Young boninitic lavas from the northern Tonga fore-arc unrelated to subduction initiation

The fore-arc crust at subduction zones is generally thought to be considerably older than the overlying arc, and to have formed during subduction initiation. We present new Ar-40/Ar-39 ages as well as geochemical and petrological data for volcanic rocks from a 500 m-high section along the wall of a rifted basin within the northern Tonga Ridge fore-arc. Compared to lavas from the adjacent Tonga Arc, the samples have relatively high SiO2 and low TiO2 at a given MgO, and the least evolved samples have boninitic bulk-rock compositions. Ar-40/Ar-39 ages of three samples are about 1.3 Ma, which is younger than the initiation of subduction in northern Tonga (possibly 20 Ma). All lavas are depleted in incompatible trace elements but enriched in fluid-mobile elements as a result of slab dehydration beneath the fore-arc. Although lavas from the adjacent arc front at this latitude have been proposed to contain Pb from the subducting Louisville Seamount Chain (LSC), the unradiogenic Pb isotope compositions (Pb-206/Pb-204 < 18.9) of the older fore-arc lavas indicate that they do not contain a contribution from LSC materials. A dike intruding the lavas is relatively enriched in fluid-immobile Nb and Ta and has higher Nb/Zr ratios than N-MORB. The Nd and Hf isotope compositions resemble those of rejuvenated Samoa lavas, and we propose that Samoa mantle flowed into fore-arc mantle after eruption of the 61ROV lavas at about 1.3 Ma. Our results show that not all boninites (including high-Si boninites) in fore-arc locations are related to subduction initiation.