An unusual occurrence of mafic accretionary lapilli in deep-marine volcaniclastics on 'Eua, Tonga: Palaeoenvironment and process

• Palagonitised mafic ash aggregates described from 'Eua, Tonga
• Concentric layers with micro-aggregates survive to identify accretionary lapilli.
• Microfauna in host Miocene volcaniclastics reflects deep marine conditions.
• Host rocks reflect sediment gravity flows with minor hemipelagites.
• Ash aggregates transported to depth by subaerial dispersal and settling

Reports of occurrences of accretionary lapilli on Earth, whether in historic time or in the geological record, are restricted to subaerial environments or to shallow marine environments when faunal evidence exists to determine palaeodepths. The proximity of the deep ocean to subduction zones/island arcs (where moist explosive volcanism conducive to ash aggregate formation is common) makes this surprising. In this paper, accretionary lapilli are reported within Middle Miocene mafic glass-rich volcaniclastics on 'Eua, the island closest to the Tonga Trench, a persistent high in the frontal arc basin. The glass in the accretionary lapilli has been subjected to advanced palagonitisation, but concentric layers marked by micro-aggregates containing shard-shaped particles survive to determine one group of occurrences as layered accretionary lapilli. The palaeoenvironment, as established by pelagic microfauna, is clearly deep marine, not less than 1600 m. The host rocks, typically gravel/sand in grain size, contain sedimentary structures (normal grading to inverse and normal-to-inverse grading, lack of grading, large-scale cross-bedding, slump bedding and sedimentary dykes) suggesting that the full spectrum of sediment gravity flow types, including less ordered debris flows, has been active. In an island arc environment, a range of sediment gravity flow types can be initiated, some by pyroclastic flows entering the sea. However, the thin beds of accretionary lapilli do not exhibit features of sediment gravity flow deposits or those of submarine pyroclastic flows. Possible transport processes must account for the matrix between the ash aggregates, which is either coarse-grained or absent. Modelling of particle descent times to 1600 m through a sea water column provides one explanation for the features displayed.