Origin and distribution of components in boninite genesis: significance of the OIB component

Abstract

Abstract Previous detailed geochemistry of lavas from the Troodos ophiolite, Cyprus, has revealed the existence in the massif of a variety of magma types ranging in composition from island-arc tholeiites (IAT; Group I) to boninites (BSV; Group III). Petrogenetic modelling, in conjunction with field evidence, suggests that the tholeiitic types were probably derived by extensive melting of relatively fertile spinel Iherzolite, in contrast to the boninitic types that seem to require small degrees of melting of variably depleted spinel harzburgite. Melting was hydrous but not H 2 O-saturated, testified by the exclusive presence of anhydrous liquidus phases and low primary water content levels in the lavas. The IAT magmas most probably represent accumulated melt compositions, whereas the BSV magmas may be interpreted as instantaneous mantle melts. Parameterization of the residual composition of the shallow upper mantle generated during isentropic upwelling of MORB-pyrolite beneath mid-ocean ridges has been undertaken and a new model for the modal stratification of variably depleted sub-oceanic lithosphere is presented; this further constrains the depth of melting of potential Troodos sources to c. 10–12 kbar for Group I and c. 3–7 kbar for Groups II and III. The above inferred depths of melting are in agreement with experimentally determined saturation pressures of the liquidus phases observed in the rocks. Consideration of trace-element abundances provides good evidence that all Troodos source regions have been affected by hydrous fluids and/or melts from subducted lithologies as well as by an ‘exotic’ OIB-like component. Additional consideration of selected major elements and isotopes suggests that the distribution of these components in the supra-subduction zone (SSZ) mantle was proportional to depth, in that the deeper sources were affected by larger volumes of component(s). The concept of a variably depleted source ‘sensing’ (i.e. relative) v. ‘seeing’ (i.e. absolute) a component and subsequently bequeathing the acquired chemical signature to the resultant melts is explored at length. It is demonstrated that, for any given amount of component added, the sensing ability of a source becomes significantly greater with increasing degree of depletion. Harzburgites can sense the presence of even negligible masses of enriched component(s), and this is particularly evident in small extent melts from such rocks. The important implications of component-sensing v. -seeing for Troodos petrogenesis are incorporated in a model of ‘cryptic’ mantle metasomatism followed by melting and preferential extraction of the contaminant with the early melt fractions. The methodology to calculate isotopic compositions in melts generated according to the scheme proposed above is also included. It is shown that heterogeneous distribution of a single OIB-like component with near-chondritic ɛ Nd in variably depleted peridotite and subsequent differential partial melting may account for most of the geochemical peculiarities of the lavas. Further calculations show distinctive variations in the distribution of specific components with depth in the SSZ mantle, with the OIB-like component dominating over the subduction fluids at shallow regions. Only minimal amounts of enriched component(s) are necessary to explain the differences in magma chemistry observed, suggesting that the extent of melting in the SSZ mantle was largely controlled by the thermal/compositional regime prevailing during magma genesis. The appreciably larger volumes of enriched component(s) previously invoked to induce melting of the more refractory Troodos peridotite sources are not required by the model and do not agree with the comparable H 2 O contents of primitive lavas from different Groups. A two-stage melt extraction/infiltration process taking place beneath spreading centres is presented as a candidate for the incorporation of the OIB component residing at the sites of boninite magmatism in the sub-oceanic lithosphere.