| With this thesis we attempt a comprehensive examination of the locations of midoceanic earthquakes based on the available high-resolution bathymetry maps. Our focus is on the use of independent geological constraints, the plate tectonics theory, and advanced multi-event relocation techniques to improve the locations of these events that are usually severely biased by poor station coverage and noise in teleseismic observations. The concentration of these events in the southern hemisphere thus is critical for three-dimensional (3D) tomography images of the Earth. After outlining in Chapter 2 the main concepts of earthquake relocation method and 3D tomography technique, we relocate about 1,500 globally distributed master events in Chapter 3 using arrival times from the International Seismological Centre (ISC) Bulletin and the high-resolution bathymetry maps. The ISC locations can be as far as 70 km away from appropriate topographic features. Our new locations support that large interplate earthquakes occur on or very close to the primary bathymetric features (published in Pan et al., 2002a). In Chapter 4, we compare the improved locations with those obtained by Autonomous Underwater Hydrophone (AUH) arrays (location errors are on the order of 4 km). This exercise illustrates that relocated events agree with AUH locations better than ISC does and our relocation method is valid (being published in Pan and Dziewonski, 2004). Then, we examine three multi-event relocation (cluster analysis) techniques in Chapter 5 on both nuclear test events and mid-oceanic events (from Pan et al., 2004a), before applying the preferred simultaneous Joint Hypercenter Determination procedure with respect to the already relocated master events (JHDME) to 5731 mid-oceanic earthquakes around the world (from Pan et al., 2004b). Fixing the master events along appropriate plate boundaries and the use of JHDME minimize the effect of poor station coverage. The resulting seismicity agrees well with the main active ridges or transforms, and the station corrections reveal the deep "continental roots." We obtain and compare our new 3D P-wave velocity models of the mantle, enhanced by using the improved P-phases in Chapter 6, to other existing ones. Even though the results are preliminary, our comparisons show significant impact of the new dataset for resolving three-dimensional deep Earth structures. |
