Item

Abstract

Quaternary dacites and trachytes from the Aird Hills and Lusancay Islands in Papua New Guinea show some of the clearest slab melt geochemical signatures (Mg# 73-93, Sr = 1520-2650 ppm, Sr/Y = 140-445, La/Yb = 135-238), yet there is no slab currently subducting beneath Papua New Guinea. Alternatively, they may be melts from orogenic mafic crustal underplate, yet they do not occur above an arc crustal keel, nor are they part of an active convergent tectonic setting. Instead, they occur at the tip of a propagating rift-tectonic system within rift-related mafic to silicic alkaline magmatic suites. Although some adakites lie up to 100 km off the present rift-front, they connect to a curved line after their relative positions are adjusted for 16° of late Cenozoic rotation that accounts for active oceanic spreading in the Woodlark Rift. The timing of rift-propagation is consistent with the Quaternary age of Papua New Guinea adakites. The geochemical signature of these rocks is similar to other modern adakites (Western Aleutians, Cerro Pampa, Cook Island). Their Mg#, Sr contents, Sr/Y and La/Yb ratios are significantly higher than those of adakitic melts from orogenic mafic underplate. Trace element modeling indicates that their high Sr/Y and La/Yb ratios requires both small partial melting degrees (<5 vol.%) with residual garnet-amphibolite and/or eclogite (30-42 vol.% garnet), and subsequent interaction with metasomatised peridotite. We present a model of rift-propagation over previously subduction-modified mantle. Papua New Guinea was influenced by rapid creation and destruction of oceanic microplates since Mesozoic times. In a new rift-tectonic regime, rift-related magmas encountered and partially melted some remnant eclogite and/or garnet-amphibolite slab fragments, and subsequently interacted geochemically with hydrated and veined peridotite, resulting in adakites in a non-convergent tectonic setting. Other rift melts traveled through this paleo-mantle wedge without encountering any slab fragments, and formed MORB and OIB-type magmatic suites. Support for our model comes from geophysical tomographic imaging which indicates a number of flat-lying, anomalously fast regions interpreted as former subduction zone remnants. Earlier ophiolite studies show (adakite-like) trondhjemitic rocks along oceanic paleo-transform faults. Papua New Guinea is a key area for testing the adakite = slab melts story, because it not only simulates the geochemical, but also the geodynamic context which presumably led to widespread continental crustal growth in the Archean. This challenges existing adakite and Archean crustal growth models which suggest that the generation of adakitic melts are restricted to convergent plate margins.