In recent years seismology has been used extensively to detect and locate the small scale (~10 km) structure of the Earth. In the mantle these structures likely represent chemical heterogeneity and are essential in our understanding of mechanical mixing processes within mantle convection. As subducted crust is chemically distinct from the background mantle, imaging the remains of the crust provides a tracer for convectional flow. In this study global and regional seismic heterogeneities in the mantle are found by processing teleseismic earthquake data through array seismology methods. Scattered energy from shallow earthquakes that arrives as PP precursors is studied in a 100 s quiet window before the main PP. Global average stacks of the PP wavefield are formed using data recorded at a global distribution of seismic arrays, for distances 70-120°. The resultant global stacks of PP have revealed that precursors to PP exist for all distances with the amplitudes increasing with distance and time. Regional stacks for the Pacific and Atlantic Oceans are found to contain very similar patterns of PP precursors suggesting scattering observed here does not vary with tectonic region. Global averages of PP precursors are modelled using a Monte Carlo phonon method that generates statistical scattering models for random media. Modelling results show that heterogeneities in the crust and mantle contribute to scattered PP wavefield. The best models found have 1% scattering strength in the crust, with reduced scattering strength of 0.8% in the uppermost mantle and an increase in scattering strength to 1% at ~700 km. Correlation length also varies from 2 km in the crust to 6 km in the mantle. The extent of the deeper mantle layer of heterogeneity is not well resolved and may be determined using larger epicentral distances. Regional patterns of heterogeneity are found from PP precursors that are scattered from small-scale heterogeneities in subduction zones. Array methods are applied to data in the epicentral distance range of 90°-110° from Eielson Array in Alaska, to calculate directivity and to enhance weak arrivals. Coherent precursors are selected automatically based on a semblance weighted beampower spectrum. Assuming single P-to-P scattering and using the directivity information from array processing, the origin of scattering is found by ray-tracing through a 1D velocity model. Most scatterers are imaged in western Pacific subduction zones with evidence for ~300 small-scale heterogeneities in the region around the present day Japan, Izu-Bonin, Mariana and West Philippine subduction zones. Most of the detected scatterers are located in the crust and upper mantle, but 6% are located deeper than 600 km. Scatterers in the transition zone correlate well with edges of fast features in tomographic images and subducted slab contours derived from slab seismicity. Deeper scatterers are located beneath the Izu-Bonin/Mariana subduction zones, which outline a steeply dipping pseudo-planar feature to 1480 km depth, and beneath the ancient (84-144 Ma) Indonesian subduction trench down to 1880 km depth. The cause of scattering is likely the underside reflection of the subducted Moho of subducted crustal material and are related to past and present subduction providing evidence that the subducted crust does descend into the lower mantle at least for these steeply dipping subduction zones. Combining the findings from both global and regional studies, it is likely that the heterogeneities detected in these studies are related to different stages of the mantle mixing cycle. As such a simple model for mid-mantle heterogeneity applicable to subduction zones has been suggested, with a well mixed mid-mantle of 6 km scale heterogeneities that have been thinned through mechanical stirring and a steeply dipping slab with attached crust penetrating to the lower mantle.
| Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:617156 |
| Date | January 2013 |
| Creators | Bentham, Hannah Louise Marie |
| Contributors | Rost, Sebastian |
| Publisher | University of Leeds |
| Source Sets | Ethos UK |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
| Source | http://etheses.whiterose.ac.uk/6913/ |
Page generated in 0.0016 seconds