| Surface waves propagate along the Earth's surface and are sensitive to the shallow S-wave velocity structure.They are widely used in shear-wave velocity tomography on near surface or crust structure.Currently,the signal used in surface wave tomography usually comes from teleseismic events and ambient noise.Recently,surface wave tomography based on ambient noise method has developed rapidly.High-resolution surface wave imaging results can be obtained from surface wave signals from ambient noise.The surface wave with higher frequency is more sensitive to the shallower structure.This method also has the advantages of safety and cheapness.Because of the relatively dense population and lack of enough and uniformly distributed earthquakes in Anhui section of Tanlu fault zone,it is not suitable to use artificial sources.Then ambient noise method seems to be a more appropriate choice.Noise with higher frequencies(e.g.above 0.5Hz)usually comes from human activities.It is still in need to be further verified on how reliable high-frequency surface wave signals can be extracted by noise correlation in areas where the noise has a great impact and the time required.On the one hand,the traditional surface wave tomographyusing ray theory based on the high frequency hypothesis cannot take into consideration of the influence of the structure off the ray path,which introduce errors to the calculation.Meanwhile,the computational cost of full waveform inversion is very expensive and the hardware requirement is high,which will bring inconvenience in practical use.Therefore,we hope that the finite frequency method which use travel time information can obtain relatively high accuracy and efficiency for the travel time calculation.However,the influence of finite frequency effect on near-surface and shallow crust in small area is still in need to be further validated,and the inversion method based on this theory in small area needs to be further developed for application.To solve the problems mentioned above,we have done some research in the following four aspects:(1)Near-surface 3-D S-wave velocity structure imaging in urban areaHuman activities have a great influence on high frequency-surface wave signals in ambient noise.In order to study the characteristics of high-frequency noise in such areas and verify the accuracy of using ambient noise method to recover near-surface structures,we collected about two weeks of continuous ambient noise data in a 5km*3 km area in Hefei urban area.We firstly obtained cross-correlation functions between arbitrary stations.Then we used the phase weighted stacking method based on S-transform to improve the signal-noise ratio of the cross-correlation functions.Finally,the group velocity and phase velocity dispersion curves in the period of 0.5-2 s are extracted,and the shear wave velocity structure to 400 m depth in the region is inverted using a direct surface wave tomography method based on ray tracing.The results are consistent with the local geological structure,which verifies the possibility of extracting reliable high frequency surface wave signals from ambient noise in urban area.(2)3-D shallow crust S-wave velocity structure imaging with complex structure in a small areaThe Lujiang section of the Tanlu fault zone is located at the junction of Tanlu fault zone,Dabie orogenic belt,Hefei basin and Luzong volcanic basin.Its geological structure is complex.Using the data of 80 stations,the dispersion curves of group velocity and phase velocity in the periods of 0.5-8s are obtained by ambient noise cross-correlation.We use the method of direct inversion of shear wave velocity based on ray tracing to obtain a three-dimensional shear wave velocity structure from the surface to 5km depth.Our results are consistent with the local geological structure,showing the coupling of velocity structure of fault zone,orogenic belt,basin and intrusive magmatic rocks.Our new model can be used to explain the evolution and dynamic mechanism of magmatic intrusive rocks in the fault zone.(3)Comparisons of the accuracy of travel time prediction from 4 different theoriesBecause ray theory is based on the assumption of high frequency approximation,it cannot reflect the influence of velocity anomaly off the ray path on travel time,and the three-dimensional inversion of full waveform will bring a lot of calculation cost.We hope to achieve a balance between computational accuracy and cost by using travel time finite frequency method.Based on the traveltime finite frequency method,we construct an empirical two-dimensional surface wave phase kernel based on ray tracing.In order to verify the applicability and accuracy of different theories in the study of shallow crustal velocity structure,we systematically analyzed the accuracy of travel time difference of surface wave phase velocity obtained from five methods:full waveform,empirical and analytical 2-D surface wave phase kernel based on finite frequency theory,ray tracing under the assumption of off-great circle path and ray theory under the assumption of great circle path.Here we obtained the travel time difference by using the widely used forward methods such as SEM3D and fast marchingmethod.By comparing the results,the accuracy of the empirical two-dimensional surface wave phase kernel is better than that from the ray theory and analytical kernel,and the computational efficiency is better than full waveform theory.The comparison validates the possibility of using finite frequency kernel for inversion(4)Direct inversion method of pseudo-3-D surface wave phase velocity empirical kernel based on ray tracing and its applicationBased on the travel time finite frequency theory of surface wave,we propose a direct inversion method of 3-D S-wave velocity structure using empirical pseudo-3-D surface wave phase kernels.This method allows us to construct a curved surface wave phase velocity kernel due to large structural variation.The real data from Taipei Basin is used to verify the reliability of the method.This method can be further extended in the future.At present,we only use the empirical kernel of surface wave phase velocity.By adding the kernel for group velocity,we can improve the efficiency and quantity of data.In addition,a more self-consistent model with higher resolution can be obtained if finite-frequency kernel of body wave is also used in the inversion. |
PDF Full Text Request