Sources of OIB magmatism

According to a popular concept proposed by A. Hofmann and colleagues (e.g. Hofmann and White, 1982; Chauvel et al., 1995), the subduction of metasedimentary and metabasaltic residues after dehydration and partial melting of the oceanic crust would have introduced "metacrustal" material of variable composition into the depleted mantle. As discussed in Chapter 25, both the initial composition of subducting slabs and their chemical development in the course of subduction depend on a number of highly variable parameters; therefore the compositions of subducted rocks introduced into the mantle are also highly variable, as are the mixing proportions of these and the ambient mantle materials. For example, hydrophilic and incompatible Pb is readily removed from the subducting slab and transferred into the continental crust by arc magmas (Fig. 25.1), thus providing enhanced U/Pb ratios in the residual slab material. If such a slab were stored in the mantle as an isolated domain, it would accumulate highly radiogenic Pb and could become a source of ocean-island basalts showing the HIMU signature (Fig. 27.3(a), (b)). Highly incompatible and moderately hydrophilic Rb is also underabundant in HIMU rocks (Fig. 24.9), leading to their low 87Sr/86Sr ratios (Fig. 27.4). The 207Pb/206Pb age, ~ 1.8 Gyr for the HIMU source (see the previous section) is then interpreted as the approximate time interval between the subduction of high-material and OIB magmatism. This would place limits on the rate of mantle convective stirring.

Other OIB end-members could be partially sourced by a subducted metased-imentary component. Thus Eisele et al. (2002) interpreted the isotopic (Os, Pb, Hf, Nd, Sr systematics) and trace-element data obtained for Pitcairn basalts, which define the EM I mantle domain, in terms of multicomponent mixing. These authors identified subducted oceanic-crust residue as the major mixing component but also emphasized the contribution of metasediments, both pelagic and terrigenous. Identification of the sedimentary material involved high time-integrated Th/U ratios (up to 14; these are also recorded by the 208Pb/206Pb ratios, Fig. 27.3(a), (b)) and somewhat enhanced Lu/Hf values (recorded by the 176Hf/177Hf ratios). The latter feature is typical of pelagic metasediments that inherited high Lu/Hf ratios from ocean water (Fig. 27.13).

From a study of the trace-element and Sr-Nd-Pb-He-Os-O isotope abundances in lavas of the Samoan hot spot, Workman et al. (2004) concluded that the major chemical and isotopic features of Samoan volcanic rocks could have resulted from the mixing of several mantle materials (enriched, depleted and high-3He). From their mixing model these authors derived a composition for the EM II end-member, i.e. the ancient metasomatized oceanic lithosphere transferred into the deep mantle via subduction or delamination.

Small-volume continental and oceanic alkali basalts (e.g. in the Cameroon volcanic chain; Fitton and Dunlop, 1985) can be strongly similar in their chemical and isotopic compositions. Based on this, McKenzie and O'Nions (1995) presented a quantitative model for their formation. In a first magmatic episode, melts which originated in the DMM via a small degree of melting (well below 0.01) and which are therefore highly enriched in incompatible elements (Eqns 12.4, 12.6)

penetrate into the subcontinental mantle lithosphere, thus producing metasoma-tized (refertilized) domains. Such domains would age, and their rock units would acquire enriched time-integrated isotopic signatures, set apart from the conservative host reservoir, on a reasonably short time scale, — 1 Gyr. The authors noted that such an enrichment could hardly be achieved within a subducted slab because of insufficient fractionation of the parent-daughter elements in slab rocks. Later, a second magmatic event, i.e. partial melting within a given domain, generates continental alkali magmas, whereas delamination of the domain, followed by melting, gives rise to ocean-island and seamount volcanoes. Along with delamination of the mantle lithosphere, delamination of the lower crust could also introduce isotopic heterogeneities into the mantle.

Recently McKenzie et al. (2004) developed a new model in which compositions related to a plume source can be distinguished from those that originated in the course of formation of plume melts. correlations between isotope ratios (unaffected during partial melting or fractionation) and elemental concentrations (depending on both source composition(s) and fractionation) are used to reveal isotopic and elemental heterogeneities in the source region. Solution of the model is possible only if extensive isotopic and trace-element data sets exist for a restricted region and if it is assumed that the plume-source geochemistry was modified in a single event. The model was applied to comprehensive data sets for Theistareykir (Iceland), Kilauea (Hawaii) and the Pitcairn hot spot (Pacific Ocean). The principal outcome from the modelling is the cannibalistic recycling of OIB materials: ancient ocean-island basalts, subducted at some time in the past (400 Myr for the Theistareykir source and 1.2 Gyr for Kilauea) and stored in the mantle appear to be the most suitable materials to generate the presently observed OIBs. There are two surprising features of this solution. The first is the preservation of the subducted OIB without its being affected by the subduction process or mantle convection. The other feature is the absence in these test cases of any obvious evidence of melts from the sediments that are expected to be subducted along with the OIB materials and whose traces emerge in Figs. 27.2 and 27.4.

Thus a number of subduction- and delamination-related scenarios would result in fits to the data. While they have not been firmly established, the most important general observation is that these isotopic-component signatures cannot have been generated by melting and melt differentiation in the mantle alone, as they depart from the geochemical and isotope correlation trends that would be expected in that case. In connection with models of mixing in the mantle, it is also important to realize that the plume source material itself is likely to be heterogeneous ("mingled" rather than mixed) and that the actual mixing probably occurs between melts extracted from the different subsources. This would also help to explain the heterogeneities observed at very small scales.

0 0

Post a comment