The Applications Of Ray Theory To Studies Of Seismic Source And Interfaces In Earth Interior

The ray theory plays key roles in seismological studies includingtomographic imaging, discontinuity study, source mechanism as well as earthquake locations.The modern seismic broadband stations deployed around the world produce abundant high quality earthquake and ambient noise records, and present unprecedented opportunites for detailed seismological studies. With the ray theory as the vital tool, this paper includes studies on various micoseismic sources, a new calibration location method, resolving the shear velocity Vs at the top of inner core and a new traveltime calculation method. Since the ray theory help to understand many issuses, we will discuss the limitation and future development of ray theory in the end.The discontinuity and lateral heterogeneity are the two complementing features of the earth seismic structure. The interior of earth is divided into four layers with first order discontinuities:the Moho (crust-mantle), the CMB (core-mantle), and the ICB(inner-outer core). The elasticicity contrast across the ICB reflects the material composition and dislocation properties of the upper most inner core which provides information about the lower boundary of outer core convection. The Vs obtained from most previous seismologic studies are obviously less than mineralogical experiments and numerical simulations. The amplitude ratio of PKiKP to P is dominated by the contrast of Vs and density across the ICB. We analyzed more than400observations of PKiKP recorded at the IL array. Taking the differential effects of inelastic attenuation between PKiKP and P into consideration, a new estimation of Vs and density are obtained. The Vs (3.2-4.0km/s) is much closer to the recent mineralogical results.Besides the earthquake signals, the seismometer also continuously record the ambient noise. The ambient noise tomography (ANT) is one of the important seismological innovations in last decade and it has been broadly applied in studies at different scales. Recently, more and more evidences indicate the impact of climate change on seismic records. The location and mechanism of ambient seismic noise source play key role in these studies. We first focused on the oceanic noise in chapter three. The wind waves are excited by wind near the sea surface(especially cyclones), and swells form after wind waves propagates further away. The strong microseismic excited by the Hurricane Katrina is observed at global seismic networks. We analyzed the signals in three frequency bands which are generated in different region. The0.2-0.45Hz SPDF is earliest observed at local stations (<15degree) and the source region is proposed to be near the eye of the hurricane. After the swell reaches the coastal area, the SF is efficiently excited in shallow water and propagates as surface wave showing up on stations around the world (<70degree). Due to strong scattering and inelastic attenuation, the LPDF correlated with the SF is only observed in regional stations (<30degree). Besides the ocean storms, the small climate event also generate ambient noise. We analyzed an anomalous microseism in Gulf of Mexico on April28th,2011. The synthetic seismograms with a wide area of wind wave as source explain both amplitude and traveltime features on the observations. This source model also agrees with the sea surface wind field revealed by satellite data. In the end of chapter3, we used the data provided by dense array in Iceland to calculate the NCR The asymmetry of NCF shows clear seasonal variation of oceanic noise.Besides the oceanic noise sources, there are several monochromatic peaks in spectrum of seismic record. These signals are considered to be excited by persistent localized sources which compromise the validity of EGF. The26s microseism signal are reported in observations at global seismic network except in South America. Previous studies located the source in the gulf of guinea, and a mirror source is situated in north Fiji basin. Based on the NCFs of station-pairs in southern Pacific and Australia, we relocate the Fiji source with a grid search location method and apply calibration from earthquake record. The new location result is outside the antipodal area of Gulf of Guinea and the temporal variation pattern is also different from the source in Gulf of Guinea. Both the earthquake record and NCFs supports the theory that there is an independent26s signal source in FIJI basin.The26s signal source in the Gulf of Guinea shows strong temporal variation in amplitude and the strongest ones could be observed on single station records without much processing. According to the constant arrivals on NCFs, we confirmed the identical locations of26s during burst event and calm days. The ocean wave height during burst event is calculated with WAVEWATCH III and no anomalos wave is detected, thus arguing against the oceanic generation mechanism. The radiation patterns of the burst events are obtained with global seismic records and shows stable azimuthal asymmetry which may be explained with two horizontal source with π/2phase difference.Very strong10s signals are present on one-month NCFs of FNET stations and on many years NCFs of GSN stations in the East Asia. Locationresults from methods of energy stacking and traveltime indicate that the source is situated on Kyushu Island, Japan. It does not show strong seasonal variation in location and amplitude which are expected for common characteristics of oceanic source. Therefore we propose that it is not an oceanic source and may be related to the magma activity of a nearby volcano ASO. Both the26s micorseismic sources and the Kyushu source could be used to monitor the temporal variation of Earth and to assess performance of timing systems in seismograph systems.To suppress contamination for earthquakes and persistent microseism sources, the temporal and spectral normalization are usually adopted in the standard procedure of NCF computation. Both normalizations in time and frequency domains are nonlinear because the procedures abandon the amplitude information. We calculated NCF in long time span without normalizations and compared it with the one obtained from standard procedure. The NCFs of three station-pairs in different separation are similar. The new procedure reduced time cost and keep as much amplitude information as possible, thus it is helpful to studies of microseism source and attenuation tomography.Most ANT studies focus on the crust structure. The mantle structure study with ANT requires longer period EGF which is limited by effective band of seismometer. The super-conduct gravimeter (SG) could record ultra long period acceleration signal. We analyze the NCF between SGs and seismometers which provide long period (>100s) EGFs. The feasibility was confirmed with both seismic and SG stations on different continents.The velocity heterogeneity is one of major factors leading to earthquake mislocations. The NCF provides reliable travel time calibration with EGF between stations.We proposed a new NCF calibration method involving of group traveltime and applied it to location of1998Zhangbei earthquake and2005Klannie earthquake sequence. Our relocation achieves GT5accuracy as compared with location from InSAR data.When the separation between two earthquakes is less than the wavelength of seismic wave, the traditional earthquake location methods are difficult to reolve their spatial interval. The coda wave interforemtry (CWI) proposed by Snieder is claimed to be able to fix this problem. With the reciprocity theorems, the records of same earthquakes in two close stations are used to assess performance of the CWI method. In small range (<<wavelength), CWI could provide reliable estimation while there is cut-off in large range.The traveltime computation is a basic issue in earthquake location, exploration, and traveltime tomography etc.. Most traveltime calculation methods are difficult to handle the model with irregular interfaces. The methods using finite difference to solve eikonal equations could calculate traveltime filed rapidly. We adopted similar framework with new finite difference operators for irregular grids to improve the computation efficiency. The traveltime field in an Earth model with a rolling660km discontinuity is demonstrated to show that topography effect is substantial.The precursors of PKPPKP (P’P’) are used to constrain the depth and sharpness of discontinuity in mantle. The waves preceding P’P’ is usually thought to be generated by corresponding mantle discontinuities, but the situation is not simple. The geometric ray theory predicts the PKPPcP will disappear beyond275degree, but the synthetic seismograms obtained from full wave theory and mode summation show that the PKPPcP is clear in low frequency band (<1Hz) even exceed280degree and the slowness is also close to the previous reported P’785P’. Therefore we infer that some P’P’precursors are actually PKPPcP and hence the785km discontinuity may not exist.