University of Southern California Dissertations and Theses

Systematic analysis of crustal anisotropy and attenuation using seismic data associated with the 1999 Chi-Chi, Taiwan, earthquake

(USC Thesis Other)

Systematic analysis of crustal anisotropy and attenuation using seismic data associated with the 1999 Chi-Chi, Taiwan, earthquake

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content NOTE TO USERS This reproduction is the best copy available. ® UMI R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. SYSTEMATIC ANALYSIS OF CRUSTAL ANISOTROPY AND ATTENUATION USING SEISMIC DATA ASSOCIATED WITH THE 1999 CHI-CHI, TAIWAN, EARTHQUAKE by Yunfeng Liu A Dissertation Presented to the FACULTY OF THE GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY (EARTH SCIENCES) August 2004 Copyright 2004 Yunfeng Liu R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. UMI Number: 3145234 INFORMATION TO USERS The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleed-through, substandard margins, and improper alignment can adversely affect reproduction. In the unlikely event that the author did not send a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion. ® UMI UMI Microform 3145234 Copyright 2004 by ProQuest Information and Learning Company. All rights reserved. This microform edition is protected against unauthorized copying under Title 17, United States Code. ProQuest Information and Learning Company 300 North Zeeb Road P.O. Box 1346 Ann Arbor, Ml 48106-1346 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. ACKNOWLEDGEMENTS I would like to express my sincere gratitude to my dissertation advisors Dr. Yehuda Ben-Zion and Dr. Ta-Liang Teng for their support, patience and encouragement during the course of my graduate study. Their technical and editorial advices are essential to the completion of this dissertation and have taught me innumerable lessons and insights on the workings of academic research in general. I would like also to thank my committee members Dr. Stephen Hass, Dr. Charlie Sammis and Dr. Vladimir Lyakhovsky. The guidance and cooperation of Dr. Lyakhovsky were very important for my early graduate research. My thanks also go to my classmates Zhegang Peng, Youlin Chen and Shoshana Levin for their help and cooperation. I am grateful for the financial support I have received from the department of earth sciences of USC. Last but not least, I am deeply thankful to my wife Hailuo Zhou, our daughter Manlu Liu and my parents. It is impossible for me to complete my Ph.D. program without their support and sacrifice. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. TABLE OF CONTENTS ACKNOWLEDGEMENTS ........ ii LIST OF TABLES AND FIGURES-— ....................................-.................................... .vi ABSTRACT — .....................................-................................................. xii INTRODUCTION.................................-................................. -.—- 1 CHAPTER I : SHEAR-WAVE SPLITTING AND SPATIAL DISTRIBUTION OF CRUSTAL SEISMIC ANISOTROPY IN TAIWAN.....................................................4 1.1 Introduction....................................................................... 4 1.2 Shear-Wave Splitting Analysis Methods and Feasibility Study — ................... 8 1.2.1 Visual Inspection M ethod ................ 8 1.2.2 Automated and Quantitative Method ......... ...............-............................9 1.2.3 Feasibility and Reliability Study................................ 10 1.3 Shear Wave Splitting Measurements and Crustal Anisotropy in Taiwan 15 1.3.1 Data processing and Analysis Results.--......................................................15 1.3.2 Depth extent of Crustal Anisotropy.............................. 16 1.3.3 Relationship between crustal anisotropy and regional tectonic stress.— 26 1.4 Conclusions- ...... ...... —------------ 30 CHAPTER II: SYSTEMATIC ANALYSIS OF SHEAR-WAVE SPLITTING IN THE AFTERSHOCK ZONE OF THE 1999 CHI-CHI, TAIWAN, EARTHQUAKE: SHALLOW CRUSTAL ANISOSTROPY AND LACK OF PRECURSORY VARIATION..........................—................................................................................... 32 2.1 Introduction ------------- 33 2.2 Data Set and Geologic Background ...... 35 2.3 Analysis Methods -......................-.............. 37 iii R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 2.3.1 Shear-Wave Splitting Measurement------------------------------------------------37 2.3.2 Estimation of Near-Surface Anisotropy from Borehole D ata---------------43 2.4 Data Processing and Uncertainty in SWS Measurements--------------------------45 2.5 Near-Surface Crustal Anisotropy-------------------------------------------------------- 49 2.6 Depth Distribution of Crustal Anisotropy--------------------------------------------- 55 2.6.1 Observation of Anisotropy in the Crust Deeper than 8 k m ------------------ 55 2.6.2 Estimation of Anisotropy in the Crust Shallower than 8 k m ---------------- 58 2.6.3 Normalization by Travel Distance versus by Travel Time------------------- 61 2.7 Temporal Change of SWS Associated with Stress Changes Induced by Large Earthquakes-----------------------------------------------------------------------------------------62 2.8 Conclusions----------------------------------------------------------------------------------- 68 CHAPTER III: NEAR-SURFACE SEISMIC ANISOTROPY, ATTENUATION AND DISPERSION IN THE AFTERSHOCK REGION OF THE 1999 CHI-CHI, EARTHQUAKE------------------------------------------------------------------------------------- 71 3.1 Introduction----------------------------------------------------------------------------------- 72 3.2 Data Set and Geological Background---------------------------------------------------73 3.3 Methods for Attenuation Analysis -------------------------------------------- 76 3.4 R esults-------------------------------------------------------------- 81 3.4.1 Estimated Q Values From Staked Waveforms of M ultiplets---------------- 81 3.4.2 Estimated Q Values from a Set of 156 Recordings--------------------------- 86 3.5 Discussion------------------------------------------------------------------------------------- 88 3.5.1 Attenuation in the Crust---------------------------------------------------------------88 3.5.2 Body Wave Dispersion----------------------------------------------------------------89 3.5.3 Attenuation Anisotropy---------------------------------------------------------------93 iv R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 3.6 Conclusions....................................................... 95 RECAPITULATION................-.............................. -................................................... 98 REFERENCES...............- ------- 100 APPENDIX: SEISMIC DATA AND ANALYSIS CODES-— ................ -........... 109 v R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. LIST OF TABLES AND FIGURES Figure 1.1. Schematic diagram showing continent-arc collision and plate tectonic setting of Taiwan, (from Taiwan Central Geological Survey website)................ 6 Figure 1.2. Location map of more than 650 TSMIP stations.....................................- 7 Figure 1.3. Two types of seismograms and noise (all taken from observed data) used to generate seismograms for reliability study................. -------- 12 Figure 1.4. Test results with different time delays for waveform type 1. We fix a - 25° in the calculation. The original PD is 60°— .....................................-........... 12 Figure 1.5. The test results with different for waveform type 1. We fix a = 25° in the calculation. For small a values, one of the projected waveforms is much weaker than the other.-............. — -------------------------------- ----------- --------- 13 Figure 1.6. The test results with different time delays for waveform 2. We fix a - 25° in calculation..........................................................................................13 Figure 1.7. The test results with different for waveform type 2. We fix TD = 0.1 sec in the calculation. -..................... 14 Figure 1.8. Distribution of event magnitude in data set used for SWS analysis.----- 17 Figure 1.9. Measured SWS parameters from data recorded by TSMIP stations in the time period from 1993 to 2000. The direction of each bar gives the PD of the FSW and its length represents the DT of the SSW. The bars are put in the mid way between the epicenter of event and the station............................................ 18 Figure 1.10. (a) Measured time delays versus depth for CHY030 and CHY036. (b) Measured SWS parameters (bars) superimposed on the hypocentral locations of associated events. The bars are oriented parallel to the PD of the FSW and scaled by the TD. Grey scale of bars indicates that data are obtained from the station marked with the same grey scale, (c) Spatial distribution of the events used to measure the time delays of (a). The events associated with time delays < 0.13 sec and >0.13 sec are represented by open and solid circles respectively. Note that the dashed lines are not the ray paths. They only serve as an association reference. Due to increasing velocity with depth, rays are curves concaving upwards. This makes the incident angle well within shear-wave window............. 20 vi R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. Figure 1.11. (a) Measured time delays versus depth for CFIY060, CHY071 and CHY091. (b) Measured SWS parameters (bars) superimposed on the hypocentral locations of events employed to SWS analysis. The bars are oriented paralleled to the PD of the FSW and scaled by the TD. Grey scale of bars indicates that data are obtained from the station marked with the same grey scale. -.............................. —........................................................................... — 21 Figure 1.12. (a) Measured time delays versus depth for CHY059and CHY077. (b) Measured SWS parameters (bars) superimposed on the hypocentral locations of events employed to SWS analysis. The bars are oriented paralleled to the PD of the FSW and scaled by the TD. Grey scale of bars indicates that data are obtained from the station marked with the same grey scale...........................— 22 Figure 1.13.(a) Measured time delays versus depth for CHY009, CHY047 and CHY073. (b) Measured SWS parameters (bars) superimposed on the hypocentral locations of events employed to SWS analysis. The bars are oriented paralleled to the PD of the FSW and scaled by the TD. Grey scale of bars indicates that data are obtained from the station marked with the same grey scale. — .................................................................... 23 Figure 1.14. (a) measured time delays versus depth for HWA009, HWA012, FIWA013 and CHY014. (b) measured SWS parameters (bars) superimposed on the hypocentral locations of events employed to SWS analysis. The bars are oriented paralleled to the PD of the FSW and scaled by the TD. Grey scale of bars indicates that data are obtained from the station marked with the same grey scale. —...............-...........................................-......................................— 24 Figure 1.15. GPS velocity vectors of Taiwan (Y uetal. 1997)-------------------------- 25 Figure 1.16. Measured SWS parameters (bars) superimposed on stations in the area covers CFIY station group (Here only the character C is plotted instead of CHY). The bars are oriented paralleled to the PD of the FSW and scaled by the T D . ---- -............................ -................26 Figure 1.17. Measured SWS parameters (bars) superimposed on stations in the area covers TCU station group (Here only the character T is plotted instead of TCU). The bars are oriented paralleled to the PD of the FSW and scaled by the TD. - 28 Figure 1.18. Measured SWS parameters (bars) superimposed on stations in the area covers TAI and ILA station groups. (Here only the characters T and I are plotted instead of TAI and ILA respectively). The bars are oriented paralleled to the PD of the FSW and scaled by the T D ........................................................29 Figure 1.19. Measured SWS parameters (bars) superimposed on stations in the area covers KAU and TTN station groups. Here only the characters K and T are vii R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. plotted instead of KAU and TTN respectively). The bars are oriented paralleled to the PD of the FSW and scaled by the T D . ----- 30 Figure 1.20. Measured SWS parameters (bars) superimposed on stations in the area covers HWA station group (Here only the character H is plotted instead of HWA). The bars are oriented paralleled to the PD of the FSW and scaled by the T D .................................................................................................-........................ 31 Figure 2.1. A location map of the study region with the Meishan fault (MSF), the Chelungpu fault (CLF) and the Chukou fault (CKF). Solid triangles indicate short-period stations including a 200 m deep downhole station CHY. Solid stars represent the September 20, 1999 Mw 7.6 Chi-Chi earthquake and its two large aftershocks. Solid and open circles represent other small aftershocks recorded by the borehole station CHY and the strong-motion station CHY073, respectively. -..................................................................... — 36 Table 2.1. The parameters of the 1999 Chi-Chi earthquake and its two large aftershocks........................-......... -------------------------- -------- ------- 37 Figure 2.2. An example of SWS analysis with the aspect ratio method using short- period data, (a) Original three-component seismograms. (b) A contour of the AR values. The maximum value indicated by the cross corresponds to the PD of the FSW (170o in this example), (c) Seismograms rotated to the fast and slow PDs. (d) The CCC between two split shear waves determined by (b). The maximum value is related to the TD (0.16 sec in this example), (e) Waveforms of the fast and slow components shifted with the measured TD. (f) The horizontal particle motion of the original seismograms. (g) The horizontal particle motion of the fast and slow components shifted with the measured TD. -............................................................................................... 40 Figure 2.3. An example of SWS analysis with the cross-correlation method using strong-motion data, (b) A contour of the CCC values. The azimuth and time delay of the maximum CCC value are indicated by the cross and give the PD of the FSW (177o in this example) and the TD (0.20 sec in this example), respectively, (d) A cross section of (b) at the resolved PD of the FSW. (a), (c), (e), (f), and (g) correspond to respective panels in Figure 2.2............................41 Figure 2.4. Four examples of three-component seismograms in the downhole station CHY. The events number, magnitude, focal depth and epicentral distance are indicated in the upper left corner of each panel. The waveforms of the resolved fast and slow shear waves, together with determined PD and TD, are given in the bottom of each panel. Here P and P indicate up-going and down-going P waves, while Sf and S1 ' indicate up-going and down-going S waves.—..............42 viii R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. Figure 2.5. Schematic geometry for the direct and surface-reflected waves in the borehole configuration.------------------ — .... -..................... 43 Figure 2.6. Measurement of TD in the top 0.2 km of the crust from the autocorrelation coefficients of the resolved FSW and SSW. The inset shows the seismograms of the FSW (upper panel) and the SSW (lower panel). The shaded regions represent direct phase (S ) and reflected phase (S'), respectively. The dt f and dtx values represent two-way travel times in the top 0.2 km of the crust of the FSW and SSW, respectively.--------------------.......................................... 44 Figure 2.7. Measured PDs of the FSW and TDs from the AR method (solid circles) and the CC method (open circles)..................................................-..................... 47 Figure 2.8. (a) TDs versus incident angles of ray paths represented by the ratios between epicentral distance and event depth, (b) TDs versus magnitudes, (c) The distribution of three categories of TDs in the NS profile, (d) The distribution of three categories of three categories of TDs in the EW profile. -48 Figure 2.9. (a) Histogram of measured TDs in the top 0.2 km of the crust, (b) Measured TDs in the top 0.2 km versus measured TDs in the deeper section. - 50 Figure 2.10. A comparison of measured results from downhole records with those from surface observations, (a) Rose diagram of the PD of the FSW measured in the downhole station CHY. (b) Rose diagram of the PD of the FSW measured in the surface station CHY073. (c) Histogram of measured TDs from downhole records, (d) Histogram of measured TDs from surface records. ............— 51 Figure 2.11. SWS analysis of the calculated data based on a two-anisotropic-layer model using the CC method. The model parameters are PD = 170° and TD = 0.16 sec in the lower layer and PD = 177° and TD = 0.04 sec in the upper layer, (a) Waveforms generated from the EW component of observed seismograms (Figure 3a) using the assumed model. Panels (b), (c), (e), (f), and (g) correspond to respective panels in Figure 2.3.------ 54 Figure 2.12. (a) Measured TDs versus depth of events, (b) Measured PDs of the FSW versus the depth of events. The solid triangles give the average value over 2 km depth intervals...................................... -.......................................... 56 Figure 2.13.(a) A model of S wave velocity from 0 to 8 km depth and corresponding travel times beginning from surface, (b) Inferred TD beginning from surface based on two models of k, two boundary conditions of k, and the S-velocity model of (a).-------------- -.............. — ......... -............................ 57 ix R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. Figure 2.14.(a) Measured TDs in the crust below 0.2 km versus time. Arrows indicate the times of the Mw 7.6 Chi-Chi mainshock and two M 6.4 and M 6.0 Chiayi aftershocks, (b) Measured PD of the FSW below 0.2 km versus time, (c) Measured TDs in the top 0.2 km versus time. The mean values and standard deviations of the results before and after the Chi-Chi mainshock are indicated by solid and dashed lines......................................-.......................... -...................63 Figure 2.15. Earthquake hypocenters before (triangles) and after (circles) the Chi-Chi mainshock. The solid symbols indicate event pairs separated by less than 2 km. ................................................................................................................................. 65 Figure 2.16. Seismograms of a cluster of earthquake multiplets. (a) The EW components of seismograms. (b) The NS components of seismograms. The time, longitude, latitude and depth of each event, and the measured PD of the FSW and TD, are indicated above each trace, (c) The FSW component of seismograms in a stacked form, (d) The SSW component of seismograms in a stacked form. —..................................................................................................— 67 Figure 3.1. A location map of the study region with the Meishan fault (MSF), the Chelungpu fault (CLF) and the Chukou fault (CKF). Solid triangles indicate short-period stations including a 200 m deep downhole station CHY. Solid stars represent the September 20, 1999 Mw 7.6 Chi-Chi earthquake and its two large aftershocks. Solid circles represent other small aftershocks recorded by the borehole station CHY......................................................... 74 Figure 3.2. Pictures of drill core samples from a hydro-geological well #200201G1 in the study area. The total depth of the well is 250 m and the depths of core samples are indicated in the figure. (From the Hydrogeology Data Bank, The Central Geological Survey of Taiwan.).................................................................75 Figure 3.3. Schematic geometry for the direct and surface-reflected waves in the borehole configuration. -.................... —................................................. 78 Figure 3.4. Stacked fast and slow components of horizontal shear waveforms for a set of 7 earthquake multiplets. The vertical shaded areas indicate portions of the seismograms used in followingattenuation and dispersion analyses, while and the horizontal boxes mark portions used for background spectral analysis in Figure 3.5................. -................... — ......................-...............................................82 Figure 3. 5. (a) Amplitude spectra of the direct and reflected shear-wave phases and amplitude spectra of background noises prior to and following the direct shear waves, (b) Amplitude spectral ratios versus frequency and linear fitting to equation (3.4) in the frequency range 2 - 1 5 Hz -...............— ................. 83 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. Figure 3.6. (a) Fitting errors between the observed wavefromss and the calculated wavefromss by equation (9) with different trail values of Q. The minimum error values indicated by arrows correspond to the estimated and , respectively, (b) comparison between between observed waveforms and calculated ones corresponding to the estimated Q values. .................. 85 Figure 3.7. (a) Direct and reflected windowed phases of the fast and slow shear wave components for 156 recordingss. (b) Amplitude spectral ratios versus frequency. The heavy solid lines represent the average values and their linear fittings to equation (3.4) in the frequency range 2 - 15 FIz give the estimates of Qf and Qs....... - ................... 87 Figure 3.8. Distributions of and calculated with the waveform fitting method for 156 recordings................................................................................................................ 88 Figure 3.9. (a) Observed phase velocities versus frequency and calculated ones based on equation (14). (b) Results of waveform fitting with equation (8) compared with the results without dispersion..................................... — 92 Table 3.1. The estimated Q from two data sets using two methods.------------------- 94 Figure 3.10.Distributions of the amplitude ratios between the reflected and direct waves in time domain for the fast and slow components. The mean and standard deviation values of the amplitude ratios are 0.65and 0.10 for the fast component, while the corresponding values for the slow components are 0.55 and 0.08....... - ........................................................................................................ 95 xi R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. ABSTRACT We analyze shear-wave splitting (SWS) in a high-quality waveform data set recorded at surface and downhole (0.2 km) seismometers in Taiwan. The data set was generated by events in the time period 1993 -2000, which includes the September 20, 1999 Mw 7.6 Chi-Chi, Taiwan, earthquake sequence. The purpose is to investigate the spatial distribution of stress-induced crustal anisotropy and its possible temporal evolution in relation to the occurrence of large earthquakes. The SWS analysis employs both the aspect ratio and cross-correlation methods to obtain robust measurements. Measured results from strong motion stations that cover well the study region show that velocity anisotropy seems to be mainly controlled by local tectonic stress. Some measurements from stations nearby active faults display a polarization direction parallel or sub-parallel to the fault strike. Measured fast polarization directions and time delays of SWS vary significantly with location. No dependence of time delay on depth has been found in various areas over depth range 5-18 km. Seismic Anisotropy in study region appears to be distributed in shallow crust. Analysis results based on recordings from a borehole station further indicate that observed anisotropy is confined primarily to the top 2-3 km of the crust and is dominated by the top few hundred meters. The observed SWS parameters and the great similarity of waveforms generated by clusters of earthquake multiplets, show no appreciable systematic temporal changes over the 2.7 year period before the 1999 xii R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. Mw 7.6 Chi-Chi mainshock and the 2.3 year period after. The observed lack of temporal change of SWS parameters is obtained from a data set recorded in the vicinity of large rupture zones that experienced stress changes that are as big as expected to occur in the brittle crust during large earthquake cycles. We have also measured the seismic attenuation in the top 0.2 km of the crust by analyzing surface-reflected waves in the borehole recordings. The results reveal a substantial difference of attenuation between the fast and slow shear wave components and show a clear evidence of attenuation anisotropy in the near-surface structure. An attenuation-associated dispersion is clearly observed and it has a significant effect on the shapes of waveforms. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. INTRODUCTION The Earth’s crust is highly damaged by tectonic and other processes. Rock damage associated with microcracks and fractures is distributed in entire crust, but concentrated in fault zones. Because microcracks and fractures are related to evolving rock deformation, it is important to investigate their distribution and behavior both within and outside the fault and earthquake rupture zones. Many studies (e.g. Krajcinovic, 1996; Hamiel et al. 2004 and references therein) describe successfully the effects of damage on the elastic properties for a deforming rock in term of damage rheology. The sensitivity of seismic wave propagation to cracks and fractures is one of the fundamental observations of rock physics. Cracks and fractures decrease the P- and S- wave velocities, increase the velocity and attenuation anisotropy, increase the velocity dispersion and wave attenuation, and increase the potential for stress-induced phenomena. When seismic waves propagate through damaged crustal rock, they may carry these seismic signatures of cracks and fractures in seismograms. Shear-wave splitting (seismic birefringence) is the most diagnostic, informative, and easily observable features of azimuthal seismic anisotropy in relation to stress- induced microcrack alignment. We can investigate the spatial distribution of cracks in the crust through shear-wave splitting analysis. Moreover, possible temporal changes could be related to stress field variations and provide a potential tool for earthquake forecasting. 1 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. Taiwan is in an active plate boundary regime. The GPS measurements show that the central and Southern Taiwan is undergoing rapid crustal deformation. It is expected that the stress filed in this region is also varying rapidly with time and space. Numerous events (including the 1999 Chi-Chi Earthquake) are well recorded by a high-density network of seismic stations built over the last decade. Great coverage of seismic data and diversity of tectonic features make Taiwan an idea place for study on the shear-wave splitting and spatial distribution of crustal anisotropy in relation to local tectonic. The analysis on temporal variation of shear wave splitting in the surrounding area of the Chi-Chi Earthquake provides a great opportunity to test possible stress change associated with shear wave splitting before and after big earthquakes. In Chapter 1 (Liu et al, 2004c) of the thesis, we first discuss methodology of shear- wave splitting and related issues and then analyze strong-motion data, which covers most of the Taiwan area. The analysis results show that velocity anisotropy seems to be mainly controlled by local tectonic stress. Some measurements from stations nearby active faults display a polarization direction parallel or sub-parallel to the fault strike. No dependence of time delay on depth has been found in various areas over depth range 5-18 km. Seismic Anisotropy in study region appears to be distributed in shallow crust. In Chapter 2 (Liu et al, 2004a), we systematically analyze SWS on a borehole short 2 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. period data set in the aftershock zone of the Chi-Chi earthquake and investigate depth distribution of anisotropy in study area and temporal variations of SWS parameters in relation to the mainshock. The results further indicate that observed anisotropy is confined primarily to the top 2-3 km of the crust and is dominated by the top few hundred meters. The measurements of SWS also reveal lack of precursory temporal variations. In Chapter 3 (Liu et al, 2004b), we accurately measure the quality factors of the fast and slow shear waves in near-surface structure from the surface reflected waves of borehole recording using different methods. The results indicate the existence of strong attenuation anisotropy and body wave dispersion in the top 200 m crust of the study region. 3 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. CHAPTER I : SHEAR-WAVE SPLITTING AND SPATIAL DISTRIBUTION OF CRUSTAL SEISMIC ANISOTROPY IN TAIWAN 1.1 Introduction Shear-wave splitting (SWS) or birefringence refers to the generation of a pair of orthogonally polarized shear wave with different travel velocities from a single initial wave upon its passage through an anisotropic medium. It has been widely observed in the continental crust (e.g. Booth et al.,1985; Crampin, 1981; Shih et al.,1990; Kaneshima, 1990; Li, 1996; Liu et al, 1997). The phenomenon of shear- wave splitting in the uppermost crust has been interpreted as the effect of distributions of stress-aligned fluid- filled microcracks and preferentially oriented pore space, known as “extensive-dilatancy anisotropy” or EDA (Crampin, 1981, Crampin et al. 1984). The polarization direction of the faster shear wave has been found to correspond well to the direction of the maximum compressional stress, which could be indicated by earthquake focal mechanism analyses, geological observations and GPS measurements in tectonic region. Kaneshima (1990) pointed out that besides the vertical aligned stress-induced microcrack, at least two phenomena, cracks or fracture associated with active faults and intrinsic rock anisotropy due to preferred orientation of minerals, should be taken into account in studying the origin of crustal anisotropy. The shear wave splitting analysis in the aftershock region of the 1995 Kobe earthquake by Tadokoro (1999) also revealed that shear faulting of the earthquake is a major cause of upper crustal anisotropy in R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. the vicinity of active faults. The fabric-related velocity anisotropy in rocks from Santa Rosa Mylonite Zone was reported from direct laboratory measurements at conditions corresponding to those in Earth by Kern an Wenk (1990). As is well known, the effect of thin layering and bedding of sediments has be long been recognized as a source of seismic anisotropy in the filed and in the laboratory (Leary at al. 1990). Taiwan is situated on a convergent and compressive boundary between the Eurasian and the Philippine Sea plates (Figure 1.1). It displays the results of an active collision of the northeast-trending Eurasian continental margin of the Eurasian plate with the north-trending Luzon arc of the Philippines Sea plate (Ho 1982; Rau et al 2000). At Taiwan, the relative plate motion between the Eurasian plate and the Philippine Sea plate is at a rate of 70-80 mm yr'1 in the direction N306°E (Annemarie et al 2003 and reference therein). The Philippine Sea plate subducts north-northwestward underneath the Eurasian plate at the Ryukyu Trench, whereas it overrides westward the Eurasian plate at the Manila Trench. Along the east coast, the NNE trending Longitude Valley Fault is generally considered as the suture zone between the two plates. Thus the interaction of these two plates gives rise to a very complex tectonic structure of Taiwan. In between these two subduction zones and in the midsection of western Taiwan, there is a series of large thrust faults that take up the bulk of the plate convergence rate. GPS measurements show that the central and southern Taiwan is undergoing rapid crustal deformation. It is expected that the stress filed in 5 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. this region is also varying rapidly with time and space. TAIWAN EURASIA PLATE EURASIA PLATE CENTRALRANGE CO A STA L R A N ( I t §& ^ ^ l i f e Ph il ip p in e sea; p l a t e Figure 1.1. Schematic diagram showing continent-arc collision and plate tectonic setting of Taiwan, (from Taiwan Central Geological Survey website) The Central Weather Bureau (CWB) began a program to densely cover all of Taiwan with modem digital strong-motion instruments in the early 1990s. This came to be known as the Taiwan Strong-Motion Instrument Program (TSMIP). More than 650 modern digital free-field strong-motion stations were in place in the mid 1990s (solid triangles in Figure 1.2). Up to 2000, there had been numerous events (including the 1999 Mw 7.6 Chi-Chi earthquake sequence) captured by TSMIP network. Great coverage of seismic data and diversity of tectonic features make Taiwan an idea place for study on the shear-wave splitting and spatial distribution of crustal anisotropy in relation to local tectonic. This is the main purpose of this chapter. In 6 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 25 h 24.5 24 H 23.5 h 23 22.5 h 22 h 4 & A K A . \ / a/ a _L 120 120.5 121 121.5 122 Figure 1.2. Location map of more than 650 TSMIP stations. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. addition, we will also discuss shear wave splitting analysis methods and related issues in the chapter. 1.2 Shear-Wave Splitting Analysis Methods and Feasibility Study Both the visual inspection method and automated method have been used in shear wave analysis. 1.2.1 Visual Inspection Method Ando et al. (1983) observe time delays from seismograms directly. In their method, the horizontal seismograms are rotated to various azimuths. The faster polarization direction (PD) is the azimuth at which the fast shear wave (FSW) and slow shear wave (SSW) appear separated most clearly. The time delay (TD) is directly measured from seismogram rotated at this direction. When the signal-to-noise ratio is good, the similarity between the fast and slow shear waves makes the splitting quite convincing. Another commonly used visual inspect method (e.g. Crampin et al. 1985) is based on analysis of the polarization diagram or hodograms. In this method, the polarization direction of the fast shear wave is determined from linear or near- linear particle motion of the first-shear wave arrival. In many case, the arrival of the slow wave is indicated by the change of particle motion from linear to elliptic or just irregular other than an orthogonal particle motion. The time delay between two the split shear waves is defined as the difference between onset of the first shear wave and the change in polarization. Visual inspection methods have problems of subjectivity or observer bias that might affect individual and overall measurements. 8 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. However, visual inspection methods are useful for checking results from automatic processing. 1.2.2 Automated and Quantitative Method Objective automated methods are required to analyze large data sets effectively, in an unbiased and systematic manner. The cross-correlation (CC) method is commonly used for SWS analysis (e.g., Fukao, 1984; Bowman and Ando, 1987). The two horizontal seismograms are rotated in the horizontal plane at a 1° increment of azimuth a from 0°to 180°. Then the cross correlation coefficient between two orthogonal seismograms is calculated for each a with different time delay t in a selected time window. When the absolute values of cross-correlation coefficient c(a, t) reach their maximum, the corresponding values of a and r are chosen as the fast polarization direction and the time delay of the slow shear wave respectively. The underlying assumption of cross-correlation method is that the fast and slow horizontal components have similar waveforms. However, for the local seismograms, the fast and slow horizontal components may or may not display similar waveforms, as the polarizations respond differently to the structure between source and receiver (Liu et al 1997; Aster et al. 1991). The displacement aspect ratio method proposed by (shih et al., 1989). In general, particle motion is elliptical, but it will be linear under two conditions: (1) one component of shear wave is zero; or 2) the phase difference between two components is nn . The maximum value of the aspect ratio in a time interval that contains only the early shear-wave arrival indicates the most linear particle motion, 9 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. i.e. condition (1), and its direction indicates the polarization of the fast shear wave. Then the time delay is determined by comparing the faster shear wave to the slower shear wave determined by condition (1). See the definition of the aspect ratio and the detailed description of the method in Chapter 2. As discussed by Silver and Chan (1991), both the cross-correlation method and the aspect ratio method could been viewed as eigenvalue-based measures of linearity of particle motion. Other various methods may also been put in this context. However, for certain seismograms, the AR method determines the PD and TD separately and is expected to have better resolution than other methods that measure both SWS parameters simultaneously. Although the AR method need measure the first arrival of shear wave accurately, it is not too difficult to pick up the first arrival of shear wave from seismograms. 1.2.3 Feasibility and Reliability Study We have used synthetic data to test the feasibility and reliability of these two automated methods. The original waveform taken from real seismogram is projected into the assumed fast and slow polarization direction to from the fast and slow shear wave components, respectively. Then we lag the slow component with an assumed time delay. We generate the EW and NS components by projecting the above fast and slow shear wave component into the EW and NS directions. To simulate the observed data, we also add noise, which is also taken from the part of the observed 10 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. seismograms before the signal arrival, to the generated EW and NS components of seismograms to produce the synthetic data for analysis. We choose two types of waveforms in the study as shown in Figure 1.3. The waveform type 1 is an ideal type of shear waveform. The peak of waveform is right after the shear-wave arrival. The waveform type 2 is the other type of shear waveform, where the peak of waveform is somewhere behind the arrival. Figures 1.3c and 1.3d also show two components of noise taken from observed seismograms. Figure 1.4 shows the test results with different time delays for waveform type 1. We fix a = 25 in the calculation, a is the angle between the polarization direction of the initial shear wave and the PD of the SSW. We can see that both the AR and CC methods can determine well the PD and TD up n/s = 0.2 , where n/s is noise-to- signal ratio, which is defined as the ratio between the maximum value of noise and the maximum amplitude of shear waves. Figure 1.5 shows the test results with different a for waveform type 1. We fix TD = 0.1 sec in the calculation. The AR method can recover both the PD and TD well for all a up to n/s - 0.20. However, The CC method can not find the PD and TD when n/s > 0.12 and for a < 15 . Figure 1.6 shows the test results with different time delays for waveform type 2. We also fix a - 25 in the calculation. The CC method can recover both PD and TD well up to n/s = 0.28. 11 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 5 waveform type 1 0 •5 18 18.5 19 19.5 20 20.5 21 21.5 22 22.5 23 5 waveform type 2 0 •5 18 18.5 19.5 21.5 22 22.5 19 20 20.5 21 23 EW 0 10 10.5 11.5 12 12.5 13 13.5 14 14.5 15 0 1 0 10.5 11.5 12.5 13 13.5 14 14.5 15 11 12 Figure 1.3. Two types o f seismograms and noise (all taken from observed data) used to generate seismograms for reliability study. cc 100 50 O TD=0,05 + TD=0.10 x TD-0.15 A TD=0.20 QOOOO99999999999 AR 0 0.1 0.2 0.3 0.4 0.5 150 100 0.1 0.2 0.3 0.4 0,5 0 0.3 0.25 0.2 > • > -O 0.15 < u ,§ S 0.1 0.05 0 ■ \A A A A A A A A A A A A A A A A A A A A A A A A x x x x x x x x x x x x x x x x x x x x x x x x x 0 0 0 0 0 0 0 0 0 +++++++++++ 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.1 0.2 0.3 0.4 0.5 S/N ratio 0.3 0.25 0.2 0.15 Jxxxxxxxxx x x x x x x x x u x x x x u x 0.1 0.05 0 X X + + + + ! P0000ooooo00oo0oo0000ooo°l 0 0.1 0.2 0.3 0.4 0.5 S/N ratio Figure 1.4. Test results with different time delays for waveform type 1. We fix a = 25° in the calculation. The original PD is 60° 12 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. c c AR x 100 x 50 100 0.1 0.2 0.3 0.4 0.5 0.4 O a =10.0 0.3 A g =30.0 0.2 u " O o E ooooo o o o o © © © © © © © © () - 0.2 0.1 0.2 0.3 0.4 0.5 S/N ratio 0.4 0.3 0.2 0.1 0 - 0 .1 - 0.2 o a = 10.0 x a =15.0 + a =20.0 A a -30 0 O x o x °A O + 0 A x A A x + + 0.2 0.3 S/N ratio Figure 1.5. The test results with different for waveform type 1. We fix a = 25° in the calculation. For small a values, one o f the projected waveforms is much weaker than the other. c c •2 150 0 TD=0.05 + TD=0.10 x TD=0.15 A -i o I I o o 150 100 Do O' 0.1 0.2 0.3 0.4 0.5 0 0.3 0.25 0.2 > > ■ 3 0.15 0 J E ~ 0,1 0.05 0 0.3 0,25 A A A A A A A A A A A A A A A A A A A A A A A A A | A A A A A A A A A A A A A A A A A A A A A ^ A x xx x xx x x x x x x x x x x x x x x x x x x x 0.15 x x x x x x x x x x x x x x x x Xxx x xxx + + + + + + + + + + + + + + + + + + + + + + + + 4 0.1 f + + + + - t - + + + + + + - M - + + + + + + + + + +-»- O O O O O O O O O O O O O O O O O O O O O O O O O 0.05 0 b o ° o o n ° 0 0 0 0 0 o ° o o n o ° o ° . ° .................. A ..... * 0.2 0.3 S/N ratio 0.1 0.2 0.3 0.4 S /N ratio Figure 1.6. The test results with different time delays for waveform 2. We fix a = 25° in calculation. 13 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. o o o o o o o $ © $ © $ o ® + + © © © o 0.1 0.2 0.3 0.4 0.5 O a =10.0 x a =15.0 + a =20.0 A a =30.0 0 + vQ A A x () O :: X 0.4 O a =10.0 0.3 x a =20.0 A g =30.0 0.2 T3 U E -0 . - 0.2 0.2 0.1 0.3 0.4 0.5 S/N ratio S/N ratio Figure 1.7. The test results with different for waveform type 2. We fix TD = 0.1 sec in the calculation. However, The AR method can only recover the PD and DT reasonably well up to n/s = 0.18. Figure 1.7 shows the test results with different a for waveform type 2. Here we fix TD = 0.1 sec in the calculation. Again, both methods can recover the PD and TD reasonably well up to n/s = 0.15. The CC method has trouble in finding the PD DT for small a and n/s >0.15. From above results, we can see that the AR method is more suitable for waveform type 1 than for waveform type 2. The reason is that in the AR method we use the beginning part of waveform to find PD and noise-to-signal ratio is much larger in 14 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. this part for waveform type 2 than waveform type 1. In the CC method we use the entire range of shear-wave signal to measure PD and TD, and it is expected that it has a better performance than the AR method for the waveform type 2. We note the AR method is not suitable for the case that time delay between two split shear waves is small. However, the CC method always has trouble when a is small. In general, CC method has a better performance than AR method in our synthetic test, because we assume that two split shear waves have the same shape of waveform. However, this assumption is not necessarily satisfied for observed data. In most situations, two split shear waves have the quite different waveforms. In this case, the AR method is expected to have a better performance than the CC method. In summary, both the AR and CC methods have their own advantages in different situations. So we may use the following strategy in measuring shear-wave splitting. We use both AR and CC methods in calculations. If the aspect ratio of the AR method is larger than a certain value (e.g., 10), we take the results of the AR method. Otherwise, we take the results of the CC method. Obviously, the value of cross-correlation coefficient (CCC) could be another criterion in judging the quality of the measurement. If the data set is large enough, we may only use the SWS parameters under the condition that both the AR and CC methods yield almost the same results. 1.3 Shear Wave Splitting Measurements and Crustal Anisotropy in Taiwan 1.3.1 Data processing and Analysis Results. The majority of recordings of TSMIP stations are generated by larger (M > 4) 15 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. earthquakes. However, there still are many small events captured by the TSMIP network. We use data within the shear-wave window (depth > epicenter distance) and with magnitude < 3.0 in our shear-wave splitting analysis. In the time period from 1993 to 2000, there are a total about 2600 recordings that satisfy above criteria. As shown in Figure 1.8, most of strong motion data used in our SWS analysis is generated by events with magnitude 2.5 - 3.0. The acceleration seismograms are filtered with the frequency band range of 1~ 15 Hz and then integrated into velocity waveforms. The later are used in our SWS analysis. The measured SWS parameters resulted from employing the aspect ratio method from 1274 recordings that satisfy the criteria: with aspect ration > 5 and cross-correlation coefficient > 0.75. These criteria are used to filter out unreliable measurements. The results are shown in Figure 1.9. The direction of each bar gives the PD of the FSW and its length represents the DT of the SSW. The bars are put at the mid way between the epicenter of the event and the station. 1.3.2 Depth extent of Crustal Anisotropy We use SWS parameters estimated from five groups of stations distributed in various locations to study depth extent of crustal anisotropy. The measurements from these selected station groups have sufficient data points that satisfy the higher criteria (with aspect ration > 8 and cross-correlation coefficient > 0.8). The first group include two station TCU030 and TCU036. Figure 1.10a shows the measured time delays versus depth. The time delays measured from CHY036 are generally larger that those from CHY030. Both results show no dependence of time delay on depth 16 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 2.2 m agitude Figure 1.8. Distribution o f event magnitude in data set used for SWS analysis. in the range from 7 to 18 km. Figure 1.10b shows relationship between the SWS parameters with the hypocentral locations of events. The results from CF1Y030 are represented by deeper grey bars and those from CHY036 are given by lighter grey bars. Although some measurements from CHY030 and CHY036 employ events in a nearby area, the observed polarization directions and time delays are significantly different from each other. This indicates that the SWS parameters determined by the station location rather than event location and, in turn , this implies that those results controlled by medium near the station. In other words, the crustal anisotropy is dominated at shallow depth. From figure 1,10a we find that the TD values from CHY030 can be classified into two groups anisotropy. We note that the time delays generated by the medium over paths from events (solid circles) to CHY030 are 17 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 22.5 \ V 22 \ 120 120.5 I _____ 121 lo n g itu d e (°) I __ 121.5 1 2 2 Figure 1.9. Measured SWS parameters from data recorded by TSMIP stations in the time period from 1993 to 2000. The direction o f each bar gives the PD o f the FSW and its length represents the DT o f the SSW. The bars are put in the mid way between the epicenter o f event and the station 18 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. (a ) 0.3 0.25 0.2 0.1 0.05 © o ° ° m © © © © © -J__________________ L, 8 10 12 14 depth (km) A CHY036 © CHY030 16 18 20 (b ) 23.75 0.1 sec 23.7 23.65 CHY030 23.6 CHY036 23.55 2 3 .5 120.65 120.4 120.45 120.5 120.55 120.6 longitude (°) R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. (c ) 0 A CHY030 23.7 23.65 23.6 la titu d e ( ° ) n n 120 56 120.58 120 52 120.54 i ^ . jo 12o.42 120.44 120.46 12048 120.5 23.55 lo n g itu d e ( ° ) Figure 1.10. (a) Measured time delays versus depth for CHY030 and CHY036. (b) Measured SWS parameters (bars) superimposed on the hypocentral locations o f associated events. The bars are oriented parallel to the PD o f the FSW and scaled by the TD. Grey scale o f bars indicates that data are obtained from the station marked with the same grey scale, (c) Spatial distribution o f the events used to measure the time delays o f (a). The events associated with time delays < 0 .1 3 sec and > 0 .1 3 sec are represented by open and solid circles respectively. Note that the dashed lines are not the ray paths. They only serve as an association reference. Due to increasing velocity with depth, rays are curves concaving upwards. This makes the incident angle well within shear-wave window. within a small range around 0.15 sec, although depth of these events varied from 7 km to 18 km. Figures 1.11-1.14 shows results from other four groups of stations. Similar to what we see in Figure 1.10, all of results show no depth dependence of time delay and strong dependence of SWS parameters on station locations rather than event locations. 20 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. (a) 0.2 0.18 0.16 0.14 0,12 u 4 2 - 0 . 1 Q H 0.08 0.06 0.04 0,02 0 8 10 12 14 16 18 20 22 depth (km) (b) 23.15 23.1 3 23.05 120.1 120.15 120.2 120.25 120.3 120.35 longitude (°) Figure 1.11. (a) Measured time delays versus depth for CHY060, CHY071 and CHY091. (b) Measured SWS parameters (bars) superimposed on the hypocentral locations o f events employed to SWS analysis. The bars are oriented paralleled to the PD o f the FSW and scaled by the TD. Grey scale o f bars indicates that data are obtained from the station marked with the same grey scale. 0.1 sec CHY060 CHY091 CHY071 * o © & o A A © CHY071 A CHY091 0 CHY060 ©A 21 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. (a) 0.25 A O CD © G A A CHY077 © CHY059 12 14 16 depth (km) 18 20 (b) 23.24 23.22 0.1 sec O O 23.2 A CHY059 23.18 A CHY077 23.16 ’2 23.14 23.12 23. 23.( 23.06 23.04 * — 1 2 0 120.25 120.05 120.1 120.15 120.2 longitude (°) Figure 1.12. (a) Measured time delays versus depth for CHY059and CHY077. (b) Measured SWS parameters (bars) superimposed on the hypocentral locations o f events employed to SWS analysis. The bars are oriented paralleled to the PD o f the FSW and scaled by the TD. Grey scale o f bars indicates that data are obtained from the station marked with the same grey scale. 22 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. CHY009 CHY073 CHY047 0.25 0.2 A A 15 0.1 0.05 0 10 12 14 16 4 6 8 depth (km ) 236 0.1 se c 23.55 <r 235 CHY073 CHY047 CHY009 23.45 1204 120.45 1205 120.55 1203 120.35 longitude (°) Figure 1.13.(a) Measured time delays versus depth for CHY009, CHY047 and CHY073. (b) Measured SWS parameters (bars) superimposed on the hypocentral locations o f events employed to SWS analysis. The bars are oriented paralleled to the PD o f the FSW and scaled by the TD. Grey scale o f bars indicates that data are obtained from the station marked with the same grey scale. 23 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. © HWA009 A HWA012 HWA013 rn HWA014 1 A 9 10 depth (km) (b) 24. 24.1 24.06 0.1 sec 24.04 24.02 23.' 23.96 23.94 23.92 23.9 1 ---- 121.54 121.56 1 21.6 121.62 121.64 121.66 121 .6 8 121.58 logitude (°) Figure 1.14. (a) measured time delays versus depth for HW A009, HW A012, HWA013 and CHY014. (b) measured SWS parameters (bars) superimposed on the hypocentral locations o f events employed to SWS analysis. The bars are oriented paralleled to the PD o f the FSW and scaled by the TD. Grey scale o f bars indicates that data are obtained from the station marked with the same grey scale. 24 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 25 N h 244 N 23 N h 22*N Figure 1.15. 0 50mm/yr VELOCITY SNM S 50km MAP SCALE soil { V y y w otn S 2 JR 5 1 0 2 $041 1 2 0 E 121 ‘E GPS velocity vectors o f Taiwan (Yu et al. 1997) 25 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 23.; 0.1 sec ^ C 0 2 5 23.7 23.6 Mcisha:'. fauflSoS 23.5 0 1 2 3 '4 C d ikou fault 23.3 1 ) 1 7 ^^C058 ^ CO 18 23.2 2 » W nC W I ~ c q u 23.1 2 2 .' 119.9 1 2 0 120.1 120.3 120.4 120.5 120.6 120.7 120.8 120.9 1 2 0 .2 longitude (°) Figure 1.16. Measured SWS parameters (bars) superimposed on stations in the area covers CHY station group (Here only the character C is plotted instead o f CHY). The bars are oriented paralleled to the PD o f the FSW and scaled by the TD. 1.3.3 Relationship between crustal anisotropy and regional tectonic stress. Figure 1.15 shows the GPS the velocity field derived from the 1990 - 1995 data of the Taiwan GPS network (Yu at al. 1997). Such a velocity field provides information on kinematics of the crustal deformation and regional tectonic stress. Figures 1.16- 1.20 show spatial distribution of crustal anisotropy revealed by SWS parameters (bars) in five regions of Taiwan. Each region covers one or two groups of TSMIP stations. In Figure 1,9, we plot SWS parameters (bars) at the middle way between stations and hypocentral locations. As discussed above, the measured SWS parameters mainly reveal the crustal anisotropy just beneath the station. So it is 26 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. reasonable to plot them on the station instead. CHY station group is located in the Chia-Nan area of Taiwan, an area includes the western costal plain and the western foothills. The Meishan fault (MSF) and the Chuko fault (CKF) are two major active faults in the area. The excellent spatial coverage and the reliable measurements of shear-wave splitting obtained in this area provide us a unique chance to study the relationship between crustal anisotropy and regional tectonic stress. As shown in Figure 1.16 the observed polarization directions of SWS in general conform to the maximum compressional filed indicated by GPS velocity vectors (figure 1.15). We note that the observed polarization directions and time delays vary significantly with location. This tells us that the crustal anisotropy may also be affected by local geological structures. Figure 1.17 shows measured SWS data in the area covers TCU station group over Taichung area. 1999 Chi-Chi earthquake ruptured an approximately 100 km segment of the Chelungpu fault of the area. Most of reliable SWS measurements are obtained in a domain between the Chelunpu fault and another major active fault to its west - Changhua fault. Nearly EW polarization directions of most SWS measurement agree with regional maximum compressional axis. Other measurements are sporadically distributed in other part of the area without a dominate orientation. We note that the polarization direction of SWS data of TCU 120 differs significantly from those of other nearby stations and is approximately parallel to the direction of the Changhua fault strike. Figure 1.18 shows measured SWS data in the area covers TAI and ILA station groups. Most of SWS measurements obtained from this area are associated with ILA050. However R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 0.1 sec 24.; 24.6 ^ 24.4 C h elu n g p y fau lt/ ~ 24,2 23.; 23.6 1 2 0 121.5 120.5 121 longitude (°) Figure 1.17. Measured SWS parameters (bars) superimposed on stations in the area covers TCU station group (Here only the character T is plotted instead o f TCU). The bars are oriented paralleled to the PD o f the FSW and scaled by the TD. the measured SWS parameters from ILA050 scatter greatly. The average polarization direction of SWS is oriented in NW direction. Other sporadically distributed measurements in Taipei basin and Ilan basin of the area have a similar orientation. We note that several measurements from TAI032, ILA025 and ILA067 are oriented in NE direction. These stations are located in central mountain range. Figure 1.19 shows SWS measurements in the area covers KAU and TTN station groups. We can find that several measured polarization directions of SWS from KAU038, KAU039 and KAU042 are oriented to NW or NWW direction and conform to local GPS motion vectors as shown in Figure 1.15. The measured 28 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 25.2 25.1 24.9 € 24.: 1 (1 3 (1 I S P 24.7 24,6 24.5 24.4 0.1 sec 24,3 121 121.1 121.2 121.3 121.4 121.5 121.6 121.7 121.8 121.9 122 longitude (°) Figure 1.18. Measured SWS parameters (bars) superimposed on stations in the area covers TAI and ILA station groups. (Here only the characters T and I are plotted instead o f TAI and ILA respectively). The bars are oriented paralleled to the PD o f the FSW and scaled by the TD. polarization directions of SWS from stations around the Central Range, Chihshang and Lichi faults are parallel or sub-parallel to the strike directions of these faults. Figure 1.20 shows SWS measurements in the area covers HWA station group. Again, several measurements from station nearby the Costal Range fault display fault-parallel polarization direction. As shown in the inset of the figure, Many reliable measurements from stations HWA009 and HWA012 show a NNE polarization direction of SWS, while those from stations HWA010, HWA011, HWA013 and HWA014 display a NNW polarization direction of SWS. These two 29 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 23.4 Central RaKgfyfaulf / ( #" ' \ j / / 23.2 22 .: Chih^hang fault/ € Xichi fault JT037 22.4 22.2 0.1 sec 120.2 120.4 120.6 120.8 121 121.2 121.4 121.6 longitude (°) Figure 1.19. Measured SWS parameters (bars) superimposed on stations in the area covers KAU and TTN station groups. Here only the characters K and T are plotted instead o f KAU and TTN respectively). The bars are oriented paralleled to the PD o f the FSW and scaled by the TD. set o f polarization direction probably are associated w ith com p lex tectonic stress or g eological structure due to transition o f the tw o subduction system s as discussed in the introduction. 1.4 Conclusions O bserved polarization directions o f SW S in general agree reasonably w ell w ith the apparent regional com pressional axis indicated by G PS velo city field s, focal m echanism analysis and geological observations. Som e m easurem ents from stations in the vicin ity o f the active faults display fault-parallel polarization directions. In 30 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 24.1 23.9 23.; •s 23.7 23.6 23.5 23.4 23.3 0.1 sec 23.2 120.7 120.8 120.9 121 121.1 121.2 121.3 121.4 121.5 121.6 121.7 longitude (°) Figure 1.20. Measured SWS parameters (bars) superimposed on stations in the area covers HWA station group (Here only the character H is plotted instead o f HWA). The bars are oriented paralleled to the PD o f the FSW and scaled by the TD. all areas, in w hich stations have sufficient reliable m easurem ents, the observed SW S parameter determ ined m ainly by station locations rather than even t locations. The observed polarization directions and tim e delays vary significantly w ith locations. The observed tim e delays show no depth dependence in depth range from about 4 km to 18 km various w ith area locations. B ased on these observations w e can conclude that crustal anisotropy in these areas is dom inated at sh allow depth. 31 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. CHAPTER II: SYSTEMATIC ANALYSIS OF SHEAR-WAVE SPLITTING IN THE AFTERSHOCK ZONE OF THE 1999 CHI-CHI, TAIWAN, EARTHQUAKE: SHALLOW CRUSTAL ANISOSTROPY AND LACK OF PRECURSORY VARIATION Abstract We analyze shear-wave splitting (SWS) in a high-quality waveform data set recorded at surface and downhole (0.2 km) seismometers in a region around the September 20, 1999 Mw 7.6 Chi-Chi, Taiwan, earthquake sequence. The data set was generated by events in a 5 year period before, during and after the mainshock. The purpose is to investigate the depth extent of stress-induced crustal anisotropy and its possible temporal evolution in relation to the occurrence of large earthquakes. Results from downhole records show a stable polarization direction of the fast shear wave which matches well the local GPS velocity field. A slightly different polarization direction of the fast shear wave is obtained from surface data. This suggests a possible anisotropy change between the top 0.2 km structure and the deeper section of the crust. Measured time delays below the downhole station have an average value 0.16 sec without systematic changes for sources from about 8 km to 20 km in depth. Estimates of time delays in the top 0.2 km of the crust based on shear waves reflected from the free surface give a constant 0.04 sec. A likely depth distribution inferred from these two types of measurements and an S-velocity model indicates that the crustal anisotropy in the region is dominated by the top 2 - 3 km. The measured polarization directions and time delays give essentially constant values 32 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. over the 2.7 year pre-seismic and 2.3 year post-seismic periods in the region adjacent to the Chi-Chi rupture and within 10 km from the epicentral region of its two large M > 6.0 aftershocks. Analysis of SWS in waveforms produced by earthquake multiplets confirms further the lack of precursory temporal variations of crustal anisotropy in the immediate neighborhood of the Chi-Chi earthquake sequence. The results raise doubts on the general usefulness of SWS measurements for earthquake forecasting. 2.1 Introduction The seismic anisotropy of the upper crust has been studied extensively, mostly based on analysis of shear-wave splitting from local earthquakes. The commonly used explanation for the upper crust anisotropy is the assumed “extensive dilatancy anisotropy” (Crampin, 1978) or its modified version “anisotropic poroelasticity” (Crampin and Zatsepin, 1997; Zatsepin and Crampin, 1997). The model ascribes the upper crustal anisotropy to a preferred orientation of vertical, fluid-filled microcracks aligned in a direction controlled by the in-situ stress. In this hypothesis, shear waves with particle motions parallel to the plane of the cracks travel faster than those polarized in the orthogonal direction. Fluid-filled microcracks, oriented parallel to the maximum horizontal compressive stress direction ( crH), will preferentially remain open and it is thus expected that the polarization direction of the fast shear wave will be parallel to a H . Other possible causes of shallow anisotropy include alignments of microcracks or minerals in fault vicinities (Leary et al., 1987; Zhang 33 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. and Schwartz, 1994; Zinke and Zobak, 2000) and rock fabric anisotropy due to preferred mineral orientation (Kern and Wenk, 1990; Aster and Shearer, 1992). In this study, we analyze SWS in seismic waveform data recorded by a downhole short-period station CHY and a surface strong-motion station CHY073 located in the vicinity of the Chelungpu fault associated with the September 20, 1999 Chi-Chi earthquake (Figure 2.1). Numerous aftershocks (with two having M > 6.0) occurred just beneath the stations about one month after the Chi-Chi mainshock. The purpose of the study is to investigate the depth extent of stress-induced crustal anisotropy and its possible evolution in time in relation to large earthquakes. High-quality borehole records are used to measure SWS more precisely than can be done with surface records. Clearly recorded surface reflections in downhole data along with surface records allow us to estimate the magnitude of near-surface crustal anisotropy. The aftershock sequence in this region lasted more than two years and there are also many recorded events during a 2.7 year period before the Chi-Chi mainshock. We analyze SWS in this data set using both the aspect ratio method (Shih et al., 1989) and the cross-correlation method (e.g., Fukao, 1984). The results indicate that the crustal anisotropy in this region is dominated by the top 2 -3 km and that the top 0.2 km accounts for about 20% of the total observed SWS time delay in the upper crust. The results do not show systematic pre-seismic or systematic post-seismic temporal changes of SWS parameters in the source regions of the examined large earthquakes over the 5 year study period. 34 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 2.2 Data Set and Geologic Background Modern digital seismic monitoring in Taiwan began in the early 1970s. The Taiwan Central Weather Bureau Seismic Network (CWBSN) now operates 75 telemetered stations. Each station has short-period and strong motion 3 component sensors. A few of these stations are installed in boreholes reaching to about 0.2 km in depth. Since the beginning of the 1990s, the Taiwan Strong-Motion Instrumentation Program (TSMIP) added more than additional 650 free field stations with modern digital instruments (Shin and Teng, 2001). The Chi-Chi earthquake sequence was unusually energetic, with many M > 6.0 aftershocks, two of which occurred in the study area. The parameters of these two events and the mainshock are listed in Table 2.1. The locations of the mainshock (larger solid star), two large aftershocks (smaller solid stars), and other aftershocks (open and solid circles) are shown in Figure 2.1. The solid triangle indicates the location of short-period stations. One of these stations, CHY, is installed in a 0.2 km deep borehole. There is also a surface strong-motion station, CHY073, in the same location as CHY. The sampling rate of strong-motion seismograms is 100 samples/sec (sps), while the short-period records of CHY are sampled at 50 sps. The data used in this study extends from January 1997 to March 2002. Our study region is on the eastern boundary of the west coast Holocene alluvium plain, southwest of the southern end of the Chelungpu Fault (CLF). The Meishan fault (MSF), a strike-slip fault associated with the 1906 M l earthquake, is located at 35 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. the northern boundary of the study region. The Chukuo fault (CKF), another well- known active structure, is about 10 km to the east of the region. The 1999 Chiayi earthquake sequence, often collectively referred to as the Chi-Chi earthquake sequence, was associated with a blind fault (dipping 45° west from 8 km down to 15 - 17 km depth (Rau, 2002)). Observations from hydro-geological drilling reveal that the fine-, medium- and coarse- grain sandstones and gravel beds interfinger to form the top 200 - 300 m crust of the study area. OM3D Figure 2.1. A location map o f the study region with the Meishan fault (MSF), the Chelungpu fault (CLF) and the Chukou fault (CKF). Solid triangles indicate short-period stations including a 200 m deep downhole station CFIY. Solid stars represent the September 20, 1999 M w 7.6 Chi-Chi earthquake and its two large aftershocks. Solid and open circles represent other small aftershocks recorded by the borehole station CHY and the strong- motion station CHY073, respectively. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. Table 2.1. The parameters of the 1999 Chi-Chi earthquake and its two large aftershocks. Date Time (UTC) Lat. (N) Long. (E) Depth (km) Ml 1999/09/20 17:47:15.85 23.852 120.816 8.00 7.30 (7.6 Mw) 1999/10/22 02:18:56.90 23.517 120.423 16.50 6.40 1999/10/22 03:10:17.46 23.533 120.431 16.74 6.00 2.3 Analysis Methods 2.3.1 Shear-Wave Splitting Measurement Several methods have been used in SWS analysis. Visual inspection methods (e.g., horizontal particle analysis) have problems of subjectivity or observer bias that might affect individual and overall measurements. However, visual inspection methods are useful for checking results. Objective automated methods are required to analyze large data sets effectively, in an unbiased and systematic manner. The cross correlation (CC) method is commonly used for SWS analysis (e.g., Fukao, 1984; Bowman and Ando, 1987). It determines simultaneously the polarization direction (PD) of the fast shear wave (FSW) and the time delay (TD) of the slow shear wave (SSW) by comparing the cross correlation coefficients (CCC) of the two horizontal seismograms with various projections. The aspect ratio (AR) method proposed by Shih et al. (1989) calculates the PD of the FSW from the waveforms before the arrival of the SSW, and then the TD is determined by comparing the FSW to the SSW. 37 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. In general, particle motion for simple shear wave splitting is elliptical but it will be linear under one of the following two conditions: (i) one shear wave component is zero, or/ii) the phase difference between the two components is nn. Following Shih et al. (1989), the AR is defined as n - 1 Z ^ | c o s ( 0 , ~ e i )I 2V ,|sin(0, -e ,)| /= i where n is the total number of samples at equal time steps, Qj is the angle between the fast trial axis and X axis, and j is the index of the angle increment. The parameters e, and di are e,= tan" — — cjzLL (2.2a) (*, -*,--i) with i being sample number and d, = -T m )2 +(x, - V i ) 2 (2.2t>) where x and y are two orthogonal velocity time series. The maximum value of the aspect ratio in a time interval that contains only the early shear-wave arrival indicates the most linear particle motion - condition (i) and its direction indicates the PD of the FSW. The ideal time window for the above aspect ratio calculation should start at the onset of the FSW and end right before the arrival of the SSW. While it is not easy to determine such a window, multiple windows R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. with different sizes can be tested to find a suitable one. Shih et al. (1989) also determine the TD by maximizing the aspect ratio as a function of time lag based on condition (ii). Here we estimate the TD by cross-correlating the FSW with the SSW waveforms as determined by condition (i). Following Gledhill (1991), the measure error of the PD of the FSW is estimated by ± tan-1 (1 / AR) . The confidence interval estimation for the cross-correlation method is discussed in Rau et al. (2000). We adopt the same scheme to estimate the measure errors for all parameters calculated in this study with a 95% confidence level. Figure 2.2 illustrates an example of SWS analysis using the AR method. Figure 2a shows original three-component seismograms from downhole velocity sensors at CHY, generated by an M 1.88 event at 10.70 km depth with 1.70 km epicentral distance (see information above the figure). The shaded areas indicate the time window for data analysis chosen to bound the direct shear wave signal. We estimate the PD of the FSW, which in this example gives 170° ± 1°, from the maximum value of the aspect ratio shown in Figure 2.2b. Then we rotate the horizontal seismograms into the determined PDs of the FSW and the SSW (Figure 2.2c). The cross correlation coefficients (Figure 2.2d) of the two split seismograms are used to determine the TD of the SSW, which in this example gives 0.16 ± 0.01 sec. Figure 2.2f shows the horizontal particle motion of the original seismograms. The abrupt change in the direction of particle motion starts at the 16th data point (indicated by 39 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. (a) 23.49 N 120.44 °EM1.88 10.70km 1.70km CHY 00 13021948 ( c ) 500 0 -500 500 0 -500 (e) 500 -500 500 -500 500 -500 19 19.5 20 20.5 21 21.5 22 22.5 time (sec) 19 19.5 20 20.5 21 21.5 22 22.5 time (sec) 500 fast slow (shift) -500 20.2 20.4 20.6 20.8 (0 (d ) 50 100 150 azimuth (°) ,(0.16 sec) 0.5 -0.5 0.5 -0.5 0 time (sec) (g) 500 0 -500 500 -500 0 500 I -500 -500 0 500 time (sec) fast Figure 2.2. An example of SWS analysis with the aspect ratio method using short- period data, (a) Original three-component seismograms. (b) A contour of the AR values. The maximum value indicated by the cross corresponds to the PD of the FSW (170o in this example), (c) Seismograms rotated to the fast and slow PDs. (d) The CCC between two split shear waves determined by (b). The maximum value is related to the TD (0.16 sec in this example), (e) Waveforms of the fast and slow components shifted with the measured TD. (f) The horizontal particle motion of the original seismograms. (g) The horizontal particle motion of the fast and slow components shifted with the measured TD. an arrow). As a check, we advance the SSW with the estimated TD and plot the shifted seismograms in Figure 2.2e and their horizontal particle motion in Figure 2.2g. The linear particle motion in Figure 2.2g and well-matched seismograms in Figures 2.2e indicate that the above measurement is valid. 40 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. (C) 1.5 0 -1.5 1.5 0 -1.5 ( e ) 23.50 °N 120.45 °EM 2.83 11.99km 2.93km (a ) 1.5 0 1.5 1 0 1 1.5 0 1.5 (b) 2 1 time (sec) 2 1 time (sec) 2 fast slow(shift) 1 0 • 2 20.5 21.5 2 1 ( f > 0.1 (d ) CHY073 99 11121927 IIIIS (1 7 7 °) -tj time (sec) EW 100 azimuth (°) 150 time delay (sec) (g) 1 0 1 2 • 2 0 fast 1 0.5 0.5 o o f c o o g -0.5 o (0.20 sec) -0.5 0 0.5 1 0 ■ 2 0 2 Figure 2.3. An example o f SWS analysis with the cross-correlation method using strong- motion data, (b) A contour o f the CCC values. The azimuth and time delay o f the maximum CCC value are indicated by the cross and give the PD o f the FSW (177o in this example) and the TD (0.20 sec in this example), respectively, (d) A cross section o f (b) at the resolved PD o f the FSW. (a), (c), (e), (f), and (g) correspond to respective panels in Figure 2.2. Figure 2.3 illustrates the SWS analysis example using the CC method. The original strong-motion data are given in Figure 2.3a and the contours of the CCC are shown in Figure 2.3b. The maximum CCC value is indicated by the cross and its corresponding azimuth and time delay give the PD of the FSW (here 177° ± 16°) and the TD (here 0.20 ± 0.02 sec). Figure 2.3d is a cross section of Figure 2.3b at the determined PD of the FSW. The other panels in Figure 2.3 correspond to those of Figure 2.2. The nearly linear particle motion in Figure 2.3g and well-matched 41 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. -12171327 M1.86 11,60km 0.90km 200 0 -200 400 200 0 -200 -400 50 UD 0 -50 200 0 -200 fust 2 0 0 0 -200 slow 8.5 9 9.5 10 10.5 11 11.5 12 12.5 13 Time (see) 5 < X ) 0 -500 1000 0 •1(X)0 500 0 -500 5 < X ) PD 173+4" TD 0.15+0.02 sec 0 -500 5(X ) slow 0 -500 1000 0 • 1000 T 1070624 M2.28 11.45km 1.80km 1000 0 -I (XX) 2 0 0 UD 0 -200 I (X)0 PD 171 +1° TD 0.17+0.01 sec 1'asl 0 1000 0 •1000 slow ' 13.5 14 14.5 15 15.5 16 16.5 17 17.5 18 Time (sec) 1000 10230434 M2.59 15.12km 2.60km 0 -KMX) 500 0 -500 400 2 0 0 0 -200 -400 500 0 -500 500 slow 0 -500 12 12.5 13 13.5 14 14.5 15 15.5 16 16.5 Time (sec) 9.5 10 10.5 11 11.5 12 12.5 13 13.5 14 Time (see) Figure 2.4. Four examples o f three-component seismograms in the downhole station CHY. The events number, magnitude, focal depth and epicentral distance are indicated in the upper left corner o f each panel. The waveforms o f the resolved fast and slow shear waves, together with determined PD and TD, are given in the bottom o f each panel. Here P ' and P indicate up-going and down-going P waves, while S' and S indicate up-going and down-going S waves. seismograms (with opposite polarity phases) in Figure 2.3e again indicate a valid measurement. 42 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. Free Surface A CHY073 A CHY Reflected Waves Direct W aves Figure 2.5. Schematic geometry for the direct and surface-reflected waves in the borehole configuration. 2.3.2 Estimation of Near-Surface Anisotropy from Borehole Data Borehole-recorded seismograms at CHY clearly show (Figure 2.4) a direct up-going phase (S 1) and a surface-reflected down-going phase (.S' ). Figure 2.5 illustrates schematically the geometry of the direct and surface-reflected waves in the borehole configuration. After the PD of the FSW is determined as discussed in the previous section, we project the horizontal seismograms into the fast and slow components as shown in the inset of Figure 6. The shaded regions indicate the direct (S ' ) and reflected (S ') phases, respectively. It is expected that the FSW and SSW travel to the free surface and are reflected back to the downhole receivers with separate velocities. We calculate the autocorrelation coefficients of both the FSW and SSW 43 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 1.4 fast slow fast 500 1 .2 -500 1000 slow 0 .8 -1000 13.5 14.5 15.5 tim e (sec) 0 .6 0.4 0 .2 0 ■ 0 .2 •0.4 -0.5 0 0.5 1.5 t (sec) Figure 2.6. Measurement o f TD in the top 0.2 km o f the crust from the autocorrelation coefficients o f the resolved FSW and SSW. The inset shows the seismograms o f the FSW (upper panel) and the SSW (lower panel). The shaded regions represent direct phase (S') and reflected phase (5 ), respectively. The d tf and dts values represent two-way travel times in the top 0.2 km o f the crust o f the FSW and SSW, respectively. and plot them in Figure 2.6. The secondary peaks of the autocorrelation coefficients of the FSW and SSW correspond to the best matches between the S 1 and S K phases and their time lags are indicated by dt f and dts respectively. Therefore, dtf and dts give the estimated two-way travel times between the receiver and the free surface for the FSW and SSW, respectively. We 44 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. can thus calculate the TD of the SSW from the borehole receiver to the free surface within the top 0.2 km crustal structure as The above estimate is based on the assumption that the PD of the FSW above the borehole station is the same as that below. Since the estimated TD is not very sensitive to the PD, the measured results will not be appreciably affected by slight variations of the PD. 2.4 Data Processing and Uncertainty in SWS Measurements Figure 2.4 shows four examples of three-component records of the downhole station CHY. The events number, magnitude, focal depth and epicentral distance are indicated in the upper left corner of each panel. These records are velocity seismograms band-pass filtered from 1 to 15 Hz. The estimated PD of the FSW and the TD are given above the 4th trace in each panel. We show the resolved FSW and SSW in the lower part of each panel for all four examples. The effect of S-P waves conversion at the ground surface can be avoided using steep rays within the shear- wave window (Nutti, 1961; Booth et al, 1985 ). The latter is defined as having ray paths with incident i0 angles less than sin ~'(V, /V ). For typical values of Vsmd Vp near the top of the crust, i0 « 35°. Considering the high velocity gradient near the surface, we use events with hypocentral distance less than focal depth to assure that (2.3) 45 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. their ray paths lie within the shear-wave window. The observed effect of S-P wave conversion is negligible for downhole records; however, we find that the incident angle does affect the accuracy of measurement. A total of 401 events within the shear-wave window satisfying the above criteria are used in this study. Of these, 375 are recorded by the downhole station CHY while the others are recorded by the surface strong-motion station CHY073. The strong-motion acceleration data are also band-pass filtered from 1 to 15 Hz, and then integrated to velocity seismograms, which are used to measure shear-wave splitting. We employ both the CC and AR methods to analyze all these short-period records and show the results in Figure 2.7. We rejected measured results with a CCC value less than 0.7. The open and solid circles give analysis results of 334 events based on the CC and AR methods, respectively. In general, both methods produce similar results. The average differences of the measured TD and PD between these two methods are 0.001 sec and 5.2° , respectively . This indicates that the measurements are robust because the AR and CC methods use different portions of seismograms in the PD calculation as discussed in the previous section. Figure 2.7 shows that the spread of the PD estimates of the FSW based on the AR method is smaller than that of the CC method, while both methods have a similar spread of measurements of the TD. We thus use in the following sections results from the AR analysis. 4 6 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 0.25 0 .2 e o ^ 0.15 o o 0.05 aspect ratio 240 2 2 0 2 0 0 £ 180 O o £ 160 o © 140 1 2 0 aspect ratio Figure 2.7. Measured PDs o f the FSW and TDs from the AR method (solid circles) and the CC method (open circles). In many cases, the coda of shear waves makes it difficult to recognize the SSW arrivals, introducing an important source of ambiguity in TD measurements (Aster et al, 1990). However, even without significant effects from coda waves there is still a large scatter in our measurement results. Figure 2.8a shows the TDs versus incident angles represented by the ratio between epicentral distance and focal depth. We see clearly that the results from larger incident angles display larger scatter. For those events with epicentral distance over focal depth less than 0.2, the measured TDs are within the range 0.13 to 0.20 sec. Figure 2.8b indicates that results for events with 47 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. epicentral distance / depth m agnitude 23.4 23.45 23.5 23.55 Latitude ( 0 ) 120.4 120.5 Longitude ( 0 ) 120.6 Figure 2.8. (a) TDs versus incident angles of ray paths represented by the ratios between epicentral distance and event depth, (b) TDs versus magnitudes, (c) The distribution of three categories of TDs in the NS profile, (d) The distribution of three categories of three categories of TDs in the EW profile. magnitude between 2.4 to 3.0 have less scatter. This is possibly due to a relatively higher signal level for these records. The measurement accuracy is affected by both the AR and CCC values. As shown in Figure 2.7, the scatter for both TDs and PDs gradually decrease as AR values increase. The TDs are restricted to the range 0.12 to 0.20 for measurements with AR > 50. The CCC represents the degree of similarity between the FSW and SSW. Factors that influence the similarity between the FSW and SSW may include scattering, attenuation, dispersion and vertical variations of 4 8 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. anisotropy. Figures 2.8c and 2.8d show the relations between hypocenter positions and TD measurements. We divide the measured TDs into the following three categories: (1) larger than 0.17 sec, (2) between 0.14 and 0.17 sec, and (3) less than 0.14 sec. Each category is distributed approximately over the entire study region. In other words, there still exists a scatter in the measured TDs from similar paths. In addition to measurement uncertainty, there may be an inherent scatter in the TDs due to inhomogenous distribution of the crustal crack density. However we can not identify heterogeneous patterns of the crack density distribution due to the limited resolution. Such inhomogeneity will not affect the PDs of the FSW, which are controlled by the in-situ horizontal stress. The PD measurement is also insensitive to coda “contamination”. Thus the measured PDs of the FSW have less sources of error and scatter. 2.5 Near-Surface Crustal Anisotropy Figure 2.9a shows the estimated TDs of split shear waves over the depth range 0 - 0.2 km from surface-reflected waves. The mean and standard deviation of TD values are 0.04 ± 0.003 sec. We thus have derived two sets of TDs from the same set of borehole records. The TD for the section deeper than 0.2 km (> 0.2 km) estimated from the direct phased7 is related to the path from the source to the borehole station CHY. The TD for the depth section 0 - 0.2 km (0 - 0.2 km) measured from the reflected phase S 7 characterizes the path from the borehole station CHY to the free surface. Due to attenuation, dispersion and coda contamination, the individual 49 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 70 60 50 40 30 2 0 1 0 0 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 TD (sec) 0.08 0.06 (J © ©^ p 0.02 0 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2 0.22 0.24 TD (>0.2km) (sec) Figure 2.9. (a) Histogram of measured TDs in the top 0.2 km of the crust, (b) Measured TDs in the top 0.2 km versus measured TDs in the deeper section. analysis error of TD from the reflected phase S s is much larger than that from the phase S'7. However, in contrast to the large scatter that exists in the measured TDs for > 0.2 km, the estimated TDs for 0 - 0.2 km are much more consistent. Figure 2.9b shows TDs for > 0.2 km versus TDs for 0 - 0.2 km. We find that there is no correlation between these two sets of TDs. This implies that the sources that cause a broader scatter in the measured TDs for > 0.2 km do not affect the measurement of TDs over 0 - 0.2 km. Since the ray paths should be nearly vertical in the range 0 - 0.2 50 - (a ) _ i-------------- _!_____________ l_ r f l _l_____________ I _____________ l_ R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. Borehole Station CHY 23.50 °N 120.42 °E-173m EL 90 200 1 0 0 240" 270 PD (° ) 1 2 0 1 0 0 0.05 0 .1 0.15 0 .2 0.25 0.3 TD (sec) Surface Station CHY073 23.50 °N 120.42 °E + 17m E L 90 20 24(T 270 PD ( ° ) 1 0 8 6 4 2 0 0.05 0.1 0.15 0.2 0.25 0.3 TD (sec) ( d ) □ Figure 2.10. A comparison o f measured results from downhole records with those from surface observations, (a) Rose diagram o f the PD o f the FSW measured in the downhole station CHY. (b) Rose diagram o f the PD o f the FSW measured in the surface station CHY073. (c) Histogram o f measured TDs from downhole records, (d) Histogram of measured TDs from surface records. km, and all reflected waves should go through a nearly identical path, we expect to have a consistent measurement of TD over the range 0 - 0.2 km. The results suggest again that the scatter of TDs results mainly from inherent heterogeneity of anisotropic medium rather than measurement uncertainty. We find that the TD per kilometer of travel distance is 0.04 sec over top 0.2 km crust, or 0.2 sec/km. Given the quality and quantity of the downhole data used in this study, the results 51 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. demonstrate clearly that highly fractured rock in the near-surface crust plays a dominant role in seismic anisotropy observed at the surface. At the surface location above station CHY, there is a strong-motion FBA station CHY073. Figures 2.10a and 2.10b show rose diagrams of the PD of the FSW from borehole and surface data, respectively. Figures 2.10c and 2.10d show the distributions of TDs based on the borehole and surface measurements. The difference between the peaks of the distributions is comparable with the TD measured from the reflected phases of the split shear waves. The results of Figures 2.10a and 2.10b indicate that there is a discernible PD difference between measurements based on surface and downhole records. The mean and standard deviation of PD measurements from borehole data are 170.2°± 4.1°, while from surface records they are 177.5° ± 4.3°. The orientations of the downhole sensors are determined by a downhole gyroscope at the time of their installation (by Teledyne/Geotech). The precision of these orientations is about 2 degrees according to installation records and the gyroscope instrumental specifications. Therefore, it can be ruled out that the PD difference results from the orientation uncertainty of the downhole sensors. Interpretations of SWS parameters in the presence of two anisotropic layers are discussed by several papers (e.g. Silver and Savage, 1994; Wolfe and Silver, 1998). Since there are insufficient surface data to resolve these different anisotropic layers, 52 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. we adopt the following scheme to help interpreting the SWS parameters derived from surface data, (i) We use the observed EW seismogram produced by event 99- 11121927 and shown in Figure 2.3(a) as the original waveforms, (ii) We project the assumed original seismograms into the FSW and SSW polarization directions of the lower layer in an assumed two-anisotropic-layer model (see caption of Figure 2.11). (iii) We advance the waveform of the FSW with the assumed TD of the lower layer. (iv) We perform the similar steps to the obtained waveforms of the FSW and SSW using the assumed PD and TD of the upper layer and get calculated waveforms ( Figure 2.11(a)) corresponding to the two-anisotropic-layer model, (v) We analyze the SWS parameters from these generated data and compare the results with the parameters of the assumed model. The analysis results indicate that the estimated TD (Figure 2.11 (d)) is the sum of the TD values of the two layers and the estimated PD (Figure 2.11 (b)) is close to that of the upper layer. We also note that the shape of the resolved slow component is very similar to that of the original observed data given in Figure 2.3 (c) and (e). As shown in those figures, in both cases there are small phases (indicated by arrows) before the arrival of the slow shear wave that are not observed in the downhole seismograms (Figures 2.1 (c), 2.1 (e) and 2.4). The earlier small phases are caused by the small vertical variation of anisotropy. As discussed earlier, the PDs from the surface observations represent primarily the PD in the upper layer, while the PDs from the downhole records give an estimate of the PD in the layer below the downhole station. The results imply that the anisotropy in the top 0.2 km of the crust differs from that of the deeper section. The average downhole PD 53 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. (c) 2 19 (e) c a lc u la te d d a ta 2 0 • 2 2 0 •2 2 0 2 20 21 tim e (se c) V 20 21 tim e (se c) 22 23 2 0 ■ 2 20.4 20.6 20.8 21 21,2 21.4 (b) CHY073 99 11121927 (178°) - 1 o (d) 50 100 a z im u th ( ° ) tim e (se c) 0.6 0,4 0.2 0 - 0.2 -0.4 150 0.5 u o f c o 8 (0.20 sec) 0.5 0 0.5 0 1 - 2 0 2 fast Figure 2.11. SWS analysis of the calculated data based on a two-anisotropic-layer model using the CC method. The model parameters are PD = 170° and TD = 0.16 sec in the lower layer and PD = 177° and TD = 0.04 sec in the upper layer, (a) Waveforms generated from the EW component of observed seismograms (Figure 3a) using the assumed model. Panels (b), (c), (e), (f), and (g) correspond to respective panels in Figure 2.3. of 170.2° is in good agreement with the direction of maximum compressional stress based on the regional GPS data (Yu et al., 2001). The small change of PD in the surface data is possibly associated with a transition from a highly-fractured loosely- cohesive material in the top 100 - 200 m of the crust to a more competent and cohesive rock below or short-scale fluctuations in the near-surface stress field. Aster and Shearer (1991, 1992) analyzed the downhole and surface PDs and TDs 54 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. using a similar experimental geometry in the Anza region of southern California with downhole arrays and local earthquake records. They noticed a PD change between the downhole and the surface receivers in station KNW-BH. They interpreted the anisotropy observed at this station as due to a fixed palaeostrain alignment of anisotropic minerals and/or microcracks, and ascribed the vertical change of PD to a highly weathered near-surface zone. Coutant (1996) studied very shallow shear-wave anisotropy in local earthquake seismograms recorded by a vertical array of accelerometers at Garner Valley in southern California. The results were interpreted in terms of two superposed anisotropy layers: An igneous-rock anisotropy due to microcrack or mineral alignment below 220 m, and stress-induced anisotropy related to the San Andreas fault system above 220 m. Results on depth-varying crustal anisotropy were also found by VSP experiments (Winterstein and Meadows, 1991a, 1991b). Shear waves with very slow velocity (e.g., 500 m/sec) in the near-surface material can have several wavelengths (e.g., 4 wavelengths with 10 Hz over 200 m). In this case, surface observations will be strongly affected by the depth-varying anisotropy that appears to exist in commonly the near-surface region. 2.6 Depth Distribution of Crustal Anisotropy 2.6.1 Observation of Anisotropy in the Crust Deeper than 8 km Several studies have attempted to constrain the depth extent of anisotropy in the crust. Zhang and Schwartz (1994) claimed that the seismic anisotropy in the Loma 55 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 0.25 CD O r " o 0 ° o <d° 0.2 , 00, 42. 0.15 0.05 6 8 10 12 16 14 18 depth (km ) 200 190 ° w 180 & £ 170 °Q o Q Ph o o 160 150 140 6 8 10 12 16 18 14 depth (km ) Figure 2.12. (a) Measured TDs versus depth of events, (b) Measured PDs of the FSW versus the depth of events. The solid triangles give the average value over 2 km depth intervals. Preita rupture zone is no deeper than 2 km. Gledhill (1991), Peacock et al. (1988), and Savage et al. (1989, 1990) concluded that anisotropy in their study regions must be confined to the upper few kilometers of the crust to explain their observations of very different PDs obtained for stations located only a few kilometers apart. In contrast, Shih and Meyer (1990) observed increasing TDs with propagation distance in the South M oat o f the Long V alley Caldera in California, su ggestin g m ore 56 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. S-velocity (km/sec) 0 1 S-velocity 2 3 4 o. 5 travel tim e 6 7 8 0 3 1 2 4 5 travel tim e (sec) (b) 4 - 5 - 4 0 40 AO AO AO A O A O A O A O A 0 A O A O A O A ( A A A A A A o m odel (6a) a m odel (6b) h=h o o o o o o a o A o A o A o AO AO AO AO AO A O A O A 3 0.1 0.2 T D (sec) 0.3 Figure 2.13.(a) A model of S wave velocity from 0 to 8 km depth and corresponding travel times beginning from surface, (b) Inferred TD beginning from surface based on two models of k, two boundary conditions of k, and the S-velocity model of (a). pervasive anisotropy than only a few kilometers. Zinke and Zobak (2000) concluded that the shallow crust below the station does not influence the PD near the Calaveras fault in California. Li et al (1994) observed seismic anisotropy in the seismogenic layer beneath the Los Angeles basin. Due to the scatter inherent in TDs, it is often difficult to constrain well the depth extent of anisotropy. However, the high-quality seismic records observed in the downhole station CHY allow us to constrain it in our 57 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. study region accurately. Figure 2.12a and 2.12b show the TDs and the PDs of the FSW versus the depth of events. The solid triangles represent the average values over 2 km intervals and the associated vertical lines represent standard deviations. For instance, the first triangle on the left in Figure 12a gives the set of TD values from 6 to 8 km of 0.14 ± 0.04 sec. The results indicate that the PDs are very stable over the entire range from 6 km to 18 km, and that the TDs are reasonably constant over the range 8 to 18 km. We conclude that the anisotropy in our study area occurs over the top 8 km of the crust. In other words, there is no appreciable crustal anisotropy from 8 to 18 km. 2.6.2 Estimation of Anisotropy in the Crust Shallower than 8 km As shown in Figure 2.12a, the average value of TDs from several events with depth 6.0 - 8.0 km is smaller than that from the deeper section. We believe that the change in this range is mainly due to lack of sufficient data points. Unfortunately, no events are recorded with depth less than 6.0 km, and thus we cannot constrain the depth extent of anisotropy directly in the range from 0.2 km to 6.0 km. However, several lines of arguments can be used to infer that the observed crustal anisotropy is actually dominated by the top 2 -3 km. The depth extent of anisotropy may be estimated through the coefficient of anisotropy k , which is defined as (e.g., Tadokoro et al. 1999) R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. k = (Vfs- V J / V fs, (2.4a) where Vfs and Va. are the velocity of the FSW and the SSW, respectively. The coefficient of anisotropy k can also be calculated by k = ('/; - Th )/T (2.4b) where Tfs and Tss are the travel time of the FSW and SSW, respectively. We have two boundary conditions for k . In the range deeper than 8km, The values Tx s = 0.52 sec and Tfs - 0.48 sec are obtained from surface-reflected waves, associated with the TD measurement. The shear wave velocity in the top 0.2 km can be estimated from the measured Tfs, which gives 0.2 km / 0.48 sec « 0.42 km/sec. The velocities in the deeper crust were estimated from surface wave dispersion analysis by Chung and Yeh (1997). The station CHY is located at the northern boundary of their study area. We combine the shear wave velocity in the top 0.2 km measured in this study with their shear wave velocity model (Figure 2.13a). The surface wave dispersion data may not resolve the layer structures accurately. However, it gives a good estimate for the average velocity distribution over the depth (2.5a) and for the top 0.2 km, k = k0 = (Tu. - Tfi )/Tss = 0.04 sec/0.52 sec « 0.077. (2.5b) 59 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. range under consideration, w hich is im portant for the follow ing analysis. To estimate the depth distribution of TD data subjected to the boundary conditions (2.5a) and (2.5b), we assume two models that relate k to the depth h: where hc is a given depth above which the crustal anisotropy fallows a constant k0, and We calculate the travel time (beginning at the surface) using the shear-wave velocity model and plot the results in Figure 2.13a. Then we calculate the TD (again beginning at the surface) based on assumptions (2.6a), (2.6b) and plot the results in Figure 2.13b. The results for model (2.6a) indicate that the TD is distributed in the top 4 km. Since the actual k at depth is probably smaller than&0, this result gives an upper limit for the TD distribution. Considering a gradual closure of microcracks with increasing depth (due to increasing of confining pressure), model (2.6b) having a linear decaying k with depth is more realistic than model (2.6a). The corresponding TD distribution shown in Figure 2.13b is similarly more realistic. The TD gradually approaches the observed value of 0.2 sec when the depth reaches 8 km. (2.6a) km km . (2.6b) 60 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. Model (2.6b) is consistent with results of Boness and Zoback (2004) from the SAFOD pilot hole in Parkfield, CA. They found that the shear wave velocity anisotropy decreases with depth from about 10% to 2%. We note that, although the anisotropy may extend beyond 4 km, for both models (2.6a) and (2.6b) 65 - 70 % of the TD is distributed within the top 2 km and 20 % is in the top 0.2 km. This is not surprising since the travel time in the top 2 km (Figure 2.13a) contributes nearly half of the total travel time in the 8 km section of the crust. 2.6.3 Normalization by Travel Distance versus by Travel Time Due to the existence of low velocity layers, the shallow structure dominates the travel time and it also plays an important role in the TD distribution. In this regard, we can see that there is an inherent drawback in the commonly used method, in which the TD is normalized by travel distance to describe the degree of anisotropy. For instance, if our observed total 0.20 sec TD is normalized by 8 km, we get 25 ms/km. The travel time per kilometer in the deep layer with a shear-wave velocity of 3.29 km/s is about 0.30 sec and the corresponding k is 0.083. In contrast, the travel time per kilometer in the shallow layer with a shear-wave velocity of 0.90 km/sec is 1.11 sec and the corresponding k is 0.023. In this case, the consequence of the above normalization is that the coefficient of anisotropy k in the deep layer is 3.6 times the value of k in the shallow layer. Obviously, this could cause confusion in interpreting measured results. Based on our shear-wave velocity model, the total travel time in the top 8 km crust is 3.7 sec (Figure 2.13a). Thus, the average k = 0.2 sec / 3.7 sec = 0.057. In this case, the corresponding TDs per kilometer over the 61 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. deep and shallow sections are 0.017 and 0.063 sec, respectively, which are much closer to the real observations. Considering the strong variations in shear-wave velocity, normalizing the TD by travel time provides a more physical way of describing the degree of anisotropy. 2.7 Temporal Change of SWS Associated with Stress Changes Induced by Large Earthquakes Temporal changes of SWS parameters were reported before and after the M 6 North Palm Springs earthquake in southern California (Peacock et al., 1988; Crampin et al., 1990, 1991). However, Aster et al. (1990, 1991) applied an automatic technique and similar event analysis of SWS for the same data set used by Crampin et al. (1990) and did not identify temporal change in TD. Temporal changes in TDs before smaller earthquake were reported by other SWS studies (e.g., Booth et al., 1990; Gao et al., 1997; Liu et al., 1997). Crampin et al. (1999) claimed that they successfully ‘stress forecast’ an M = 5 earthquake in Iceland by using variations in TDs of the SWS. It is difficult to identify possible temporal changes from measured TDs with a large scatter. Since analysis of temporal variations often use TDs normalized by travel distance, inferred temporal patterns in TDs could be contaminated by different travel paths, i.e., spatial changes. In many cases it is difficult to measure unambiguously TDs due to the complex nature of shear-wave seismograms recorded at the surface. All these factors contribute to existing controversy regarding temporal changes of seismic anisotropy. 62 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 9/20/1999 Chi- Chi M7.6 -O s 0 X 0.2 S ( N 1 01 Q H 0 220 200 10/22/1999 Chiayi M6.4, M6.0 1 (a) “111 i .............i i ' i ' i i - r f -*- -it4 _ _i _ l r ~ ^ 1 y i f e " _i ■_ rrvte 2t _ < f? - _ JL _ i J 4 n ;# W v h $ y . .. . . 1 -1000 - (b) -800 -600 -400 -200 < T 180 I f - - o - i p - -1,1 i> - - # - 4 - - f - 4 j £ 160 t t _ 140 120 0.08 0.06 0.04 0.02 0 _J_ -1000 (c) -800 -600 -400 -200 ■ = p ,ir .di =<5hi5r£-5 _ _ ! *I ZT J _ \ ~ -1000 -800 -600 -400 -200 0 tim e (day) Figure 2.14.(a) Measured TDs in the crust below 0.2 km versus time. Arrows indicate the times o f the Mw 7.6 Chi-Chi mainshock and two M 6.4 and M 6.0 Chiayi aftershocks, (b) Measured PD o f the FSW below 0.2 km versus time, (c) Measured TDs in the top 0.2 km versus time. The mean values and standard deviations o f the results before and after the Chi- Chi mainshock are indicated by solid and dashed lines. As shown in Figure 2.1, station CHY is located off the southern end of the Chelungpu fault, on which the Mw 7.6 Chi-Chi earthquake occurred. The M 6.4 Chiayi earthquake and another large M 6.0 aftershock occurred just beneath the station about one month after the Chi-Chi mainshock. Thus, CHY is well situated to detect possible changes of SWS associated with crustal stress adjustment before, during and after the Chi-Chi mainshock and two nearby large aftershocks. Most of 63 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. the well-recorded events by CHY are aftershocks of those large earthquakes and there are also a number of events before the Chi-Chi mainshock. Figures 2.14a and 2.14b show the measured TDs and PDs of the FSW, respectively, in the crust below the 0.2 km deep borehole station versus time, while Figure 14c shows the calculated TDs in the crust above the 0.2 km deep station versus time. We average these parameters over two periods spanning 2.7 year before the mainshock and 2.3 year after it. The average values and standard deviations are represented by solid and dash lines, respectively. None of the results show systematic changes either before or after the Chi-Chi mainshock. In addition, the sets of PD values for depth > 0.2 km and TD values for the 0 - 0.2 km depth range show small variations over the entire period, while the set of TD values for depth > 0.2 km changes from 0.144 ± 0.028 to 0.161 ± 0.027sec at the time of the Chi-Chi mainshock. The observed apparent co-seismic temporal change of TDs for > 0.2 km is likely to be affected by spatial changes associated with different event locations before and after the Chi-Chi mainshock. Figure 2.15 shows the hypocenter locations of events before (triangle) and after (circle) the Chi-Chi mainshock. To estimate the possible impact of the different event locations on the TD values, we assign for each event before the mainshock the closest event after the mainshock. Fifteen pairs of such events with catalog distances less than 2 km are marked in Figure 15 with fdled symbols. The mean and standard deviation of TD values from these 15 pairs are 0.146 ± 0.031 before and 0.155± 64 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. & 14 ^ 23.6 23.55 120.55 120.5 23.5 120.45 120.4 23.45 120.35 23.4 120.3 latitude (°) Longitude (° ) Figure 2.15. Earthquake hypocenters before (triangles) and after (circles) the Chi-Chi mainshock. The solid symbols indicate event pairs separated by less than 2 km. 0.034 after the mainshock. The difference between those values is significantly smaller than that associated with the entire data set (solid line in Figure 2.14a). If we decrease the maximum separation distance to 1.5 km, the number of event pairs drops to 11 and the set of TD values are characterized by 0.146± 0.026 before and 0.151 ± 0.038 after the Chi-Chi mainshock. The results indicate that spatial variations of parameters affect strongly the observed co-seismic temporal changes of TDs for the section deeper than 0.2 km. The SWS analysis of clusters of earthquake doublets (multiplets) may be the most robust way to identify temporal variation of anisotropy. Doublets and multiplets are 65 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. sets of earthquakes with similar waveforms, and indications of similar hypocentral locations, focal mechanisms, and ray paths to the stations. Temporal changes extracted from analyzing cluster of earthquake doublets or multiplets should have less contamination due to spatial variations. Figurel6 shows a set of eight multiplets from borehole records at CHY spanning a time period from about 1.5 year before the Chi-Chi mainshock (02/25/1998) to seven months after it (04/13/2000). The pertinent parameters of these events and the measured TDs and PDs are given on the left of Figure 16b above the traces of the NS components. The arrows mark the traces right before the mainshock. Figures 2.16c and 2.16d display the resolved FSW and SSW of these events in a stacked form. The results from these similar events clearly demonstrate a lack of significant TD and PD changes. The well-synchronized reflected waveforms in Figure 2.16 also indicate that there are no significant changes in near-surface crustal anisotropy over this period (e.g., a change of TD larger than about 10 ms). An important result of this study is that there are no systematic changes of SWS parameters observed within the 2.7 year pre-seismic period of the Chi-Chi earthquake. The measured lack of temporal changes in anisotropy characterizes the vicinity of several large earthquakes during a highly active tectonic period, which involved a 100 km rupture with as much as 8 m surface slip and a very energetic aftershock sequence including many M > 6 events. The results do not support claims that SWS measurements can provide a general tool for forecasting impending large earthquakes. 66 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. EW 0.5 1.5 2.5 3.5 4.5 0 2 3 4 NS 991121 23.495_L20.441 11.48 171.00.16 991109 23.495 120.449 10.72 171.0 0.16 0.5 1.5 2.5 3.5 4 4.5 0 1 2 3 tim e (sec) fast 0.5 ■0.5 0 0.5 1.5 2 2.5 3 slow 0.5 -0.5 0.5 1.5 2.5 3 0 1 2 tim e (sec) Figure 2.16. Seismograms o f a cluster o f earthquake multiplets. (a) The EW components of seismograms. (b) The NS components o f seismograms. The time, longitude, latitude and depth o f each event, and the measured PD o f the FSW and TD, are indicated above each trace, (c) The FSW component o f seismograms in a stacked form, (d) The SSW component o f seismograms in a stacked form. 67 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 2.8 Conclusions We have analyzed systematically shear-wave splitting of seismic data recorded in the aftershock zone of the September 20, 1999 M W 1.6 Chi-Chi earthquake at a 0.2 km deep borehole short-period station and a surface strong-motion station. The analysis employs both the aspect ratio and cross-correlation methods. The results show that the two methods provide almost the same TDs and similar PDs of the FSW. Since the AR and CC methods use different portions of seismograms, this indicates that the measured SWS parameters are robust. We have also measured the TDs between the fast and slow shear waves in the top 0.2 km of the crust by analyzing surface- reflected waves in the borehole records. We found that the incident angle of a ray path might influences the accuracy of SWS measurements, even when the events are within the shear-wave window. The measured PD of the FSW in the crust below the 0.2 km deep borehole station is 170.2°± 4.1°. This matches very well the surface deformation field during 1992- 1999 based on GPS data (Yu, et al., 2001). The results indicate that the anisotropy in the section deeper than 0.2 km is controlled by local tectonic stress field. The mean and standard deviation of TD measurements in the top 0.2 km of the crust are 0.04 ± 0.003 sec. The analysis reveals strong near-surface (0.2km) anisotropy in the study region with a coefficient of anisotropy k = 0.077. The mean and standard deviation of PD measurements from the surface strong motion data are 177.5° ± 4.3°. The change of PDs between the downhole and surface observations implies that the 68 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. anisotropic axis of symmetry varies slightly with depth. The measured TDs in the crust below the 0.2 km deep borehole station show essentially a constant value of 0.16 sec for sources from 8 km to 20 km, implying a zero coefficient of anisotropy in this depth range. We use measured k values (zero below the 8 km and 0.077 at the top 0.2 km) as boundary conditions to infer on a possible depth distribution of crustal anisotropy based on an S-wave velocity model. The assumption of a constant k leads to the result that the anisotropy is limited to the top 4 km of the crust, while the assumption of a linearly decaying k with depth gives a slightly deeper extent of anisotropy. Both results indicate that the top 2 km of the crust produces 65 - 70 % of the total TD. The dominance of the shallow crust in shear-wave anisotropy is partly due to the existence of low shear-wave velocity layers. Considering the strong variation of shear-wave velocity over depth, we suggest that measured TDs should be normalized by travel time rather than by travel distance. The observed PDs of the FSW and TDs in both the near-surface and deeper crust, and the great similarity of waveforms generated by multiplets, show no appreciable systematic temporal changes over the 2.7 year period before the 1999 Mw 7.6 Chi- Chi mainshock and the 2.3 year period after. The observed apparent co-seismic temporal change of TDs is affected strongly by spatial variations of seismic anisotropy, and at most indicates increasing crack density associated with the mainshock. 69 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. We note that our data set is recorded in the vicinity of large rupture zones that experienced stress changes that are as big as expected to occur in the brittle crust during large earthquake cycles. Nevertheless, these stress changes apparently did not leave precursory or post-seismic temporal signatures that can be mapped by our SWS analysis of the observed high-quality borehole seismograms. The lack of such temporal changes of SWS parameters and the dominant influence of the shallow structure on the data suggest that the seismic anisotropy in our study region may be associated with a preferred closure of randomly-distributed cracks due to the existing anisotropic stress field. A similar cause of seismic anisotropy is suggested by Boness and Zoback (2004). Acknowledgments We thank Zhigang Peng for useful discussions and anonymous referee, Paul Silver and Andy Michael for many helpful comments. The research was supported by the National Science Foundation (grant EAR-0124926) and the Taiwan Central Weather Bureau (MTOC-CWB-93-E-06) 70 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. CHAPTER III: NEAR-SURFACE SEISMIC ANISOTROPY, ATTENUATION AND DISPERSION IN THE AFTERSHOCK REGION OF THE 1999 CHI-CHI, EARTHQUAKE Summary Seismograms from local aftershocks of the 1999 Chi-Chi, Taiwan, earthquake recorded at a 200 m deep downhole station CHY of the Taiwan Central Weather Bureau Seismic Network (CWBSN) have clear direct up-going shear waves and their surface-reflected down-going phases. Measurements of time difference between the direct and reflected phases of the fast and slow components of split shear-waves show approximately 8% velocity anisotropy in the top 200 m of the crust. The phase velocities extracted from the direct and reflected waveforms display clear evidence of attenuation-related dispersion. Taking the dispersion and geometrical spreading factor into account, we estimate the Q value of the shear waves by fitting calculated results to the observed reflected waveforms. The amplitude spectral density ratios between the direct and reflected phases are approximately linear within the frequency range 2-15 Hz. This allows us also to estimate the Q value from the slope of the amplitude spectral ratio (in dB/Hz) in this range. The estimated Q values with both methods, based on a set of similar waveforms and additional 156 high-quality records, are 61-68 for the fast components and 43 - 52 for the slow components. The observed attenuation anisotropy may be, similarly to velocity anisotropy, a manifestation of microcracks alignment and their response to in-situ stress. Strong 71 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. attenuation anisotropy (23 - 30% in this study) along with attenuation-related dispersion in the shallow crust can affect significantly the behaviors of shear waves and should be taken into account in studies employing surface and shallow borehole records of shear-wave waveforms. Key words: seismic anisotropy, attenuation, borehole, crust. 3.1 Introduction The heavily damaged material in the top few hundred meters of the crust, with a low overburden pressure, contains a high density of cracks, pores and other defects. The near-surface material can thus have a large influence, in term of attenuation, dispersion and anisotropy, on seismic recordings made at the surface. Many studies (e.g. Aster and Shearer 1991; Fletcher et al. 1990; Hauksson et al. 1987) employed borehole observations to investigate the seismic attenuation and anisotropy in the shallow crust, and their effects on surface recordings. In this paper we perform such a study using high quality seismograms from local aftershocks of the 1999 Chi-Chi earthquake recorded at a 200 m deep downhole station CHY of the Taiwan Central Weather Bureau Seismic Network (CWBSN). The borehole seismograms we use have clear direct up-going shear waves Su p and surface-reflected down-going phases Sdow n■ In a previous study based on this data set (Liu et al. 2004) we found a strong anisotropy of shear-wave velocity in the top 200 72 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. m, which contributes about 20% of the total shear-wave splitting (SWS) time delay in the upper crust. In the present work we employ two methods to calculate the quality factors (Q) of both the fast and slow shear-wave components determined from the previous SWS analysis of Liu et al. (2004). In the first method, the quality factor is estimated from the slope of the amplitude spectral ratio of the direct and reflected waves versus frequency. In the second method, we estimate the quality factor by comparing the observed reflected waveform with a calculated one, generated by applying an attenuation equation to the observed direct waveform. The quantity and quality of the recordings enable us to measure the quality factor reliably. The results show clear evidence of attenuation anisotropy in the near surface structure. The phase velocities extracted from the direct and reflected waveforms indicate the existence of attenuation-related dispersion. The inferred dispersion curves fit the theoretical logarithm dispersion equation (Aki and Richards, 2002) well. Because of the difficulty of isolating the reflected P wave phases, we only analyze the attenuation and dispersion properties of shear waves. 3.2 Data Set and Geological Background Modern digital seismic monitoring in Taiwan began in the early 1970s and at present the Taiwan Central Weather Bureau Seismic Network (CWBSN) has 75 telemetered stations (Shin and Teng 2001). One of these short-period stations, CHY, is installed in a 0.2 km deep borehole. The 1999 Chi-Chi earthquake sequence was highly 73 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. O M 3 0 Figure 3.1. A location map o f the study region with the Meishan fault (MSF), the Chelungpu fault (CLF) and the Chukou fault (CKF). Solid triangles indicate short-period stations including a 200 m deep downhole station CHY. Solid stars represent the September 20, 1999 Mw 7.6 Chi-Chi earthquake and its two large aftershocks. Solid circles represent other small aftershocks recorded by the borehole station CHY. energetic, with many M > 6.0 aftershocks, two of which and many other smaller aftershocks occurred in the area close to CHY. The data used in this study extends from January 1997 to March 2002 and the sampling rate of the employed seismograms is 50 sps. 74 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 92 91 90 89 TKTM 140 139 138 137 Figure 3.2. Pictures of drill core samples from a hydro-geological well #200201G1 in the study area. The total depth of the well is 250 m and the depths of core samples are indicated in the figure. (From the Hydrogeology Data Bank, The Central Geological Survey of Taiwan.) As shown in Figure 3.1, the borehole station CHY is located in the eastern boundary area of the west coast Holocene alluvium plain, southwest of the southern end of the Chelungpu Fault (CLF), which ruptured during the Chi-Chi main shock. The Meishan fault (MSF), a strike-slip fault associated with the 1906 M l earthquake, is located at the northern boundary of the study region. The Chukuo fault (CKF), 75 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. another w ell-know active structure, is about 10 km to the east o f the region. Observations from hydro-geological drilling reveal that the top 200 - 300 m crust of the study area consists of inter-fingered fine-, medium- and coarse - grain sandstones and gravel beds. Figure 3.2 shows drill core samples from a nearby hydro-geological well #200201G1 (The Hydrogeology Data Bank, The Central Geological Survey of Taiwan) that illustrate visually some characteristics of the near-surface material. The core samples from 88 m to 90 m are coarse- grain (0.50 mm - 1.00 mm) sandstone, and from 90 m to 92 m are medium- and fine- grain (0.125 mm - 0.5 mm) sandstone. From 136 m to 140 m they contain finer siltstones or mudstones These rock types are highly likely to produce microcracks under low confining pressure in the shallow crust. 3.3 Methods for Attenuation Analysis The amplitude spectrum A, ( / ) recorded at the ix h station for a given earthquake can be expressed as (Bath, 1974) A,(f) = G ,^ ( /) I S ,( / ) / , ( / ) e x p ( - ^ , / e v ) , (3.1) where G: is the geometrical spreading factor, K( /') is the source spectrum, S: ( / ) is the site response, /, ( / ) is the instrumental response, R, is the travel distance of the seismic wave, v is the average wave velocity and Q is the assumed frequency 76 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. independent quality factor averaged along the path. Because the travel times can be measured directly from the recordings, they can be used to determine interstation Q values without making assumptions about velocity structure. For two stations along the essentially the same ray path, we approximately have where tx and t2 are the travel times from the source to the first and second stations, respectively, and Q is the average quality factor along the path between these two stations. Here, the source spectrum k(f ) is eliminated, since the both recordings are from the same source. In this study, we adopt this method for downhole recordings in which reflected waves from the free surface are viewed as waveforms that are recorded by another virtual station. Figure 3.3 illustrates schematically the geometry of the direct and surface-reflected waves in the borehole configuration. Since the up- and down-doing phases are both recorded by the same physical station, we have Si ( / ) = S2 ( / ) and /, ( / ) = I2 (/). Thus, we can eliminate the site and instrumental response terms as well and get M f ) Gx 5 ,( / ) / ,( / ) r f { h - t x) (3.2) A2(f) G2S 2 ( / ) / 2 ( / ) Q 4*™ (/) G, ■ A U p (y ) g r f {td o w n t ) 7---- 777 = 77---- exP(------------- —) , (3.3) where subscripts ‘up’ and ‘down’ replace ‘1’ and ‘2’, respectively. We can develop the following two procedures for estimating Q based on equation (3.3): 77 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. Free Surface Reflected Waves CHY Direct Waves Figure 3.3. Schematic geometry for the direct and surface-reflected waves in the borehole configuration. (a) Amplitude spectral ratio method Taking denary (base 10) logarithms on both sides of equation (3.3), we get , . G ft ( fd o w n t u p ) r log— = log— — + loge — / (3.4)The first item in the right side of equation (4) is independent of the frequency / . Thus a plot of log Aup(f)/Adow n(f) versus/ gives a line with a slope m , from R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. w hich the Q value can be estim ated as Q = ^'if -tuP)K^ge\0-m) (3.5) A similar method is adopted by Hauksson et al. (1987) and Aster and Shearer (1991) for attenuation analysis in borehole experiments. (b) Waveform fitting For a layered velocity model, the geometrical spreading factor G for a shallow depth range is from ray theory approximately proportional to 1/7? , where R is the distance from source to receiver (Hauksson, 1987). Therefore, equation (3.3) can be rewritten as As will be discussed later, for a frequency-independent Q there must be a frequency-dependent phase velocity c(f ), which can be represented (Aki and Richards, 2002) as where /„ is a reference frequency. In this case we can combine the amplitude and -V... ( / ) = ~ ( / ) . ^dow n * 1 (3.6) = -------+ ------------log — , C(f) c(/o) nQc(f0) \ f ) (3.7) 79 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. phase spectra and get • X d o w n ( / ) = ~ ~ ~ X u p ( / ) exp[- down down R R, ’down u p X u P(f)ex P h down down ~7X~~ X u p ( / ) exp[- ^ / / / l l l f H down + ' 2 “^)]» down (3.8) where ( / ) and ^ r f o w „ ( / ) are the Fourier spectra of the direct waveform xu P (0 ar* d the reflected waveform xd o w n (t) , respectively In this study we use equation (8) as follows. We first calculate X up{f )from waveform xup( t). Then we calculateX k dow n(f) using equation (3.8) with an assumedQk (where k is the index of different trial values of Q). In a third step we calculate xk U o w r l ( / ) by inversely transforming X k d o w n ( / ) and compare the result with the observed waveform xd o w n (/). The fitting errors between the calculated and observed waveforms are defined as where N is the number of data points in the waveforms. The minimum value of (3.9) E(k) indicates the best fitting result and the corresponding Qk gives an estimate of 80 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 3.4 Results 3.4.1 Estimated Q Values From Staked Waveforms of Multiplets Multiplets are a set of earthquakes with similar waveforms, and by implication similar locations, focal mechanisms, and ray paths to the station. Several sets of multiplets are identified by cross-correlating the observed horizontal waveforms with each other for all 360 events. Figure 3.4 shows the stacked horizontal seismograms of a set of 7 multiplets. These seismograms are projected into the resolved fast and slow directions of shear-wave splitting (SWS), as determined by the detailed SWS analysis of Liu et al. (2004). As expected, the average waveforms for all these events display a high signal-to-noise ratio. The nearly ideal impulse-like waveforms of the shear-wave phases allow us to window the direct and reflected phases properly. A cosine taper is used to reduce the effect of data truncation (Kanasewich 1981). We calculate the amplitude spectra of the direct and reflected phases for both the fast and slow components and show them in Figure 3.5a. Since the data sample rate is 50 sps, the spectra are cut off at the Nyquist frequency range 2-15 Hz. We also present corresponding results based on the data prior to and following the first S wave (“boxes” on the seismograms of Figure 3.4). It appears that the amplitude spectra of the direct and reflected S phases are sufficiently larger than that of background signals within the frequency range 2-15 Hz. We note that the above amplitude spectra are obtained from stacked waveforms, which usually have a higher signal-to- noise ratio. The amplitude spectral ratios between the reflected and direct phases for 81 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. (a) fast 0.5 0.5 (b ) slow 0.5 0.5 3.5 4 5 2 2.5 3 time (sec) Figure 3.4. Stacked fast and slow components of horizontal shear waveforms for a set of 7 earthquake multiplets. The vertical shaded areas indicate portions of the seismograms used in followingattenuation and dispersion analyses, while and the horizontal boxes mark portions used for background spectral analysis in Figure 3.5. the fast and slow components are shown in Figure 3.5b. As discussed in section 3.3, we can estimate Q values of the fast and slow shear waves by fitting the observed amplitude spectral ratios versus frequency (in db/Hz) to equation (3.4). We fit the curves to the equation within the 2 - 15 Hz range and calculate Q values from the estimated slope m through equation (3.5). The measured values are Qf = 68 ± 8 for the fast shear wave and Qs = 52 ± 3 for the slow wave. 82 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. s fast up fast, dow n slow.up slow,down 10 signal prior to signal 3 10 following signal 4 1 0 ' o 1 0 * 10 frequency (Hz) (b) 10 o 10 10 15 20 0 5 slow Q = 52 10 o 10 20 10 15 0 5 frequency (Hz) Figure 3. 5. (a) Amplitude spectra of the direct and reflected shear-wave phases and amplitude spectra of background noises prior to and following the direct shear waves. Amplitude spectral ratios versus frequency and linear fitting to equation (3.4) in the frequency range 2 - 15 FIz R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. The main advantage of estimating Q values from amplitude spectral ratios is that we can avoid making assumptions on geometrical spreading factor and considering the effect of dispersion. However, the strong dependency of the estimated results on the employed frequency range reduces from the robustness of the measurements. Using the same waveforms, we also estimate Qf and Qs by fitting calculated reflected waveforms to the observed ones as discussed in method (b) of section 3.3. The fitting errors, defined by equation (3.9), with different trial Q values are shown in Figure 3.6a. As indicated by arrows in the figure, the minimum values of the fitting errors are associated with Q f =61 for the fast shear wave and Qs = 43 for the slow component, respectively. Figure 3.6b shows the corresponding best fitting waveform results. This method is not strongly affected by background noise since multiple reflections and scattering signals have much smaller amplitudes than the direct or free-surface reflected phases, and therefore contribute less to the fitting errors than those main phases do. As a consequence, the estimated results are insensitive to factors such as the fitting frequency range, and are therefore relatively robust. Figure 3.6a shows that the analysis of individual measurements is not very sensitive to small changes in the estimated Q value, in our cases with relatively high Q and corresponding relatively small attenuation. Nevertheless, we can use the outlined procedure to derive automatically “best” estimated Q values from a large data set and then estimate the error of the obtained values from the standard deviation of the results. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. ( a ) 0.14 fast — - slow 0.12 ^ 0.08 0.06 0.04 0.02 0 10 20 30 40 50 60 70 80 90 100 Q (b) observed S up _ observed S . dow n calculated S J dow n fast 0.5 © > -0.5 0.3 0.4 0.5 0.6 0.7 0.8 0.9 slo w 0.5 & u © 0 ) -0.5. 0.3 0.4 0.5 0.6 0.7 0.8 0.9 time (sec) Figure 3.6. (a) Fitting errors between the observed wavefromss and the calculated wavefromss by equation (9) with different trail values of Q . The minimum error values indicated by arrows correspond to the estimated and , respectively, (b) comparison between between observed waveforms and calculated ones corresponding to the estimated Q values. 85 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 3.4.2 Estimated Q Values from a Set of 156 Recordings There is typically a large scatter in Q measurements made from individual recording. This raises doubts on the reliability of Q values estimated from few observations. In this section we employ 156 events that produce clear direct and reflected phases in our borehole data set for additional attenuation analysis. Since these events are located within the shear-wave window (Liu et al. 2004, and references therein), the waves generated by them should have nearly vertical ray paths when they approach the free surface. Therefore, the reflected waves propagate to the borehole receiver through essentially identical paths. This implies that the amplitude spectral ratios between the reflected and direct phases for these waveforms should follow the same relationship. The direct and reflected phases of these seismograms for the fast and slow components are windowed with a cosine taper and shown in Figure 3.7a. We then calculate the amplitude spectral ratios from these phases and give the stacked results in Figure 3.7b. We find that the amplitude spectral ratios for frequencies larger than 15 Hz display a large scatter. We estimate Q values by fitting the average amplitude spectral ratios versus frequency curves to equation (4) in the range 2 - 15 Hz and obtain Qf =62 + 5 and Qs = 45 ± 5 for the fast and slow shear wave components, respectively. We also estimate Q values with the waveform fitting method for each of these recordings. The distributions of the estimated Q values are shown in Figure 3.7. The average value of Qf is 62 with a 86 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 0.5 -0.5 0.4 0.5 0.6 0.7 0.8 0.9 0.5 -0.5 0.6 0.7 0.8 0.9 0.4 0.5 m m 0.4 0.5 0.6 0.7 0.8 0.9 0.4 0.5 0.6 0.7 0.8 0.9 (b) time (sec) time (sec) 1 0 ' fast Q = 62 10° 20 i 10 slowQ =45 o 10 10 15 frequency (Hz) Figure 3.7. (a) Direct and reflected windowed phases o f the fast and slow shear wave components for 156 recordingss. (b) Amplitude spectral ratios versus frequency. The heavy solid lines represent the average values and their linear fittings to equation (3.4) in the frequency range 2 - 15 Hz give the estimates o f Qf and Qs. 87 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 35 30 25 20 15 10 5 0 20 35 30 25 a 20 ° 15 10 5 30 40 50 60 Q f 70 80 90 100 0 20 30 40 90 100 Figure 3.8. Distributions of and calculated with the waveform fitting method for 156 recordings. standard deviation of 11 for the fast shear wave component and the average value of Qs is 48 with a standard deviation of 11. 3.5 Discussion 3.5.1 Attenuation in the Crust Several mechanisms have been identified to contribute to seismic attenuation and velocity dispersion (Winkler and Murphy III, 1995). It appears that in homogeneous rocks attenuation and dispersion are dominated by viscous fluid/solid interactions. 88 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. In heterogeneous rocks, scattering can lead to dispersion and energy diffusion. A frictional mechanism is only important at large strain amplitudes in the near field of seismic sources. The shapes of seismograms are strongly affected by seismic attenuation. This obscure source properties that are very important for earthquake physics studies. To characterize the source properties of earthquakes, it is important to separate the source from the path and site effects. Moreover, attenuation analysis also provides a tool to probe rock properties along the ray path. Downhole experiments provide the most reliable information on attenuation properties of the shallow crust. It is usually difficult to estimate near-surface attenuation from surface observations because of the free surface amplification and other complexities. To avoid the interference of near-surface amplification, it is desirable to use clear surface-reflected waves in downhole recordings (Hauksson, 1987). The results of this study based on such surface-reflected waves provide robust estimates of near surface attenuation of fast and slow shear waves over the frequency range 2 - 15 Hz for the study area. 3.5.2 Body Wave Dispersion Dispersion of body waves is a consequence of any causal theory of absorption. Aki and Richards (2002) show that the assumptions of constant Q and linearity of seismic waves lead without dispersion to non-causality. Since observed data indicate that those assumptions typically characterize the solid earth materials, a dispersion must exist to preserve causality of a propagating wave. From Figures 3.5b and 3.6b we see that the curves of amplitude spectral ratio versus frequency are well represented by 89 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. a linear line in the approximate frequency range 2-15 Hz. This indicates that the assumption of frequency-independent Q also characterizes well that frequency range in our data. Various theories of dispersion (Lomnitz,1957; Futterman 1962 ; Kolsky, 1957; Liu et al. 1976) have the logarithmic dispersion form A waveform distortion due to dispersion of local earthquake records can affect studies based on waveform shape, such as shear-wave splitting analysis. Since seismic waves attenuate greatly when propagating through the near-surface crust, they are expected also to be distorted by corresponding dispersion. Figure 3.4 shows that the shape of the surface-reflected phase differs significantly from that of the direct one. Similar results can been found in the borehole observations by Hauksson et al. (1987). To estimate the dispersion of shear waves in the top 200 m of the crust, we first calculate the phase difference ( AOr f o v w l ) between the direct and reflected waves for both the fast and slow components. In the calculation, the direct and reflected waves are aligned using cross-correlation. Considering the phase change related to wave propagation, we have the total phase change between the direct and reflected wave as c(a> 2) jQ y(o2) (3.10) down,up + 2 4 ( t down (3.11) From equation (3.8), we have 90 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. Thus, we can calculate phase velocity as 'down (3 .1 3 ) down,up Similar to equation (8), from equation (10), we can get c ( / ) = c ( / 0) + ^ l o g ( - f ) . kQ j o (3.14) The dispersion curves extracted from the stacked waveforms (Figure 3.4) using equation (3.13) are shown in Figure 3.9a. The measured phase velocities for the fast and slow components are represented by small circles and triangles, respectively. The solid and dashed lines in the figures give the theoretical dispersion curves based on equation (3.14) with Q = Qf and Q = Qs, respectively. The theoretical curves with Q = Qf fit observed ones well within the frequency rang 4-15 Hz for both the fast and slow components. Beyond 15 Hz, the noise-to-signal ratio is too high to extract stable phase information. The measured phase velocities drop rapidly in the frequency range less than 4 Hz and theoretical results can not fit this portion data well. Similar phenomena can be seen in (Wuenschel, 1965). Figure 9b shows the impact of the dispersion on the waveform shape. The calculated waveform without dispersion is significantly different from the observed one; however the calculated results with dispersion can improve fit. We note that although the fast and slow shear waves attenuate differently, they disperse almost identically and for both of them the theoretical curves with Q = Qf (solid lines) fit the measured data better than those 91 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. (a) 044 0.43 0.42 0.41 fast slow 0.4 ■ H . 0.39 0.38 O measured c(f) (fast) A measured c(f) (slow) — - calculated c(f) with Q=50 calculted c(f) with Q-65 0.37 0,36 frequency (Hz) (b) observed S , last, up _ . observed S , . tast, dow n calculated S _ , , with dispersion fast, dow n r calculated S „ , without dispersio n • — tast, d ow n r 0.6 0.4 0.2 o -0.4 - 0.6 0.4 0.45 0.5 0.55 0.7 0.8 0.85 0.6 0.65 0.75 time (sec) Figure 3.9. (a) Observed phase velocities versus frequency and calculated ones based on equation (14). (b) Results of waveform fitting with equation (8) compared with the results without dispersion. 92 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. with Q = Qs . It seems that the mechanisms that result in additional attenuation for slow shear wave has no contribution to body-wave dispersion. 3.5.3 Attenuation Anisotropy It has been observed from VSP data that transmitted amplitudes display a systematic variation with azimuth (Liu et al., 1993; Horne and MacBeth, 1997). The amplitude variation is commonly interpreted as attenuation that is related to the fractures. Anisotropic attenuation has been also observed in laboratory measurements on rock samples containing aligned crack (Thomsen, 1995). Attenuation anisotropy is one of the main seismic signatures of cracks that could be used for fracture detection. Aster and Shearer (1991) found evidence for preferential attenuation of the slow horizontal component relative to the fast horizontal component in borehole experiments near the San Jacinto Fault Zone, Southern California. They suggested the existence of Table 3.1. The estimated Q from two data sets using two methods. Data 7 multiplets 156 events Method I II I II Qf 68 61 62 62 Q. 52 43 45 48 Attenuation anisotropy 24% 30% 24% 23% 93 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. anisotropic shear-wave attenuation between 150 and 300 m, which is below the top weathered layer. They also argued that such phenomena may be partially responsible for the clear fast shear waves and the general lack of distinct slow shear waves. Our high quality data set provides excellent conditions to study systematically the attenuation anisotropy in the near-surface crust. Based on our pervious SWS analysis (Liu et al. 2004), the horizontal shear wave can been separated clearly into fast and slow shear wave components. We can therefore estimate separate Q values for those two components of anistropic shear waves. The estimated results from the two employed data subsets using the two methods of section 3 are listed in Table 3.1. The value of Qf for the fast shear wave ranges from 61 to 68, while the value of Qs for the slow shear component ranges from 43 to 52. The attenuation anisotropy ranges from 23% to 30%. In the above analysis, we assumed that the polarization directions are the same for the sections above and below the borehole station. As discussed in Liu et al. (2004), there is probably a 6 degree difference between the polarization directions for these two sections. However, we found that such slight vertical variation of polarization direction could changes the estimated Q value by 1- - 3, which is much smaller than the uncertainties of our estimated results. Because of attenuation anisotropy, the amplitude decays differently for different polarization directions. We calculate the amplitude ratios, defined as the ratio of peak to trough amplitude, between the direct and reflected waves in the time domain. The results for both the fast and slow shear waves 94 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. (a) 40 30 I 20 10 0 0.2 fast I l l iiiiiiiii i l l ' -v* > ^ % ■ n 0.3 0.4 0.5 0.7 0.9 0.2 0.3 30 slow 25 20 15 10 5 0 0.4 0.5 0.6 0.7 am plitude ratio 0.8 0.9 Figure 3.10. Distributions of the amplitude ratios between the reflected and direct waves in time domain for the fast and slow components. The mean and standard deviation values of the amplitude ratios are 0.65and 0.10 for the fast component, while the corresponding values for the slow components are 0.55 and 0.08. obtained from 156 events are shown in Figure 3.10. The distribution of the amplitude ratios for the fast shear wave components peaks at around 0.65, while the distribution of the amplitude ratios for the slow components peaks at around 0.55. 3.6 Conclusions Estimating Q values from the direct and reflected seismic phases in downhole recordings has the advantage of using signals that have the same instrument and site responses. In addition, the analysis avoids near-surface amplification, which can 95 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. complicate measurements of attenuation due to the near-surface structure from surface seismic data. We estimate Q values from a set of earthquake multiplets and 156 events with high- quality recordings using both the amplitude spectral ratio and a waveform fitting method. The estimated value of Qf for the fast shear wave component is 61 - 68, and the estimated value of Qs for the slow shear wave component is 43 - 52 . The results reveal a substantial difference of attenuation between the fast and slow shear wave components and show a clear evidence of attenuation anisotropy in the near surface structure. The observed attenuation anisotropy may be a manifestation, of microcracks alignment and their response to in-situ stress, as is commonly assumed for the velocity anisotropy. An attenuation-related dispersion is clearly observed and it has a significant effect on the shapes of waveforms. The observed dispersion curves fit the theoretical logarithm dispersion equation well in the frequency range 4-15 Hz. The mechanisms that result in additional attenuation for slow shear do not likely contribute to body-wave dispersion. The observed strong attenuation anisotropy (23% - 30% in this study) along with attenuation-related dispersion is likely to characterize the near-surface structure in other locations. These effects can modify significantly the properties of observed 96 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. shear-wave seismograms and should be taken into account in studies employing such data. Acknowledgements The research was supported by the National Science Foundation (grant EAR- 0124926) and Taiwan Central Weather Bureau (MTOC-CWB-93-E-06) 97 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. RECAPITULATION In the thesis, we systematically analyze shear-wave splitting in the surrounding area of the 1999, Chi-Chi, Taiwan earthquake. The purpose of the study is to investigate the spatial and temporal distribution of stress-induced crustal anisotropy in relation to tectonic activity of the crust in relation to the occurrence of a big earthquake. The high- quality data set recorded by a downhole short period station and numerous surface strong motion stations provide reliable measurements of SWS parameters. The results show that the polarization directions of the fast shear waves match well the stress field, indicated by GPS velocity fields, in the study region. Some measurements from stations nearby active faults display a polarization direction parallel or sub-parallel to the fault strike. Measured fast polarization directions and time delays of SWS vary significantly with location. No dependence of time delay on depth has been found in various areas over depth range 5-18 km. Analysis results based on recordings from a borehole station further demonstrate that the crustal anisotropy in the region is dominated by the top 2 -3 km while the top 200 highly damaged crust contributes about 20% of the total shear-wave splitting time delay in the upper crust. The observed SWS parameters in the near-surface and deeper crust, and the great similarity of waveforms generated by multiplets, show no appreciable systematic temporal changes over the 2.7 year period before the 1999 Mw 7.6 Chi-Chi 98 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. mainshock and the 2.3 year period after. The lack of such temporal changes and the dominant influence of the shallow structure on the data suggest that the seismic anisotropy in our study region may be associated with a preferred closure of randomly-distributed cracks due to the existing anisotropic stress field. We also systematically measure the near-surface attenuation based on the surface - reflected waves in borehole recordings. The results reveal a substantial difference of attenuation between the fast and slow shear wave components and show clear evidence of attenuation anisotropy in the near-surface structure. The observed attenuation anisotropy may be a manifestation of microcracks alignment and their response to in-situ stress, as is commonly assumed for the velocity anisotropy. The observed strong velocity and attenuation anisotropy along with attenuation-related dispersion are likely to characterize the near-surface structure in other locations. 99 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. REFERENCES Aki, K. and Richards, P.G., (2002). Quantitative seismology, University Sciences Books. Ando, M. Y. Ishikawa, and F. Yamazaki, (1983). Shear wave polarization anisotropy in upper mantale beneath Flonsu, Japan, J. Geophys. Res. 88, 5850-5864. Aster, R.C. and P.M. Shearer (1991). High frequency borehole seismograms recorded at the San Jacinto fault zone, southern California, part 1. Polarizations, Bull. Seism. Soc. Am. 81, 1057-1080. Aster, R.C. and Shearer, P.M., (1991). High-frequency borehole seismograms recorded in the San Jacinto fault zone, Southern California, Part 2. Attenuation and site effects, Bull. Seism. Soc. Am., 88, 1081-1100. Aster, R.C. and P.M. Shearer (1992). Initial shear wave particle motions and stress constraints at Anza seismic network, Geophys. J. Int., 108, 1-9. Aster, R., P. Shearer, and J. Berger (1990). Quantitative measurements of shear wave polarizations at Anza Seismic network, southern California: Implications for shear splitting and earthquake predication, J. Geophys. Res., 95, 12,449-12,474. Aster, R., P. Shearer, and J. Berger (1991). Reply, J. Geophys. Res., 96, 6415-6419. Bos, A. G., W. Spakman, and M.C.J. Nyst, Surfac deformation and tectonic setting of Taiwan inferred from a GPS velocity field, J. Geophys. Res. 108(B10), 2456, doi: 10.1029/2002JB002336, 2003. 100 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. Bath M., (1974). Spectral analysis in geophysics, Elsevier scientific publish company, Amsterdam, The Netherlands, 563 pp. Blenkinsop, T.G. (1990). Correlation of Paleotectonic Fracture and Microfracture Orientations in Cores With Seismic Anisotropy at Cajon Pass Drill Hole, Southern California. J. Geophys. Res, 95, 11,143-11.150. Boness, N. and M.D. Zoback, (2004). Stress-induced seismic velocity anisotropy and physical properties in the SAFOD pilot hole in Parkfield, CA., Geophys. Res. Lett., in press. Booth, D. C., S. Crampin, R. Evans, and G. Roberts, (1985) Shear-wave polarization near the north Anatolian Fault, I, Evidence for anisotropy-induced shear-wave splitting. Geophys. J., R. Astron. Soc., 83, 61-73. Booth, D. C., Crampin, S., Lovell, J.H. and Chiu, J.-M. (1990). Temporal change in shear-wave splitting during an earthquake swan in Arkansas, J. Geophys. Res., 95, 11,151-11,164. Bowman, J. R. and M. Ando (1987). Shear-wave splitting in the upper-mantle wedge above the Tonga subduction zone, Geophys. J. R. Astron. Soc., 88, 25- 41. Chung, J.-K. and Yeh, Y.-T. (1997). Shallow crustal structure from short-period Rayleigh-wave dispersion data in southwestern Taiwan, Bull. Seism. Soc. Am., 87, 370-383. 101 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. Coutant, O. (1996). Observation of shallow anisotropy on local earthquake records at the Garner Valley, southern California, downhole array, Bull. Seism. Soc. Am. 86, 477-488. Crampin, S. (1978). Seismic wave propagation through a cracked solid: polarization as possible dilatancy diagnostic, Geophys. J. R. Astron. Soc., 53, 467-496. Crampin, S. (1981). The geological and industrial implication of extensive-dilatancy anisotropy , Nature,3,343-391. Crampin, S., Booth, D.C., Evens, R., Peacock, S. and Fletcher, J.B. (1990). Change in shear wave splitting at Anza near the time of the North Palm Springs earthquake, / Geophys. Res., 95, 11,197-11,212. Crampin, S., Booth, D.C., Evens, R., Peacock, S. and Fletcher, J.B. (1991). Comment on ‘Quantitative measurements of shear wave polarizations at the Anza seismic network, Southern California: Implications for shear wave splitting and earthquake predication’ by R.C. Aster et al, J. Geophys. Res., 96, 6403-6414. Crampin, S. R. Evans, and B.K, Atkinson (1984).Earthquake predication: A new physical basis, Geophys. J. R. Astron. Soc., 76,147-156. Crampin, S., Volti, T. and Stefansson, R. (1999). A successfully stress-forecast earthquake, Geophys. J. Int., 138, F1-F5. Crampin, S. and Zatsepin, S.V. (1997). Modeling the compliance of crustal rock: II response to temporal changes before earthquake, Geophys. J. Int., 129, 495- 506. 102 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. Delay, T.M., T.V. McEvilly, and E.L. Majer (1988) Multiple Polarized shear VSPs from Cajon Pass drill hole, Geophys. Res. Lett. 15,1001-1004. Fletcher, J.B., Fumal, T., Liu, H.-P., and Carroll, L.C. (1990). Near-surface velocities and attenuation at two boreholes near Anza, California, from logging data, Bull. Seism. Soc. Am., 80 , 807-731. Fukao, Y. (1984). Evidence from core-reflected shear waves anisotropy in the earth mantle, Nature, 309, 695-698. Futterman, W.I., (1962). Dispersive body waves, J. Geophys. Res. ,67, 5279-5291 Gledhill, K.R. (1991). Evidence for shallow and pervasive seismic anisotropy in the Wellington region, New Zealand, J. Geophys. Res. 96, 21,503-21,516. Gao, Y., Wang, P., Zheng, S., Wang, M. & Chen, Y.-T. (1998). Temporal changes in shear-wave splitting at an isolated swarm of small earthquakes in 1992 near Dongfang, Hainan Island, Southern China, Geophys. J. Int., 135, 102-112. Hauksson, E., Teng, T.-L., Henyey, T.L. (1987). Results from a 1500 M deep, three- level downhole seismometer array: site response, low Q values, and fm a x , Bull. Seism. Soc. Am., 77, 1883-1904. Ho, C.S., Tectonic evolution of Taiwan, Explanatory text of the tectonic map of Taiwan, Published by The ministry o f economic affair, Taipei, 1982. Horne, S. and MacBeth, C. (1997). AVA observations in walkaround VSPs, 67th Ann. Internat. Mtg: Soc. OfExpl. Geophys. , 290-293. Kaneshima, S. (1990). Origin of crustal anisotropy: Shear wave splitting studies in Japan, J. Geophys. Res., 95, 11,121-11,133. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. Kanasewich, E.R. (1981). Time sequence analysis, The University of Alberta Press, Third Edition. Hamiel, Y., Liu, Y., Lyakhovsky , V., Ben-Zion, Y. and Lockner, D. (2004). A visco-elastic damage model with applications to stable and unstable fracturing. Geophys. J. Int Kern, El. and Wenk, EL-R. (1990). Fabric-related velocity anisotropy and shear wave splitting in rocks from the Santa Rosa Mylonite Zone, California, J. Geophys. Res., 95, 11,213-11,224. Kolsky, H. (1956). The propagation of the stress pulses in viscoelastic solids, Phil. Mag., 8, 1,673-710. Krajcinovic, D. (1996). Damage Mechanics, Elsevier, Amsterdam. Leary, P.C. Crampin, S. and McEvilly, T. (1990). Seismic Fracture Anisotropy in the Earth’s Crust: An Overview. J. Geophys. Res, 95, 11,105-11,114. Leary, P.C., Li, Y.G. and Aki, K. (1987). Observations and modeling of fault-zone fracture seismic anisotropy, I. P, S V and SH travel times, Geophys. J. R. Astron. Soc., 91, 461-484. Li, Y.-G., Leary,P.C., and T.L. Henyey (1988). Stress orientation inferred from shear waves splitting in basement rock at Cajon Pass, Geophys. Res. Lett. 15,997- 1000. Li, Y.-G. (1996). Shear wave splitting observations and implications on stress regims in the Los Angeles Basin, California, J. Geophys. Res., 101, 13,947-13,961. Li, Y.G., Teng, T-L, Henyey, T.L. (1994). Shear- wave observations in the Northern 104 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. Los Angeles Basin, Southern California, Bull. Seism. Soc. Am., 84, 307-323 Liu, Y., Crampin, S. and Main, I. (1997). Shear-wave anisotropy: spatial and temporal variations in time delays at Parkfield, central California, Geophys. J. Int., 130, 771-785. Liu, E. Crampin, S. Queen, J.,H., and Rizer, W.D. (1993). Velocity and attenuation anisotropy caused by microcracks and macrofractures in a multiazimuth reverse VSP, Can. J. Expl. Geophys. 19, 162-176. Liu, H.P., Anderson, D.L. and Kanamori, H. (1967). Velocity dispersion due to anelasticity; implications for seismology and mantle composition, Geophys. J.R. Astron. Soc., 47, 41-56. Liu, Y., Teng, T.-L., Ben-Zion, Y. (2004a). Systematic analysis of shear-wave splitting in the aftershock zone of the 1999 Chi-Chi, Taiwan, Earthquake: Shallow crustal anisotropy and lack of precursory variations, Bull. Seism. Soc. Am. Liu, Y., Teng, T.-L., Ben-Zion, Y. (2004b). Near-surface seismic anisotropy, attenuation and dispersion in the aftershock region of the 1999 Chi-Chi, earthquake, Geophys. J. Int. Liu, Y., Teng, T.-L., Ben-Zion, Y. (2004c). Shear-Wave Splitting and Spatial Distribution of Crustal Seismic Anisotropy in Taiwan , in preparing. Lomnitz, C. (1957). Linear Dissipation in solids, J. Appl. Phys., 28, 201-205. Nutti, O. (1961). The effect of earth’s surface on the S-wave particle motion, Bull. Seism. Soc. Am. 51, 237-246. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. Peacock, S., S. Crampin, D.C. Booth, and J.B. Fletcher (1988). Shear-wave splitting in the Anza seismic gap southern California: temporal variations as possible precursors, J. Geophys. Res., 93, 3339-3356. Rau, R.-J., Liang, W.-Z., Kao, H. and Huang, B.-S., (2000). Shear wave anisotropy beneath the Taiwan orogen, Earth Planet. Set. Le., 177, 177-192. Rau, R.-J. (2002). Seismicity and fault structures in Taiwan, 2002 APEC Workshop on Active Fault Study, Central Geological Survey, Taiwan. Savage, M.K., W.A. Peppin, and U.R. Vetter (1990). Shear wave anisotropy and stress direction in and near Long Valley caldera, California, 1979-1988, J. Geophys. Res. 95, 11,1165-11,117. Savage, M.K., X.R. Shih, R.P. Mayer, and R.C. Aster (1989). Shear-wave anisotropy of active tectonic region via automated S-wave polarization analysis, Tectonophysics, 165, 279-292. Silver, P.G. and Chan W.W. (1991). Shear wave splitting and subcontinental mantle deformation, / Geophys. Res., 96, 16429-16454. Silver, P.G. and Savage, M.K., (1994). The interpretation of shear-wave splitting parameters in the presence of two anisotropic layers, Geophys. J. Int., 119, 949-963. Shih, X.R., R.P. Meyer, and J.F. Schneider (1989). An automated and analytical method to determine shear wave splitting, Tectonophysics, 165, 271-278. Shih,X.R., R.P. Meyer (1990). Observation of shear wave splitting from nature events: South Moat of long valley Caldera, California, June29 to August 12, 106 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 1982, J. Geophys. Res., 95, 11,179-196. Shin, T.-C. and Teng, T.-L. (2001). An overview of the 1999 Chi-Chi, Taiwan, earthquake, Bull. Seism. Soc. Am. 91, 5, 895-913. Tadokoro, K., Ando, M., and Umeda, Y. (1999). S wave splitting in the aftershock region of the 1995 Hyogo-ken Nanbu earthquake, J. Geophys. Res. 104, 981 - 991. Thomsen, L. (1995). Elastic anisotropy due to aligned cracks in porous rock, Geophys. Prosp., 43, 805-829 Wang, C-Y. and Sun, Y. (1990). Oriented Microfractures in Cajon pass Drill Core: Stree Field Near the San Andreas Fault, J. Geophys. Res, 95, 11,135-11.142. Winkler K.W. and Murphy III W.F. (1995). Acoustic velocity and attenuation in porous rock, Rock physics and phase relation, A handbook of physical constants, AGU Reference shelf 3, 20-34. Winterstein, D.F. and M.A. Meadows (1991a). Shear-wave polarizations and surface stress directions at Lost Hills field, Geophysics, 56, 1331-1348. Winterstein, D.F. and M.A. Meadows (1991b). Changes in shear-wave polarization azimuth with depth in Cymric and Railroad Gap oil field, Geophysics, 56, 1349-1364. Wolfe, C.J. and Silver, P.G. (1998). Seismic anisotropy of oceanic upper mantle: Shear wave splitting methodologies and observations, J. Geophys. Res. 103, 949-972 Wu, F. T. and R-J Rau, (1998). Seismotectonics and Identification of Potential 107 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. Seismic Source Zones in Taiwan, TAO, 9, 4, 739-754 Wuenschel, C. P., (1965). Dispersive body waves - an experimental study, Geophys., Vol. 15, 539-551. Yu. S.-B., Chen, H.-Y., and Kuo, L.-C. (1997). Velocity field of GPS stations in the Taiwan area, Tectonophys. 274. 141-59. Yu, S.-B., Kao, L.-C., Hsu, Y.-J., Su, H.-H., Liu, C.-C., Hou, C.-S., Lee, J.-F., Lai, T.-C., Liu, C.-C., Liu, C.-L., Teng, T.-F., Tsai, C.-S., and Shin, T.-C. (2001). Preseismic deformation and coseismic displacements associated with the 1999 Chi-Chi, Taiwan, earthquake, Bull. Seism. Soc. Am. 91, 995-1012. Zatsepin, S.V. and Crampin, S. (1997). Modelling the compiance of crustal rock: I. Response of shear-wave splitting to differential stress, Geophys. J. Int. 129, 477-494. Zinke, Jens C. and Zoback, Mark D. (2000). Structure-related and stress-induced shear-wave velocity anisotropy: Observations from microearthquakes near the Calaveras fault in central California, Bull. Seism. Soc. Am. 90, 1305-1312. Zhang, Z. and Schwartz, S.Y. (1994). Seismic anisotropy in the shallow crust of the Loma Prieta segment of the San Andreas fault system, J. Geophys. Res., 99, 9651-9661. 108 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. APPENDIX: SEISMIC DATA AND ANALYSIS CODES strong motion waveform data from 1993 to 2000 are sequentially in directories /home/terra-01/ yunfengl/wrk2 0/ SWS/ SMDATA/19 93/SELECTED/ /home/terra-01/yunfengl/wrk2 0/SWS/SMDATA/1994/SELECTED/ /home/terra-01/yunfengl/wrk20/SWS/SMDATA/l995/SELECTED/ /home/terra-01/yunfengl/wrk20/SWS/SMDATA/1996/SELECTED/ /home/terra-01/yunfengl/wrk20/SWS/SMDATA/1997/SELECTED/ /home/terra-01/yunfengl/wrk20/SWS/SMDATA/1998/SELECTED/ /home/terra-01/yunfengl/wrk20/SWS/SMDATA/l999/SELECTED/ /home/terra-01/yunfengl/wrk20/SWS/SMDATA/2000/SELECTED/ All data are in sac form. The data under preprocessing are in the corresponding directories /home/terra-01/yunfengl/wrk20/SWS/SMDATA/1993/DATA /home/terra-01/yunfengl/wrk20/SWS/SMDATA/1994/DATA /home/terra-0l/yunfengl/wrk20/SWS/SMDATA/1995/DATA /home/terra-01/yunfengl/wrk20/SWS/SMDATA/1996/DATA /home/terra-01/yunfengl/wrk20/SWS/SMDATA/1997/DATA /home/terra-01/yunfengl/wrk20/SWS/SMDATA/1998/DATA /home/terra-01/yunfengl/wrk20/SWS/SMDATA/1999/DATA /home/terra-01/yunfengl/wrk20/SWS/SMDATA/1999/DATA Shear-wave splitting analysis matlab files for 1993 are in directory. /home/terra-01/yunfengl/wrk20/SWS/SMDATA/1993/MFILE For aspect ratio method, codes include following files analar.m arcdp.m r sac . m sac .m xcorrO .m For cross-correlation method, codes include following files 109 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. analcdp.m cdp .m rsac.m sac .m xcorrO.m both methods use inport file recinf.dat and waveform data in the directory /home/terra-01/yunfengl/wrk2 0/SWS/SMDATA/1993/DATA Analysis results (figures) are put in directory home/terra-01/yunfengl/wrk2 0/SWS/SMDATA/l998/OPFIG eps file /home/terra- 01 /yunfengl/wrk2 0/SWS/SMDATA/1999/OPFIG/fig_ar/CHY073_12 1602 35test.e ps is sample of analysis results using aspect ratio method. Similarly for other years. Short period waveform data of CHY station from 1997 to 2000 are in directory /home/terra-01/yunfengl/wrk20/SWS/SPCHY/DATA Shear-wave splitting analysis matlab files for 1993 are in directory, /home/terra-01/yunfengl/wrk2 0/SWS/SPCHY/MFILE> and Analysis results (figures) are in directory /home/terra-01/yunfengl/wrk2 0/SWS/SPCHY/OPFIG 110 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Investigation of arc processes: Relationships among deformation, magmatism, mountain building, and the role of crustal anisotropy in the evolution of the Peninsular Ranges Batholith, Baja California

PDF

Management of large earthquake data using the Antelope Relational Database and seismicity analysis of the 1999 Turkey earthquake sequences

PDF

Paleoseismology of the North Anatolian fault, Turkey: Towards a long-term record of seismic moment release in time and space

PDF

Using small events to measure the evolution of the strain and stress fields before the 1992 Landers, California earthquake

PDF

A unified methodology for seismic waveform analysis and inversion

PDF

Response of deep-ocean ostracodes to climate extrema of the Paleogene: Ecological, morphological, and geochemical data from the Eocene -Oligocene transition and late Paleocene thermal maximum

PDF

Changing characteristics of deformation, sedimentation, and magmatism as a result of island arc -continent collision

PDF

Systematic high -resolution imaging of fault zone structures

PDF

Oxygen isotope constraints on the petrogenesis of the Sybille intrusion of the Proterozoic Laramie Anorthosite Complex

PDF

Contact aureole rheology: New constraints from fieldwork in selected Cordilleran aureoles and from numerical modeling

PDF

Early Jurassic reef eclipse: Paleoecology and sclerochronology of the "Lithiotis" facies bivalves

PDF

The clinical significance, detection and biological characteristics of occult metastasis in breast cancer

PDF

Modeling and imaging asperities on a fault plane and characterizing spatial and temporal patterns of precursory seismicity

PDF

Numerical modeling of long period events at the Piton de la Fournaise by the use of object-oriented scientific computing

PDF

Expression and regulation of B7 on antigen presenting cells

PDF

Investigating the influence of atmospheric changes on the variability of the North Pacific using paleoproxy data and a fully coupled GCM

PDF

Part i: 3d simulations of short-period surface-wave propagation in southern California; and part ii: recursive stochastic deconvolution in the estimation of earthquake source parameters

PDF

The origin of quartz grain-shape facies and their variations as a function of grain size and transport distance, San Diego County, California

PDF

Synthesis, devices and applications of carbon nanotubes

PDF

Role of reactive oxygen species in iron-induced and age-related neuronal damage

Asset Metadata

Creator Liu, Yunfeng (author)

Core Title Systematic analysis of crustal anisotropy and attenuation using seismic data associated with the 1999 Chi-Chi, Taiwan, earthquake

Degree Doctor of Philosophy

Degree Program Earth Sciences

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag geophysics,OAI-PMH Harvest

Language English

Contributor Digitized by ProQuest (provenance)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c16-406902

Unique identifier UC11340261

Identifier 3145234.pdf (filename),usctheses-c16-406902 (legacy record id)

Legacy Identifier 3145234.pdf

Dmrecord 406902

Document Type Dissertation

Rights Liu, Yunfeng

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA