anevaluation of major element heterogeneity in the mantlesources of basalts, Philos. Trans. R. Soc. London A, 297(1431),383–407.

Langmuir, C. H., E. M. Klein, and T. Plank (1992), Petrolog-ical systematics of mid‐ocean ridge basalts: Constraints onmelt generation beneath ocean ridges, in Mantle Flow andMelt Generation at Mid‐Ocean Ridges, Geophys. Monogr.Ser., vol. 71, edited by J. P. Morgan, D. K. Blackman, andJ. M. Sinton, pp. 183–280, AGU, Washington, D. C.,doi:10.1029/GM071p0183.

Laporte, D., M. J. Toplis, M. Seyler, and J.‐L. Devidal (2004),A new experimental technique for extracting liquids fromperidotite at very low degrees of melting: Application to par-tial melting of depleted peridotite, Contrib. Mineral. Petrol.,146, 463–484, doi:10.1007/s00410-003-0509-3.

Li, A., and R. S. Detrick (2006), Seismic structure of Icelandfrom Rayleigh wave inversions and geodynamic implica-tions, Earth Planet. Sci. Lett., 241, 901–912, doi:10.1016/j.epsl.2005.10.031.

Maclennan, J. (2008a), Lead isotope variability in olivine‐hosted melt inclusions from Iceland, Geochim. Cosmochim.Acta, 72, 4159–4176.

Maclennan, J. (2008b), Concurrent mixing and cooling ofmelts under Iceland, J. Petrol., 49(11), 1931–1953,doi:10.1093/petrology/egn052.

Maclennan, J., D. McKenzie, K. Grönvold, and L. Slater(2001a), Crustal accretion under northern Iceland, EarthPlanet. Sci. Lett., 191, 295–310.

Maclennan, J., D. McKenzie, and K. Grönvold (2001b),Plume‐driven upwelling under central Iceland, Earth Planet.Sci. Lett., 194, 67–82.

Maclennan, J., D. McKenzie, F. Hilton, K. Gronvöld, andN. Shimizu (2003a), Geochemical variability in a single flowfrom northern Iceland, J. Geophys. Res., 108(B1), 2007,doi:10.1029/2000JB000142.

Macpherson, C. G., D. R. Hilton, J. M. D. Day, D. Lowry, andK. Gronvöld (2005), High‐3He/4He, depleted mantle andlow‐d18O, recycled oceanic lithosphere in the source of centralIceland magmatism, Earth Planet. Sci. Lett., 233, 411–427.

Mattsson, H., and N. Oskarsson (2005), Petrogenesis of alka-line basalts at the tip of a propagating rift: Evidence fromthe heimaey volcanic centre, south Iceland, J. Volcanol.Geotherm. Res., 147, 245–267.

McKenzie, D., and M. J. Bickle (1988), The volume and com-position of melt generated by extension of the lithosphere,J. Petrol., 29, 625–679.

McKenzie, D., and R. K. O’Nions (1991), Partial melt distribu-tions from inversion of rare earth element concentrations,J. Petrol., 32(5), 1021–1091.

McKenzie, D., A. Stracke, J. Blichert‐Toft, F. Albaréde,K. Grönvold, and R. K. O’Nions (2004), Source enrichmentprocesses responsible for isotopic anomalies in oceanic islandbasalts, Geochim. Cosmochim. Acta, 68(12), 2699–2724.


30 of 32

Melson, W. G., T. O’Hearn, and E. Jarosewich (2002), A databrief on the smithsonian abyssal volcanic glass data file,Geochem. Geophys. Geosyst., 3(4), 1023, doi:10.1029/2001GC000249.

Meyer, P., H. Sigurdsson, and J. Schilling (1985), Petrologicaland geochemical variations along Iceland’s neovolcaniczones, Geophys. Res., 90(B12), 223–226.

Murton, B. J., R. N. Taylor, and M. N. Thirlwall (2002),Plume‐ridge interaction: A geochemical perspective fromthe Reykjanes Ridge, J. Petrol., 43(11), 1987–2012.

Nichols, A. R. L., M. R. Carroll, and A. Höskuldsson (2002),Is the Iceland hot spot also wet? Evidence from the watercontents of undegassed submarine and subglacial pillowbasalts, Earth Planet. Sci. Lett., 202, 77–87.

Nicholson, H., and D. Latin (1992), Olivine tholeiites fromKrafla, Iceland: Evidence for variations in melt fractionwithin a plume, J. Petrol., 33(5), 1105–1124.

Nicholson, H., M. Condomines, J. G. Fitton, A. E. Fallick,K. Gronvöld, and G. Rogers (1991), Geochemical andisotopic evidence for crustal assimilation beneath krafla,Iceland, J. Petrol., 32(5), 1005–1020.

Niu, Y., K. D. Collerson, R. Batiza, J. I. Wendt, andM. Regelous(2008), Origin of enriched‐type mid‐ocean ridge basalt atridges far from mantle plumes: The East Pacific Rise at11°20′N, J. Petrol., 49(4), 633–664.

Parman, S. W., and T. L. Grove (2004), Harzburgite meltingwith and without H2O: Experimental data and predictivemodeling, J. Geophys. Res., 109, B02201, doi:10.1029/2003JB002566.

Peate, D.W., J. A. Baker, S. P. Jakobsson, T. E.Waight, A. J. R.Kent, N. V. Grassineau, and A. C. Skovgaard (2009), Historicmagmatism on the Reykjanes Peninsula, Iceland: A snap‐shotof melt generation at a ridge segment, Contrib. Mineral.Petrol., 157(3), 359–382, doi:10.1007/s00410-008-0339-4.

Peate, D. W., K. Breddam, J. A. Baker, M. D. Kurz, A. K.Barker, T. Prestvik, N. Grassineau, and A. C. Skovgaard(2010), Compositional characteristics and spatial distributionof enriched Icelandic mantle components, J. Petrol., 51(7),1447–1475, doi:10.1093/petrology/egq025.

Pertermann, M., and M. M. Hirschmann (2003a), Anhydrouspartial melting experiments on morb‐like eclogite: Phaserelations, phase compositions and mineral‐melt partitioningof major elements at 2–3 GPa, J. Petrol., 44(12), 2173–2201.

Pertermann, M., and M. M. Hirschmann (2003b), Partial melt-ing experiments on a MORB‐like pyroxenite between 2 and3 GPa: Constraints on the presence of pyroxenite in thebasalt source regions from solidus location and melting rate,J. Geophys. Res., 108(B2), 2125, doi:10.1029/2000JB000118.

Phipps Morgan, J. (2001), Thermodynamics of pressurerelease melting of a veined plum pudding mantle, Geochem.Geophys. Geosyst., 2(4), 1001, doi:10.1029/2000GC000049.

Pickering‐Witter, J., and A. D. Johnston (2000), The effects ofvariable bulk composition on the melting systematics of fer-tile peridotitic assemblages, Contrib. Mineral. Petrol., 140,190–211.

Press, W. H., S. A. Teukolsky, W. T. Vetterling, and B. P.Flannery (1992), Numerical Recipes in FORTRAN: The Artof Scientific Computing, 2nd Ed., Cambridge Univ. Press,Cambridge, U. K.

Prytulak, J., and T. Elliot (2007), Tio2 enrichment in oceanisland basalts, Earth Planet. Sci. Lett., 263, 388–403,doi:10.1016/j.epsl.2007.09.015.

Putirka, K., F. J. Ryerson, M. Perfit, and W. I. Ridley (2011),Mineralogy and composition of the oceanic mantle, J. Petrol.,52(2), 279–313, doi:10.1093/petrology/egq080.

Ringwood, A. E. (1962), A model for the upper mantle,J. Geophys. Res., 67(2), 857–867.

Ritsema, J., W. Xu, L. Stixrude, and C. Lithgow‐Bertelloni(2009a), Estimates of the transition zone temperature in amechanically mixed upper mantle, Earth Planet. Sci. Lett.,277, 244–252.

Ritsema, J., P. Cupillard, B. Tauzin, W. Xu, L. Stixrude, andC. Lithgow‐Bertelloni (2009b), Joint mineral physics andseismic wave traveltime analysis of upper mantle tempera-ture, Geology, 37(4), 363–366.

Robinson, J. A. C., B. J. Wook, and J. D. Blundy (1998), Thebeginning of melting fertile and depleted peridotite at1.5 GPa, Earth Planet. Sci. Lett., 155, 97–111.

Schiellerup, H. (1995), Generation and equilibration of olivinetholeiites in the northern rift zone of Iceland. A petrogeneticstudy of Bláfjall table mountain, J. Volcanol. Geotherm.Res., 65, 161–179.

Schilling, J. G. (1973), Iceland mantle plume: Geochemicalstudy of reykjanes ridge, Nature, 242, 565–571.

Schilling, J. G., M. Zajac, R. Evans, T. Johnston, W. White,J. D. Devine, and R. Kingsley (1983), Petrology and geo-chemical variations along the Mid‐Atlantic Ridge from29°N to 73°N, Am. J. Sci., 283, 510–586.

Schilling, J. G., R. Kingsley, D. Fontignie, R. Poreda, andS. Xue (1999), Dispersion of the jan mayen and Icelandmantle plumes in the arctic: A He‐Pb‐Nd‐Sr isotopetracer study of basalts from the Kolbeinsey, mohns, andknipovich ridges, J. Geophys. Res., 104(B5), 10,543–10,569.

Schwab, B. E., and A. D. Johnston (2001), Melting systemat-ics of modally variable, compositionally intermediate perido-tites and the effects of mineral fertility, J. Petrol., 42(10),1789–1811.

Shorttle, O., J. Maclennan, and S. M. Jones (2010), Control ofthe symmetry of plume ridge interaction by spreading‐ridgegeometry, Geochem. Geophys. Geosyst., 11, Q0AC05,doi:10.1029/2009GC002986.

Sigurdsson, H. (1981), First‐order major element variation inbasalt glasses from the Mid‐Atlantic Ridge: 29°N to 73°N,J. Geophys. Res., 86(B10), 9483–9502.

Sigurdsson, H., J. Schilling, and P. Meyer (1978), Skagi andlangjokull volcanic zones in Iceland: 1. Petrology and struc-ture, J. Geophys. Res., 83(B8), 3983–4002.

Sinton, J., K. Grönvold, and K. Sæmundsson (2005), Postgla-cial eruptive history of the Western Volcanic Zone, Iceland,Geochem. Geophys. Geosyst., 6, Q12009, doi:10.1029/2005GC001021.

Skovgaard, A. C., M. Storey, J. Baker, J. Blusztajn, and S. R.Hart (2001), Osmium‐oxygen isotopic evidence for arecycled and strongly depleted component in the Icelandmantle plume, Earth Planet. Sci. Lett., 194, 259–275.

Slater, L., D. McKenzie, K. Gronvöld, and N. Shimizu (2001),Melt generation and movement beneath Theistareykir, NEIceland, J. Petrol., 42(2), 321–354.

Sobolev, A. V., and N. Shimizu (1993), Ultra‐depleted pri-mary melt included in an olivine from the Mid‐AtlanticRidge, Nature, 363, 151–154.

Sobolev, A. V., A. W. Hofmann, S. V. Sobolev, and I. K.Nikogosian (2005), An olivine‐free mantle source of Hawai-ian shield basalts, Nature, 434, 412–417.

Sobolev, A. V., et al. (2007), The amount of recycled crust insources of mantle derived melts, Science, 316, 590–597.

Sobolev, A. V., A. W. Hofmann, G. Brügmann, V. G.Batanova, and D. V. Kuzmin (2008), A quantitative linkbetween recycling and osmium isotopes, Science, 321(5888),536.


31 of 32

Sobolev, S. V., A.V. Sobolev, D.V.Kuzmin,N.A.Krivolutskaya,A. G. Petrunin, N. T. Arndt, V. A. Radko, and Y. R. Vasiliev(2011), Linking mantle plumes and, large igneous provincesand environmental catastrophes, Nature, 477, 312–316,doi:10.1038/nature10385.

Spandler, C., G. Yaxley, D. H. Green, and A. Rosenthal(2008), Phase relations and melting of anhydrous K‐bearingeclogite from 1200 to 1600°C and 3 to 5 GPa, J. Petrol., 49,771–795, doi:10.1093/petrology/egm039.

Stracke, A., V. J. M. Salters, and K. W. W. Sims (1999),Assessing the presence of garnet‐pyroxenite in the mantlesources of basalts through combined hafnium‐neodymium‐thorium isotope systematics, Geochem. Geophys. Geosyst.,1(1), 1006, doi:10.1029/1999GC000013.

Stracke, A., A. Zindler, V. J.M. Salters, D.McKenzie, J. Blichert‐Toft, F. Albarède, and K. Grönvold (2003), Theistareykirrevisited, Geochem. Geophys. Geosyst. , 4(2), 8507,doi:10.1029/2001GC000201.

Sun, S.‐S., R. W. Nesbitt, and A. Y. Sharaskin (1979), Geo-chemical characteristics of mid‐ocean ridge basalts, EarthPlanet. Sci. Lett., 44, 119–138.

Takahashi, E. (1986), Melting of a dry peridotite KLB‐1 up to14 GPa: Implications on the origin of peridotitic upper man-tle, J. Geophys. Res., 91(B9), 9367–9382.

Takahashi, E., and N. Katsuji (2002), Melting process in theHawaiian plume: An experimental study, in Hawaiian Vol-canoes: Deep Underwater Perspectives, Geophys. Monogr.Ser., vol. 128, edited by E. Takahashi et al., pp. 403–418,AGU, Washington, D. C.

Takahashi, E., T. Shimazaki, Y. Tsuzaki, and H. Yoshida(1993), Melting study of a peridotite KLB‐1 to 6.5 GPa,and the origin of basaltic magmas, Philos. Trans. R. Soc.London A, 342(1663), 105–120.

Thirlwall, M. F., M. A. M. Gee, R. N. Taylor, and B. J. Murton(2004), Mantle components in Iceland and adjacent ridgesinvestigated using double‐spike Pb isotope ratios, Geochim.Cosmochim. Acta, 68(2), 361–386.

Tsuruta, K., and E. Takahashi (1998), Melting study of analkali basalt JB‐1 up to 12.5 GPa: Behaviour of potassiumin the deep mantle, Phys. Earth Planet. Inter., 107, 119–130.

Van der Hilst, R. D., and H. Kárason (1999), Compositional het-erogeneity in the bottom 1000 kilometers of earth’s mantle:Toward a hybrid convectionmodel, Science, 283, 1885–1888.

Vidal, P. (1992), Mantle: More HIMU in the future, Geochim.Cosmochim. Acta, 56, 4295–4299.

Villager, S., P. Ulmer, O. Müntener, and A. B. Thompson(2004), The liquid line of descent of anhydrous, mantle‐derived, tholeiitic liquids by fractional and equilibrium

crystallization–An experimental study at 1.0 GPa, J. Petrol.,45(12), 2369–2388, doi:10.1093/petrology/egh042.

Walter, M. J. (1998), Melting of garnet peridotite and the ori-gin of komatiite and depleted lithosphere, J. Petrol., 39(1),29–60.

Wang, W., and E. Takahashi (1999), Subsolidus and meltingexperiments of a K‐rich basaltic composition to 27 GPa:Implications for the behaviour of potassium in the mantle,Am. Mineral., 84, 357–361.

Wasylenki, L. E., M. B. Baker, A. J. R. Kent, and E. M. Stolper(2003), Near‐solidus melting of the shallow upper mantle:Partial melting experiments on depleted peridotite, J. Petrol.,44(7), 1163–1191.

White, R. S., D. McKenzie, and K. O’Nions (1992), Oceaniccrustal thickness from seismic measurements and rare earthelement inversions, J. Geophys. Res., 97(B13), 19,683–19,715.

White, R. S., J. W. Bown, and J. R. Smallwood (1995), Thetemperature of the Iceland plume and origin of outward‐propagating V‐shaped ridges, J. Geol. Soc. London, 152,1039–1045.

White, W. M., and A. Hofmann (1982), Sr and Nd isotopegeochemistry of oceanic basalts and mantle evolution,Nature, 296, 821–825.

Yasuda, A., T. Fujii, and K. Kurita (1994), Melting phase rela-tions of an anhydrous mid‐ocean ridge basalt from 3 to20 GPa: Implications for the behaviour of subducted oceaniccrust in the mantle, J. Geophys. Res., 99(B5), 9401–9414.

Yaxley, G. M., and G. P. Brey (2004), Phase relations of car-bonate bearing eclogite assemblages from 2.5 to 5.5 GPa:Implications for petrogenesis of carbonatites,Contrib. Mineral.Petrol., 146, 606–619, doi:10.1007/s00410-003-0517-3.

Yaxley, G. M., and D. H. Green (2000), Reactions betweeneclogite and peridotite: mantle refertilisation by subductionof oceanic crust, Contrib. Mineral. Petrol., 139, 326–338.

Yaxley, G. M., and A. V. Sobolev (2007), High‐pressure par-tial melting of gabbro and its role in the Hawaiian magmasource, Contrib. Mineral. Petrol., 154, 371–383, doi:10.1007/s00410-007-0198-4.

Zindler, A., S. R. Hart, and F. A. Frey (1979), Nd and Sr iso-tope ratios and rare earth element abundances in ReykjanesPeninsula basalts evidence for mantle heterogeneity beneathIceland, Earth Planet. Sci. Lett., 45, 249–262.

Zindler, A., E. Jagoutz, and S. Goldstein (1982), Nd, Sr, andPb isotopic systematics in a three‐component mantle: Anew perspective, Nature, 298, 519–523.


32 of 32

Compositional trends of Icelandic basalts: Implications...

Documents

Transcript of Compositional trends of Icelandic basalts: Implications...