Integrating deep Earth dynamics in paleogeographic reconstructions of Australia

Abstract

It is well documented that the Cenozoic progressive flooding of Australia, contemporaneous with a eustatic sea level fall, requires a downward tilting of the Australian Plate towards the SE Asian subduction system. Previously, this large-scale, mantle-convection driven dynamic topography effect has been approximated by computing the time-dependent vertical shifts and tilts of a plane, but the observed subsidence and uplift anomalies indicate a more complex interplay between time-dependent mantle convection and plate motion. We combine plate kinematics with a global mantle backward-advection model based on shear-wave mantle tomography, paleogeographic data, eustatic sea level estimates and basin stratigraphy to reconstruct the Australian flooding history for the last 70 Myrs on a continental scale. We compute time-dependent dynamic surface topography and continental inundation of a digital elevation model adjusted for sediment accumulation. Our model reveals two evolving dynamic topography lows, over which the Australian plate has progressively moved. We interpret the southern low to be caused by sinking slab material with an origin along the eastern Gondwana subduction zone in the Cretaceous, whereas the northern low, which first straddles northern Australia in the Oligocene, is mainly attributable to material subducted north and northeast of Australia. Our model accounts for the Paleogene exposure of the Gulf of Carpentaria region at a time when sea level was much higher than today, and explains anomalous Late Tertiary subsidence on Australia's northern, western and southern margins. The resolution of our model, which excludes short-wavelength mantle density anomalies and is restricted to depths larger than 220 km, is not sufficient to model the two well recorded episodes of major transgressions in South Australia in the Eocene and Miocene. However, the overall, long-wavelength spatio-temporal pattern of Australia's inundation record is well captured by combining our modelled dynamic topography with a recent eustatic sea level curve. We suggest that the apparent Late Cenozoic northward tilting of Australia was a stepwise function of South Australia first moving away northwards from the Gondwana subduction-related dynamic topography low in the Oligocene, now found under the Australian–Antarctic Discordance, followed by a drawing down of northern Australia as it overrode a slab burial ground now underlying much of the northern half of Australia, starting in the Miocene. Our model suggests that today's geography of Australia is strongly dependent on mantle forces. Without mantle convection, which draws Australia down by up to 300 m, nearly all of Australia's continental shelves would be exposed. We conclude that dissecting the interplay between eustasy and mantle-driven dynamic topography is critical for understanding hinterland uplift, basin subsidence, the formation and destruction of shallow epeiric seas and their facies distribution, but also for the evolution of petroleum systems.