Research And Application Of Extracting Body Waves And Long-period Love Waves From Ambient Seismic Noise

The development of Ambient Noise Technology(ANT)has greatly improved the imaging capability of the crustal and upper mantle structures.Theoretically,complete Green’s function between a pair of stations can be retrieved from cross-correlation functions(CCFs)of ambient noise.In reality,only surface wave part of Green’s function can be restored in most cases.Rayleigh wave are mainly sensitive to the SV wave velocity and Love wave to SH wave velocity.Radial anisotropy can be obtained through a joint inversion of Rayleigh wave and Love wave.However,to date,long-period Love wave still cannot be retrieved from ambient noise,which makes it difficult to study radial anisotropy of greater depths using ambient noise data.On the other hand,surface wave is insensitive to velocity interfaces.To reduce ambiguities when inverting for velocity structures with interfaces,body wave information needs to be included.However,the retrieval of body wave signals from ambient noise is a big challenge.To expand the application fronts of ANT,this thesis focuses on the retrieval of body wave and long-period Love wave from ambient noise.To extract long-period Love surface wave and body wave from ambient noise data,We developed several new methods in this thesis.We proposed a root-mean-square ratio selection stacking method(RMSR＿SS)to reduce the length of required time series during the signal recovery process of CCFs.We conducted simulation experiments and field data applications,which both confirmed the effectiveness and reliability of our method.Then Using our method,we successfully recovered Love wave signals with a period of 50-150 s from ambient noise data recorded by stations in the United States.After calculating long-period dispersion curves,we analyzed the misfit of phase velocity.To extract body wave signals from ambient noise data,we used a method called equal distance stacking.We conducted a series of synthetic experiments to verify the feasibility of our method.Based on this method,we obtained direct and reflected signals of the body wave from a dense array.Considering that the distribution of noise sources is difficult to meet the conditions for retrievals of body wave Green’s function,we also developed a Common Receiver Station(CRS)stacking method to extract interface information from the CCFs.We successfully obtained reflected signals from Moho based on our method.The main research contents and results of this thesis can be summarized as follows:(1)A new stacking method is developed to recover the CCF efficiently.This method is named as root-mean-square ratio selection stacking(RMSR＿SS).In the RMSR SS method,we first judged if a short-duration CCF constructively contributes to the recovery of EGFs,and then we only stack those CCFs which constructively contribute to the convergence of EGFs.By applying this method to synthetic noise data,we demonstrates that this method works very well in enhancing the signal-to-noise ratios of CCFs by rejecting noise sources which do not positively contribute to the recovery of EGFs.Then,this method is applied to practical noise data recorded in western USA.Comparisons have shown that reliable and accurate phase velocities can be measured from 15-d long ambient noise data using our RMSR SS method.By applying this method to ANT,one can reduce the deployment duration of seismic stations from several months or years to a few tens of days,significantly improving the efficiency of ANT in imaging crust and upper-mantle structures.The reduction of data duration requirements for signal recovery makes it possible to extract weak signals that cannot be extracted by the traditional method.(2)To retrieve long-period surface waves from ambient noise,we apply our RMSR＿SS method to ambient noise data recorded by stations in the United States.Love wave signals at the period of 50-150 s are recovered.Through the comparison of the results obtained with different duration data,it is found that the continuous waveform data required for the recovery of the long-period Love surface wave is much larger than the data required for Rayleigh waves.Statistical analysis show that the energy of the long-period Love surface wave may come from large earthquake events.High-intensity local noise near the station hinders the recovery of long period Love wave signals.This thesis also found that there is a phase deviation between the CCF of the horizontal component and the CCF of the vertical component.Since this deviation is systematic,the impact of this systematic deviation needs to be considered when applying Love wave signals from the horizontal component CCFs.(3)We develop an approach to retrieve body wave empirical Green’s functions from ambient seismic noise.We conduct a series of simulation to verify the feasibility of extracting body wave signals from ambient noise.Simulation experiments have shown that for the extraction of body wave signals,time-domain normalization is necessary.At the same time,the normalization parameters need to be assigned to the frequency of the body wave signals.Targeted pre-filtering is also necessary during pre-process.After stacking CCFs with the same station separations,we can extract direct P wave and S wave and their reflection signals from Moho.(4)We propose a method called Common Receiver Station stacking method to retrieve body waves.In this method,by shifting body wave signals generated from the ocean storms in the CCF,the positive and negative causal parts of the CCF can be regarded as two equivalent auto-correlation functions.We improve the signal-to-noise ratio and stability of the reflected signals by stacking those equivalent auto-correlation functions.We use the shifted CCF to replace the traditional auto-correlation function,which can effectively avoid the interference of surface wave energy.The method of CRS stacking can effectively improve the utilization of data,making it possible to obtain stable and reliable results.Through a case study of applying this method to a dense array across the Tanlu Fault in the Shandong Peninsula,we demonstrate our method can retrieve PmP reflected signals very well.In conclusion,the recovery of long-period Love wave and the extraction of body waves from ambient noise data further enrich the research content of ANT,and provide a new data source and technical support for the study of the fine velocity structure and deformation of the lithosphere.