Search for a command to run...
Seismic tomography is one of the most important means to image the internal structure of the Earth. Restricted array teleseismic tomography determines the 3D-distribution of velocity perturbations relative to a given background model on a regional scale, even down into the lower mantle (> 660 km depth). This enables the imaging of ongoing tectonic and geodynamic processes e.g. at subduction zones, hot spots and active volcanoes. The resulting models allow estimations of variations of the density, the elastic constants and the temperature.The objective of this thesis is the development of a new method of regional tomography (JI-3D) to obtain models of the subsurface structure with high spacial and mathematical resolution at the same time, even at large depth. The method is based on optimizing the inversion problem itself, in order to avoid non- or underconstrained model parameters. The aim is to obtain stable and more unique inversion results, with high resolution, both in a spacial and a mathematical sense. This is realized by a variable parameterization with many model parameters where the ray distribution is good, while in areas with poor sampling a model parameter rather represents a stable average over a larger area. A second important part of the method is 3D raytracing, providing realistic raypaths, so that the resulting velocity perturbations are projected to the correct locations in the model space. Since the near surface (~50 km depth) resolution of regional tomography is strongly affected by the missing ray crossings in that depth range, additional information is applied to constrain the inversion result. These additional constraints are derived from classical a priori information about the model and from the use of an additional data set (Bouguer gravity data) which is jointly inverted together with the teleseismic delaytimes.The method is applied to a synthetic data set and to real data from the German Eifel region. In the latter case, a low velocity anomaly is found, extending from beneath the lithosphere to the transition zone, which is interpreted as a mantle plume. Since this mantle plume seams to penetrate a major mantle discontinuity in 410 km depth, the method is used to examine the influence of deflected mantle discontinuities on the data and the corresponding inversion results.