| Seismic tomography has been playing a key role in studing the earth interior and its dynamic process.However,its resolution is still limited by data coverage,data type and so on.In order to improve the resolution in seismic tomography,we developed a wavelet based multi-scale seismic tomography method and a joint inversion method using both body and surface waves.Compared to conventional method,wavelet-based method has advantages of being data adaptive,which results in higher resolution model when data coverage is good,while keeps large scale features in regions with sparse data coverage.We have applied the joint inversion to the southern California region and the inverted results show good consistence with local geological structures.What's more,the obtained model outperforms the community velocity model in terms of data fitting.Moreover,we developed a ray-tracing based direct inversion method for surface wave dispersion data in order to make it straigtforward to jointly invert both body and surface waves.Application to Taipei Basin show low wavespeed anomaly in basin region which extends to about 0.5 km.The obtained model,especially the depth of the sediment,can be important for strong ground motion simulations.This thesis has achieved seismic tomography developments in the following 4 as-pects:1)Multi-scale seismic traveltime tomography based on wavelet analysis.The resolution of seismic traveltime tomography relies on events distribution,sta-tions distribution and so on.Usually the distribution of events and stations is far from homogenuous with a lot of events clustering along faults,and stations mainly deployed on the land with human access.In order to obtain a model with resolution varies with data coverage,we propose a wavelet-based seismic tomography method.Compared to traditional methods which is based on damping or smoothing,our method has the ad-vantage of multi-scale imaging with high resolution in regions with good data coverage,while keep large scale features in regions with sparse ray coverage.2)Direct inversion of surface wave dispersion data.Usually the invertion of surface wave dispersion data is divided into two steps.Firstly,constructing 2D phase/group velocity maps from the dispersion data;then in-verting a 1D model at every surface grid point and combining the inverted 1D models into a pseudo 3D model.Here we propose a new inversion strategy to invert disper-sion data directly to 3D velocity model.Compared to conventional two-step inverison method,our method can take full advantage of dispersion data within each period,even there are only a few available.Besides,the off-great-circle effect has been taken into consideration by ray tracing,which proves to be necessary in complex media.What's more,it makes joint inversion of body and surface wave data straightforward due to the similar formulation in direct inversion of surface wave data.3)Joint inversion of body and surface wave data.With the help of direct inversion of surface wave data,it is straightforward to in-coporate surface wave data into body wave inversion framework.Joint inversion can take advantage of the complementary strengths of two data sets and obtain a unified Vp and Vs model.We will compare inversion results from joint inverison and separate inversion using body or surface waves and prove its fidelity using waveform simulation.4)3D Vp/Vs tomography from body and surface wave data.The model of Vp/Vs plays a very important role in geological interpretation due to its high sensitivity to petrology,partial melting,water/gas saturation and so on.How-ever,the raliablity of Vp/Vs model from body wave travel time inversion is often low due to the difference in resolvabilty for P and S waves.The resolving difference comes from different quantity and quality and usually P wave data outweigh S wave data in both aspects.The incorporation of surface wave data makes it possible to invert Vp/Vs since surface wave data are most sensitive to Vs and thus provide a better constraint for Vs. |
PDF Full Text Request