Geochemistry of the Early Miocene volcanic succession of Northland, New Zealand, and implications for the evolution of subduction in the Southwest Pacific

Abstract

Latest Oligocene and Early Miocene volcanic rocks occur on the Northland Peninsula, New Zealand, and record the inception of Cenozoic subduction-related volcanism in the North Island that eventually evolved to its present manifestation in the Taupo Volcanic Zone. This NW-striking Northland Arc is continuous with the Reinga Ridge and comprises two parallel belts of volcanic centres ca. 60 km apart. A plethora of tectonic models have been proposed for its origins. We acquired new trace element and Sr–Nd isotope data to better constrain such models. All Northland Arc rocks carry an arc-type trace element signature, however distinct differences exist between rocks of the eastern and western belt. Eastern belt rocks are typically andesites and dacites and have relatively evolved isotope ratios indicating assimilated crustal material, and commonly contain hornblende. Additionally some eastern belt rocks with highly evolved isotope compositions show fractionated REE compositions consistent with residual garnet, and some contain garnetiferous inclusions in addition to schistose crustal fragments. In contrast, western belt rocks are mostly basalts or basaltic andesites with relatively primitive Sr–Nd isotope compositions, do not contain hornblende and show no rare earth element evidence for cryptic amphibole fractionation. Eastern and western belt rocks contain comparable slab-derived fractions of fluid-mobile trace elements and invariably possess an arc signature. Therefore the difference between the belts may be best explained as due to variation in crustal thickness across the Northland Peninsula, where western belt centres erupted onto a thinner crustal section than eastern belt rocks. The consistent arc signature throughout the Northland arc favours an origin in response to an actual, if short-lived subduction event, rather than slab detachment as proposed in some models. No Northland Arc rocks possess a convincing adakite-like composition that might reflect the subduction of very young oceanic lithosphere such as that of the Oligocene South Fiji Basin. Therefore we favour a model in which subduction of old (Cretaceous) lithosphere drove subduction.