Structure and evolution of the upper mantle of the Australian Plate and North Atlantic Ocean from waveform tomography

Abstract

With the expansion of regional and global seismic networks and the advancing of tomographic techniques, tomography models are able to reveal the heterogeneities in the Earth's interior with an increasing amount of detail. Illuminating the seismic velocity anomalies within the Earth contributes to our understanding of the structure, dynamics and evolution of the Earth. The lack of seismic data that covers the world's oceans, however, is a factor that limits the resolution of tomography models. In this thesis, two new regional upper mantle tomography models of S-wave velocity and azimuthal anisotropy are presented. These tomography models are computed using a global dataset of 1.7 million waveforms. New seismic data was obtained by the deployment of an array of ocean bottom seismometers (OBSs) in the North Atlantic Ocean. The waveforms were successfully fit using the Automated Multimode Inversion (AMI), which fits S-, multiple S- and Rayleigh waves. AMI computes a set of linear independent equations for each seismogram, describing the perturbations of P- and S-waves within their sensitivity kernels between the source and the receiver. These equations are combined in a large linear system and inverted for the 3D global distribution of P- and S-waves and S-wave azimuthal anisotropy in the upper mantle. To avoid errors in the model, the least consistent measurements of the dataset were identified and removed from the global inversion. The two tomography models were each computed using a subset of the global database, discarding the waveforms that have both their source and their receiver outside the hemisphere surrounding the research area. The models were optimized for the region of interest by optimizing the parameter settings for the area and by performing area-specific outlier analysis. A new tomography model of the Australian plate and its boundaries, Aus22, is presented. The model is validated by performing resolution tests and, for specific locations, by computing Rayleigh phase velocities using independent interstation measurements. The model reveals the S-wave and azimuthal anisotropy structure of the upper mantle of the Australian plate and its boundaries. The high velocities below Australia show the presence of thick, cold, cratonic lithosphere below western and central Australia, showing significant lateral heterogeneities. The cratonic lithosphere extends up to the northern boundary of the plate, where it collides with the island arcs, without subducting. The cratonic lithosphere underlays the vast majority of the diamondiferous outcrops. The eastern boundary of the cratonic lithosphere provides a lithospheric definition of the Tasman Line. East of this Tasman Line, the seismic velocities reveal an area of intermediate lithospheric thickness, underlying the Tasmanides. The eastern part of Australia is underlain by low seismic velocities, evidence for a thin, warm lithosphere. All the Cenozoic volcanism present in eastern Australia is underlain by this thin lithosphere. The upper mantle below the Tas-manid and Lord Howe hotspots shows low seismic velocities down to the base of the transition zone, revealing the presence of a deep mantle upwelling feeding these hotspots, which possibly also feeds the East Australia hotspot. High seismic velocities at the plate boundaries show the presence of actively subducting slabs and reveal the presence of slab remnants. In the transition zone below northeast Australia, high seismic velocities indicate the presence of subducted lithospheric fragments, possibly related to the former northern margin of the Australian continent. From Aus22, a new hypothesis is developed of the formation and evolution of the oceanic crust between Australia and Antarctica, including the Australian-Antarctic Discordance (AAD). This theory can explain the seismic structure beneath the area as revealed by Aus22, the anomalously deep residual basement depth that characterises the area and other observations in and around the AAD. High seismic velocities are present in the upper mantle below the ocean between Australia and Antarctica. These seismic anomalies are explained by the presence of remnants of the rheological sublayer, formerly present below the cratonic lithosphere of Australia. The presence of these remnants causes a lack of melt supply at the AAD, both due to a decrease in temperature in the asthenosphere and due to the partial barrier it creates, disturbing the upward heat flow from the deep mantle towards the ridge, resulting in the formation of the chaotic ridge system. The remnants of the rheological sub-layer also cause an acceleration of cooling of the oceanic lithosphere above, creating subsidence of the residual basement of the oceanic crust. A high-velocity anomaly is observed below Zone A, the part of the SEIR located east of the AAD, which is identified as a stranded piece of cratonic lithosphere, delaminated from the southern Galwer craton around 24 Ma. In the shallow upper mantle, low seismic velocities are present east of the AAD, which are linked to lateral flow from the Balleny Plume, supplying additional melt to Zone A. This results in a shallower residual basement depth and an isotopic boundary between Zone A and the AAD. Our hypothesis is tested by targeted resolution tests and petrological modelling, verifying the inferences made. A new tomography model of the North Atlantic Ocean, NA23, is presented. Besides the hemisphere dataset, this tomographic model is computed using new data from a network of OBSs deployed in the eastern North Atlantic Ocean, as part of the SEA-SEIS project. The data is thoroughly preprocessed and the compliance and tilt noise is removed to get the best quality data from the instruments. The tomography model reveals the S-wave velocity and azimuthal anisotropy structure below the North Atlantic Ocean, including the Iceland Plume. Low seismic velocities are present below Iceland and the adjacent Kolbeinsey Ridge and northern Reykjanes Ridge down to 260 km depth, below which they merge into a single low-velocity anomaly west of Iceland. The low-velocity anomaly moves further west to eastern Greenland in the transition zone. This implies that the source of the Iceland plume is located below eastern Greenland and suggests a lateral spread of the plume as it flows towards Mid-Atlantic ridge in the upper mantle. Strong ridge-parallel azimuthal anisotropy in the shallow upper mantle below the Reykjanes Ridge indicates a channelled flow of plume material from Iceland towards the southern end of the Reykjanes Ridge. Any sign of a radially outward flow from the plume in the upper mantle is absent. No low-velocity anomaly is present directly below the Jan Mayen Fault Zone, arguing against a separate plume conduit here. A low-velocity anomaly in the transition zone below the Mohns Ridge could indicate a possible second plume source further north.