UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Teleseismic receiver function analysis of the crust and upper mantle of southwestern British Columbia Cassidy, John Francis 1991

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-UBC_1991_A1 C37.pdf [ 10.18MB ]
Metadata
JSON: 831-1.0052621.json
JSON-LD: 831-1.0052621-ld.json
RDF/XML (Pretty): 831-1.0052621-rdf.xml
RDF/JSON: 831-1.0052621-rdf.json
Turtle: 831-1.0052621-turtle.txt
N-Triples: 831-1.0052621-rdf-ntriples.txt
Original Record: 831-1.0052621-source.json
Full Text
831-1.0052621-fulltext.txt
Citation
831-1.0052621.ris

Full Text

TELESEISMIC RECEIVER FUNCTION ANALYSIS OF T H E CRUST AND UPPER M A N T L E OF SOUTHWESTERN BRITISH COLUMBIA By John Francis Cassidy B.Sc. Physics (Honours), University of Victoria M.Sc. Geophysics, University of British Columbia  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY  in THE FACULTY OF GRADUATE STUDIES DEPARTMENT OF GEOPHYSICS AND ASTRONOMY  We accept this thesis as conforming to the required standard  THE UNIVERSITY OF BRITISH COLUMBIA  April 1991 © John Francis Cassidy, 1991  In presenting this thesis in partial fulfilment  of the requirements for an advanced  degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department  or  by  his or  her  representatives.  It  is understood that  copying or  publication of this thesis for financial gain shall not be allowed without my written permission.  Department The University of British Columbia Vancouver, Canada  Date  DE-6 (2/88)  2-i isq I  Abstract  The northern Cascadia subduction zone has been the site of numerous geophysical studies during the past two decades. However, little is known of the deep structure (> 40 km) or S-velocities throughout this region. In this study, locally generated P-to-S conversions (Ps) contained in -100 teleseismic P-wave coda have been analysed to determine the S-velocity structure to upper mantle depths. Prior to the analysis, the applications and limitations of this technique as applied to a dipping layer environment have been examined. It is concluded that strict stacking bounds (< 10° in A and BAZ) should be applied. It is demonstrated that dipping boundaries which could not be detected using this technique (e.g. AV = 0.08 km/s), may significantly alter the S  amplitude and arrival time of reverberations from deeper interfaces. Therefore, such phases should not be quantitatively modelled. As reverberations are an important constituent of receiver functions, formal inversion of these waveforms is not justified in this environment. Only arrivals which exhibit the amplitude and arrival time characteristics of primary P-to-S conversions are considered in this study. Finally, most studies have normalised receiver functions to unit amplitude prior to modelling. However, synthetic data demonstrate that undetected dipping boundaries may alter Ps/P ratios and lead to inaccurate earth models. A recent modification to this technique (Ammon, 1991) which provides 'absolute' amplitudes is examined. In addition to providing information on the near-surface velocity structure and on dipping layers, this modification provides for a more accurate image of the earth structure. Three 3-component broadband event triggered seismic stations were deployed in a 90 km long linear array oriented perpendicular to the continental margin of southwestern British  ii  Columbia. Between December 1987 and October 1989 approximately 100 teleseisms covering a wide azimuthal and distance range were recorded and analysed. The two largest phases observed in data from the westernmost station ALB-B reveal a prominent low-velocity zone extending from 37-41 km depth beneath central Vancouver Island. This feature correlates well with the reflective ' E ' zone, a region which also exhibits high electrical conductivity. Combining the S-velocity estimates with refraction P-velocities yields a high Poisson's ratio for this layer. The low P- and S-velocities and high Poisson's ratio and electrical conductivity are supportive of the recent interpretation of this feature as a fluid-saturated shear zone above the subducting Juan de Fuca (JdF) plate. Analysis of data at the mid-array and easternmost sites, LAS and EGM respectively, permits this zone to be mapped northeastward to a depth of 54 km beneath the British Columbia mainland, approximately 250 km from the locus of subduction. The subducting oceanic crust is imaged at 47-53 km depth dipping 15°±5° in the direction N30°E±20° beneath central Vancouver Island. The dip angle increases to 22° ±5° at a depth of 60-65 km beneath the Strait of Georgia. The results of this analysis provide the first definitive evidence for the location of the subducting plate in this region and indicates that the seismicity at depth occurs within the oceanic crust. Further, the dip direction of N30°E supports the theory (Rogers, 1983) that the JdF plate is arched upwards as it subducts in this region. Finally, the continental Moho is imaged at 36 km depth beneath LAS, and there is evidence at both this site and EGM for a low-velocity zone in the lower crust. A similar feature is imaged beneath Vancouver Island and coincides with the reflective ' C zone. The depth estimated to the top of this layer denotes the lower limit of shallow seismicity suggesting a significant structural or compositional change at a depth of 20-26 km.  iii  Table of Contents  Abstract  ii  List of Figures  viii  List of Tables  xi xii  Acknowledgements 1 INTRODUCTION  1  1.1 Thesis Outline  4  1.2 Tectonic Setting  4  1.3 Previous Geophysical Studies  6  1.3.1  Seismic Reflection Data  .  6  1.3.2  Seismic Refraction Data  8  1.3.3  Seismicity Data  11  1.3.4  Geothermal, Magneto telluric and Potential Field Data . . . . . . . .  11  1.3.5  Shear-wave Data  12  1.3.6  Summary and Problems to be Addressed  14  1.4 Receiver Function Analysis  15  1.5  19  Synthetic Receiver Functions  2 DATA ACQUISITION AND PROCESSING  21  2.1 Instrumentation  21  2.2 Station Description and Operation  24 iv  2.3  Preliminary Data Processing  28  2.4  The Data  28  2.5  Receiver Function Estimation  32  3 RECEIVER FUNCTION ANALYSIS OF DIPPING STRUCTURE  38  3.1  Introduction  38  3.2  Receiver Function Analysis and Dipping Layers  39  3.2.1  Modelling Absolute Amplitudes of Receiver Functions  42  3.2.2  Stacking Procedure  46  3.2.3  Lateral Sampling Range of P and Reverberations  52  3.2.4  Stability of Reverberations  54  3.3  3.4  s  Resolution Capability of Receiver Functions  58  3.3.1  Minimum Detectable AVs  58  3.3.2  Transition Zones vs I ' Order Velocity Discontinuities  60  3.3.3  Resolution of Thin Layers  60  3.3.4  Very Shallow Structure  63  3.3.5  Resolution and Modelling  67  3.3.6  Ps Sensitivity to AVs, AVp and p . . .  68  3.3.7  Broadband vs Short-period Receiver Functions  69  s  Summary.  71  4 MODELLING RECEIVER FUNCTIONS  73  4.1  Introduction  73  4.2  Forward Modelling  73  4.3  Comparison of Reflection Data and Receiver Functions  75  4.4  Outline  76  4.5  ALB-B Interpretation  77 v  4.6  4.7  5  4.5.1  The Data  77  4.5.2  Comparison With LITHOPROBE Reflection Data  84  4.5.3  Initial Considerations  87  4.5.4  Crustal S-Velocity Structure  89  4.5.5  Upper Mantle S-Velocity Structure  96  4.5.6  A L B - B Summary  97  L A S Interpretation .  104  4.6.1  The Data  104  4.6.2  Initial Considerations  109  4.6.3  Crustal S-Velocity Structure  Ill  4.6.4  Upper Mantle S-Velocity Structure  4.6.5  L A S Summary  :  . .  112 114  E G M Interpretation  122  4.7.1  The Data  122  4.7.2  Comparison With LITHOPROBE Reflection Data . . . . . . . . . .  127  4.7.3  Initial Considerations  129  4.7.4  Crustal S-Velocity Structure  131  4.7.5  Upper Mantle S-Velocity Structure  133  4.7.6  E G M Summary  134  I N T E R P R E T A T I O N A N D DISCUSSION  141  5.1  Regional S-Velocity Structure  141  5.1.1  144  5.2  Comparison With Previous Results  Results and Implications  147  5.2.1  Juan de Fuca Plate Subduction Geometry  147  5.2.2  The ' C and ' E ' Reflective Zones  151  vi  5.3  Summary and Conclusions  156  5.4  Future Studies  158  References  161  Appendix A  168  Appendix B  172  vii  List of Figures  1.1  Geometry of the Cascadia subduction zone  2  1.2  Station location and tectonic setting of the study area  3  1.3  Summary of previous studies  7  1.4  Seismic refraction models  10  1.5  Results of previous S-wave studies  13  2.1  Broadband velocity response of the Guralp seismometer systems  22  2.2  Geology of the study area.  2.3  A typical station set-up  2.4  Examples of high and low quality seismograms and receiver functions  2.5  Data distribution at A L B - B  31  2.6  Examples of teleseismic P-wave power spectra  33  2.7  A n example of the effectiveness of source equalisation  34  2.8  Examples of various quality receiver functions  2.9  Example of a receiver function stack  37  3.1  Azimuthal variation in receiver functions caused by a dipping interface . . .  41  3.2  Absolute versus relative receiver function amplitudes - shallow structure effects 43  3.3  Absolute versus relative receiver function amplitudes - dipping interface effects 45  3.4  Ps amplitude and arrival time variation as a function of B A Z and A . . . . .  48  3.5  Potential stacking errors  50  3.6  Examples of reverberations or scattered energy  51  . . .  25 26  viii  . . .  29  .  35  3.7  The lateral sampling range of Ps and reverberations at horizontal and dipping interfaces  53  3.8  Simplified ray diagram and synthetic receiver functions  55  3.9  Resolution of transition zones . . .  61  3.10 Resolution of thin layers  62  3.11 Effects of horizontal shallow structure  64  3.12 Effects of shallow, dipping structure  66  3.13 A comparison of short-period and broadband receiver functions  70  4.1  A L B - B receiver functions: B A Z = 130°  79  4.2  A L B - B receiver functions: B A Z = 300°  80  4.3  A L B - B receiver functions: A = 90°  81  4.4  A L B - B receiver functions: A < 65° . . . . .  82  4.5  A L B - B comparison of receiver functions and reflection data  85  4.6  A L B - B : A comparison of observed and synthetic data generated using previous models  88  4.7  A L B - B shallow structure effects  90  4.8  Comparison of observed and synthetic data for select receiver functions . . .  93  4.9  A L B - B final velocity model  98  4.10 A L B - B synthetic and observed data: B A Z = 130°  99  4.11 A L B - B synthetic and observed data: B A Z = 300°  100  4.12 A L B - B synthetic and observed data: A = 90° .  101  4.13 A L B - B synthetic and observed data: A < 65° . . .  102  4.14 L A S receiver functions: B A Z = 300° and 130° . . .  106  4.15 L A S receiver functions: A = 9 0 ° . . . . .  107  4.16 L A S receiver functions: A < 65°  108  ix  4.17 L A S data and synthetics generated using the starting model  110  4.18 A comparison of select L A S data and synthetics generated using the final model 113 4.19 L A S final velocity model  115  4.20 L A S synthetic and observed data: B A Z = 130°  '. . 117  4.21 L A S synthetic and observed data: B A Z = 300°  118  4.22 L A S synthetic and observed data: A = 90°  119  4.23 L A S synthetic and observed data: A < 65°  120  4.24 E G M receiver functions: B A Z = 1 3 0 °  123  4.25 E G M receiver functions: B A Z = 300°  124  4.26 E G M receiver functions: A = 90°  125  4.27 E G M comparison of receiver functions and reflection data  128  4.28 A comparison of E G M data and synthetics generated using the starting model 130 4.29 Fit provided to select data by the final model  132  4.30 E G M final velocity model  135  4.31 E G M synthetic and observed data: B A Z = 130°  137  4.32 E G M synthetic and observed data: B A Z = 300°  138  4.33 E G M synthetic and observed data: A = 90°  139  5.1  Comparison of S-velocity models  142  5.2  Tectonic summary and low-velocity zones  146  5.3  Comparison of JdF plate depth and hypocentres  148  5.4  Subduction geometry of the Juan de Fuca plate  150  5.5  Poisson's ratio and AV versus porosity  154  p  x  List of Tables  2.1  Station locations and recording periods  27  2.2  Summary of events recorded  30  3.1  Reference model used for Ps stacking tests  47  3.2  Arrival times and amplitudes of Ps and reverberations  57  3.3  Ps amplitude as a function of S-velocity contrast and A  59  3.4  Reference model used for Ps amplitude tests  68  4.1  ' E ' zone V /V , o, and reflection coefficients versus AV and AV  5.1  Porosity vs aspect ratio for resistivity, AV (km/s) and a  153  A. l  Teleseisms recorded  168  B. l  A L B - B final model  B.2  L A S final model  173  B.3  E G M final model  174  p  s  p  p  .  S  96  172  xi  Acknowledgements  For Lynn: with my heartfelt thanks and admiration.  I would like to thank my research supervisor Dr. Bob Ellis for his constant support and encouragement during the course of this research. Thanks also go to Bob for allowing me to get at least one (1) point in (most of) our badminton games. Dr. Ron Clowes has also been a source of inspiration. Despite his busy schedule with LITHOPROBE and numerous graduate students, he always gave freely of his time and provided valuable comments and advice. Very special thanks go out to Bruce and Evelyn Knowles of Port Alberni, Nancy and Kevin Monahan of Lasqueti Island, and Dick and Nan White of Egmont. They allowed me to deploy seismographs in their homes, and they maintained and diligently operated this equipment for two years, providing me with a first class data set. I also wish to thank R.D. Meldrum whose expertise and assistance were invaluable in the preparation and deployment of the instruments. M y committee members, Drs. Bob Ellis, Doug Oldenburg, Garry Rogers and Wayne Savigny have each contributed valuable comments during the course of this research. Also, Dr. Roy Hyndman provided insight into aspects of deep crustal fluids and porosity. The contribution of analysis programs by Dr. T.J. Owens is gratefully acknowledged. I also would like to thank Drs. Tom Owens, Charles Langston, George Zandt and Charles Ammon for sharing with me their knowledge and experience with receiver function analysis. The students, staff, and faculty in the. Department of Geophysics and Astronomy are second-to-none and make this a truly wonderful place to be. I am particularly indebted to Lydia and Carol for providing sunshine each day in sometimes rainy Vancouver; Mary, who  xii  is truly a great cook; Doug Oldenburg, who I believe is the cheeriest man on the planet Earth; John Amor, who provided invaluable assistance with computer programs, hardware problems, and climbing Mt. Baker; and Sonya Dehler for providing some of the computer plots used in this thesis and her valuable comments and advice. The contributions to this thesis provided by Colin and Barry Zelt, the 'Dave's' (Aldridge, Dalton and Lumley), and Ken Whittall will not be forgotten. Thanks also go to John and Judy Hole for their friendship, Tony Endres for the walks on the beach which helped to maintain my sanity while finishing this thesis, Yaoguo L i for swapping baby stories with me, and of course my fellow golfers. Those not mentioned here have not been forgotten -1 thank you all. Finally, thanks go to my family; my wife Lynn for her endless support (both financial and moral) and for encouraging me to pursue my dreams, my son Alexander for allowing me to keep the 'big picture' in perspective, my sister Alice for great bottles of wine, and my parents B i l l and Joy for a lifetime of support and for always encouraging me to ask 'why?'. Funding for this program has been provided by Energy, Mines and Resources Canada Research Agreement 89-84, the Natural Science and Engineering Research Council of Canada Grant A2617 and the British Columbia Hydro and Power Authority. The author was partially supported by a University Graduate Fellowship.  xiii  Chapter 1 INTRODUCTION  The world's largest earthquakes and much of its active volcanism are associated with subduction zones. The Cascadia subduction zone lies along the west coast of northern California, Oregon, Washington and southwestern British Columbia (Figure 1.1). Rogers (1988) estimates that this zone has the potential to generate earthquakes as large as any recorded in the world. Mounting geological evidence (Heaton, 1990) suggests that megathrust earthquakes have occurred here in the past. Knowledge of the geometry and physical properties of the subduction zone provides the key to understanding and predicting the potential strong ground motion which may be associated with such earthquakes. In addition, much of western North America consists of a series of accreted terranes (Jones et al., 1982). The earth structure from Vancouver Island to the British Columbia mainland reflects the history of the accretion of the last major terrane, Wrangellia, with the former continental margin. Constraints on the structure of this region are required to better understand the processes which have resulted in the westward growth of North America.  "  Little is known of the deep structure (> 40 km) of southwestern British Columbia. In 1987, a program was initiated to use locally generated P-to-S conversions (Ps) contained in the P coda of teleseisms to determine S-velocity to upper mantle depths with an overall objective to gain an improved understanding of the subduction process. Three 3-component broadband digital seismic stations were deployed in a 90 km long linear array oriented perpendicular to the continental margin (Figure 1.2). Between August 1987 and October 1989 approximately 100 teleseismic P-waveforms were recorded at each site. This dissertation  1  Chapter 1. INTRODUCTION  2  Figure 1.1: Geometry of the Cascadia subduction zone and related features (after Hyndman et al., 1990; Gabrielse and Yorath 1989). In the British Columbia region, the boundaries o f the major tectonic belts are indicated by dashed lines. The box encloses the area illustrated in Figure 1.2.  Chapter 1. INTRODUCTION  3  Figure 1.2: Tectonic setting of the study area and location of seismic stations (squares). The dashed line between Vancouver Island and the mainland represents the approximate boundary between the Insular and Coast Belts. The terranes which comprise the Insular belt are labelled.  Chapter 1. INTRODUCTION  4  describes the experiment, the analysis technique, the S-velocity models derived and the tectonic implications of this research.  1.1  Thesis Outline  This chapter describes the tectonic setting and current understanding of the study area, the results of previous geophysical studies, the specific goals of this research and the analysis technique. Chapter 2 describes the data acquisition, including instrumentation, site selection, station operation and the data processing steps required to generate stacked receiver functions which provide the basis for modelling. As this is one of the first broadband receiver function studies in a complex tectonic setting, Chapter 3 examines the application of this technique to the study of dipping interfaces. Specifically, the application of absolute amplitudes, the stability of reverberations, and stacking bounds suitable for use in a dipping layer environment are examined. In Chapter 4 the modelling method, and a simple technique which was derived to facilitate the comparison of receiver functions with seismic reflection data, are described. I then discuss the interpretation of the receiver functions at each station and present the S-velocity models. In Chapter 5 the structure which is common to each of the sites, and the implications of this research are discussed. In addition, a comparison is made with the results of previous studies. Finally, suggestions for future studies in this region are provided.  1.2  Tectonic Setting  The Cascadia subduction zone results from the convergence of the Juan de Fuca plate system with the North America plate (Figure 1.1). In the study area, the Juan de Fuca plate converges with the North America plate at the rate of 40-50 mm/a in the direction ~N56°E. It is noteworthy that the orientation of the margin changes from N - S to N W - S E between latitudes 47°N and 49°N. Rogers (1983) suggested that this may result in an upward arching of  Chapter 1. INTRODUCTION  5  the subducting plate in this region. Seismicity patterns in Washington state support this hypothesis (Crosson and Owens, 1987; Weaver and Baker, 1988). The region considered in this study lies above the north flank of the proposed arch; thus the subducting plate may have a more northerly dip than the plate convergence vector suggests. Like most of western North America, southwestern British Columbia has resulted from the collision of numerous crustal blocks (known as terranes or microplates) with the ancient continental margin during the past 200 M a (Jones et al., 1982). Most of the smaller microplates had amalgamated into 2 superterranes, the Intermontane Belt and the Insular Belt (Figure 1.1), prior to their accretion to North America in Middle Jurassic and Early Cretaceous time, respectively (Monger et al., 1982). In southwestern British Columbia the Insular Belt consists primarily of the Wrangellia Terrane - several sequences of igneous, sedimentary and metamorphic rocks of mid-Paleozoic to Mesozoic age (Muller, 1977). Paleomagnetic results indicate that Wrangellia originated several thousand kilometres to the south of its present position relative to the North American craton (Yole and Irving, 1980). The Coast Belt, dominated by granitic plutons and batholiths and high grade metamorphic rocks, lies between these 2 superterranes. The extensive plutonism, which may have been initiated by the accretion of the Insular Belt to the former continental margin (Monger et al., 1991), lasted at least 80 M a (-120-40 Ma). As a result the eastern boundary of the Insular Belt is not clearly defined. However, it is believed that Wrangellia extends beneath the westernmost part of the Coast Belt (Monger, 1991; Dehler, 1991). Subsequent to the accretion of the Insular Belt, thrusting and northward translation along strike-slip faults has occurred in this region (Coney et al., 1980). The Pacific Rim and Crescent terranes (Fig. 1.2) were accreted to Wrangellia in the early Eocene. A change in the plate tectonic regime about 43 M a established the present convergent margin. Thrust faults which penetrate very close to the top of the oceanic crust indicate that most or all of the incoming sediments are presently being scraped off the subducting Juan de Fuca plate  Chapter 1. INTRODUCTION  6  (Hyndman et al, 1990).  1.3  Previous Geophysical Studies  During the past two decades numerous geophysical studies have been conducted across the northern Cascadia subduction zone in the vicinity of Vancouver Island. Geophysical data for this region include seismic reflection and refraction, magneto telluric, gravity, magnetic, heat flow and seismicity. Results of previous studies are summarised here, with emphasis on those which are most relevant to this research (i.e. refraction P-velocity models, seismic reflection, seismicity and S-velocity studies).  1.3.1  Seismic Reflection Data  Numerous onshore and offshore reflection lines have been recorded in the vicinity of this study. Those pertinent to this research are shown in Figure 1.3a. Detailed interpretations and discussions of these, given in Hyndman et al. (1990), Clowes et al (1987a, b), Green et al (1986) and Yorath et al (1985), have provided much information on the structure and tectonic history of the region, including the location of the subducting Juan de Fuca plate and the emplacement of allochthonous terranes. More recent interpretations describe the growth of the accretionary prism (Davis and Hyndman, 1989) and the importance of fluids in this region (Hyndman, 1988; Calvert and Clowes, 1991). The most consistent features observed in the data set are the top of the subducting Juan de Fuca plate and two bands of reflectors known as the ' C and ' E ' reflective zones (Figure 1.3b). The top of the plate is clearly imaged on the offshore lines 85-01, 85-02 and 85-05, and the westernmost section of lines 84-01, 84-02 and 84-04. West of the deformation front, the plate dips at 3-4°, increasing to 8-10° at a depth of 25-30 km beneath western Vancouver Island.  ALB-B  LAS  EGM  Figure 1.3: (a) Location map of previous studies. Squares denote the location of seismic stations, circles represent other stations discussed in text. LITHOPROBE reflection lines pertinent to this study are labelled. The dashed line A - B denotes the location of the cross-section of (b), and is coincident with the refraction line discussed in text, (b) Cross-section across the northern Cascadia subduction zone {after Hyndman, 1988). Thick solid and dashed lines represent the observed and inferred location respectively of the subducting Juan de Fuca plate. Isotherms are thin solid lines. Earthquake hypocentres (recorded within a 100 km wide corridor centred on A - B ) are shown as dots and the stippled area represents a region of high electrical conductivity determined from magnetotelluric studies. Short lines are reflectors.  Chapter 1. INTRODUCTION  8  Reflective zone ' E ' , the most prominent of the reflective bands, is imaged approximately 510 km above the top of the subducting plate. It is approximately 3-5 km thick, dips -9-13° to the NE, and has a strongly laminated character (Green et al., 1986). The depth to the top of this reflective zone along line 84-01 ranges from 23 km beneath western Vancouver Island to 34 km beneath central Vancouver Island. The reflective ' E ' zone has also been imaged in the offshore reflection lines 85-05 and 85-01 (Hyndman et al, 1990; Calvert and Clowes, 1990) and appears to merge with the subduction decollement. Calvert and Clowes (1990) estimated reflection coefficients as high as 11-22% for this zone. The reflective ' C zone is very similar to the ' E ' reflective zone. It also has a laminated character, it is 3-5 km thick, and dips -5-8° to the NE. The top of this zone lies at a depth of 11 km beneath western Vancouver Island, and 20 km beneath central Vancouver Island (Green et al, 1986). Reflection data collected on the mainland (line 88-16) reveal numerous reflective bands extending to upper mantle depths (Figure 1.3b). Clowes (1990) speculates that the deepest band of reflectors may correlate with the ' E ' reflective zone beneath Vancouver Island. The reflections end abruptly at the northeast end of the line where the geothermal data indicate a rapid rise in the heat flux (Lewis et al, 1988).  1.3.2 Seismic Refraction Data The interpretation of seismic refraction data collected in the vicinity of Vancouver Island suggests that an "atypical" crust underlies this region. Based on the lack of a clear P„ arrival to distances of 360 km, White and Savage (1965) and Tseng (1968) concluded that beneath Vancouver Island the crustal thickness is at least 50 km. In 1980, a major onshore-offshore refraction experiment was conducted in the region. Interpretations of various portions of this data set are presented by Waldron et al. (1990), Drew and Clowes (1990), McMechan and Spence (1983) and Spence et al. (1985). In addition to providing constraints on the  Chapter 1. INTRODUCTION  9  position of the subducting plate offshore, the Spence et al. (1985) study interpreted a wedge of high-velocity material (V -7.7 km/s) at depths of 20-25 km beneath western and central p  Vancouver Island (Figure 1.4a) as accreted or underplated oceanic lithosphere. In addition, although only weakly constrained, the preferred model of McMechan and Spence (1983) has a broad negative gradient low-velocity zone ( V = 6.95-6.2 km/s) in the lower crust and a p  fiat lying continental Moho at 37 km depth. The final velocity model of Spence etal. (1985) which incorporates the interpretations of Waldron et al. (1990) and McMechan and Spence (1983) is given in Figure 1.4a. Drew and Clowes (1990) re-interpreted portions of the refraction data set with the goal of incorporating the LITHOPROBE reflection data. It was recognised that the high-velocity wedge coincided with the region bounded by the reflective ' C and ' E ' zones. They interpreted the thin reflective layers as low-velocity zones (V = 6.35 km/s) above and below the p  • wedge, and preferred a V of 7.15 km/s for the wedge. However, it is noted that the velocity p  contrast associated with the ' C and ' E ' low-velocity zones is based on reflected rays from these layers, and is poorly constrained. In addition, they favoured a positive rather than a negative gradient for the low-velocity zone in the continental crust. The final velocity model of Drew and Clowes (1990) is presented in Figure 1.4b. With the exception of the high-velocity wedge beneath western Vancouver Island, both the Spence et al. (1985) and the Drew and Clowes (1990) interpretations are weakly constrained at depths greater than - 20 km, with the easternmost portion of these models (near the station EGM) being the most poorly constrained. Finally, it should be noted that the refraction data only provide constraints on the position of the plate offshore. The depth of the plate beneath Vancouver Island and the mainland is inferred from the results of other studies including seismicity and reflection data.  Chapter 1. INTRODUCTION  10 CQ • CD _J  a  DISTANCE (KM)  West  CO  <  <  V a n c o u v e r Island 80  120  160  |200  240  (3 W East Mainland 320  360  2  CL UJ Q  m1  CO <  CQ  k  West  DISTANCE (KM) 160  V a n c o u v e r Island |200  240  I  2801  o LU. ~ East Mainland 320  360  2  a. UJ  a  Figure 1.4: (a) Refraction interpretation of Spence et ai, 1985 (after Drew and Clowes, 1990). (b) Refraction interpretation of Drew and Clowes (1990) along a line coincident with the cross-section A - B of Figure 1.3. P-velocities (in km/s) are shown. Vertical lines denote the approximate position of the coastlines and (M) signifies the continental Moho.  Chapter I. INTRODUCTION  1.3.3  11  Seismicity Data  Earthquakes occurring within the study area fall into two categories; shallow events in the depth range 0-25 km, and a thin band of deeper, dipping events (Figure 1.3b). The lower bound of the shallow events is close to the 400°C isotherm. The temperature limit for brittle failure in the crust is typically 350° ± 100°C (Wong and Chapman, 1990), thus this lower boundary may be thermally controlled (Rogers et al, 1990). The deeper events lie in an approximately 10 km thick NE dipping band. The top of this zone is -25 km beneath the west coast of Vancouver Island and 60 km beneath central Georgia Strait. The dip angle increases gradually across this region from 5-10° to 20-25°. Off the west of Vancouver Island this band of events lies just below the 'F' reflector, interpreted as the top of the Juan de Fuca plate. Beneath eastern Vancouver Island and the Strait of Georgia the deep events define the position of the cold core of the downgoing plate (Rogers et al., 1990); however the exact depth of the top of the oceanic crust is uncertain. Taber and Smith (1985) interpret a similar deep band of earthquakes in the Puget Sound region as being below the oceanic Moho. 1.3.4 Geothermal, Magnetotelluric and Potential Field Data The geothermal heat flux across the northern Cascadia margin is described by Lewis et al. (1988) and Davis et al. (1991). These data indicate that the heat flow decreases smoothly from the shelf to a point approximately 20 km seaward of the volcanic arc, where it rises abruptly by a factor of three. Isotherms estimated from these data (Hyndman, 1988) are given in Figure 1.3b. He interprets the top of the 'E' reflective zone to be nearly isothermal at 400°-450°C and the top of the ' C reflective zone to be -250°. Magnetotelluric (MT) data collected across Vancouver Island near reflection line 84-01 have been interpreted by Kurtz et al. (1986). They image a dipping conductive layer which  Chapter 1. INTRODUCTION  12  correlates with the reflective ' E ' zone (Figure 1.3b) beneath Vancouver Island. Their data also requires a conductive region in the lower crust beneath the mainland. In a more detailed analysis Kurtz etal. (1990) conclude that the electrically conductive zone interpreted beneath Vancouver Island must extend beneath the Strait of Georgia and be in electrical contact with the mainland crustal conductor. Magnetic and gravity data collected throughout the northern Cascadia subduction zone have been interpreted to delineate the positions of the Pacific Rim and Crescent Terranes as well as differing rock units of the Insular and Coast Belts (Dehler, 1991). In a previous gravity interpretation, Riddihough (1979) required the presence of high-density material (-3.3 g/cm ) 3  as shallow as 25-30 km beneath Vancouver Island. Dehler (1991) concludes that much of this material may be concentrated in the crust below 10 km depth.  1.3.5 Shear-wave Data S-velocities for southwestern British Columbia have been estimated by Wickens (1977) based on the inversion of long-period surface wave phase velocity data which have sampled travel paths several hundred kilometres in length. A profile along a path extending the length of Vancouver Island (PHC-VIC, see Figure 1.3a) resolves an upper crustal S-velocity discontinuity (AVs = 0.7 km/s) and two less pronounced boundaries (AV -0.3 km/s and 0.5 km/s) S  near depths of 30 and 50 km, respectively (Figure 1.5a). Although the long-period nature of this data set precludes the opportunity to examine relatively thin low-velocity zones, there is some suggestion in these results for lower S-velocities above the latter 2 discontinuities. There is significant scatter in this data set and uncertainties in the depth estimates are at least +5 km. In the only other S-velocity study in this region, Langston (1981) modelled P-to-S conversions contained in long-period P-coda recorded at Victoria (VIC, see Figure 1.3a). The  Chapter 1. INTRODUCTION  V (km/s)  a  s  2  3  ~i——'  13  V (km/s) s  4r~  AO  PHC-VIC 80  £ 120 CL  a;  Q  160  200  240  Figure 1.5: Some results of previous S-wave studies in the region, (a) S-velocity depth profile (with error bars) for the travel path P H C - V I C determined by Wickens (1977). Offset vertical bars show the depth resolution, horizontal bars denote the variance in S-velocities. (b) S-velocity model of Langston (1981) which satisfies long-period body wave data recorded at VIC.  Chapter 1. INTRODUCTION  14  main feature of his model is a high-velocity contrast interface (AVs = 1.3 km/s) near 45-50 km depth forming the base of a pronounced low-velocity zone (Figure 1.5b).  1.3.6 Summary and Problems to be Addressed To summarise, the top of the subducting Juan de Fuca plate is well defined by reflection data offshore and beneath western Vancouver Island (Figure 1.3b). Beneath eastern Vancouver Island and the Strait of Georgia the deep seismicity defines the cold core of the downgoing plate; however the exact depth of the top of the oceanic crust is unknown. Prominent reflective zones are observed in the crust beneath Vancouver Island and to upper mantle depths beneath the mainland. The ' E ' reflective zone, 30-35 km beneath central Vancouver Island, is associated with a region of high electrical conductivity and is approximately isothermal. Although the wealth of information provided by the numerous geophysical and geological surveys of the northern Cascadia subduction zone has greatly increased our knowledge and understanding of this region, the deep structure (> 40 km) is still largely unknown and poses many interesting questions. The depth of the top of the Juan de Fuca plate beneath central Vancouver Island and the Strait of Georgia is poorly defined, and the subduction geometry (dip direction and dip angle) is unknown. Refraction data indicate an 'accretionary' or 'underplated' type structure beneath Vancouver Island and a more typical 'continental' structure beneath the mainland; however, the exact location and nature of the boundary between the two is unknown. The continental Moho beneath Vancouver Island remains enigmatic. Finally, the deep reflective zones are the most puzzling features of this region. Are they continuous from the locus of subduction to the upper mantle beneath the mainland? Are they associated with regions of anomalous seismic velocity? Some interpretations for the ' E ' reflectors include a zone of decoupling between the North America and Juan de Fuca plates (Clowes et al., 1987b), saline fluids trapped in a porous layer overlain by an  Chapter 1. INTRODUCTION  15  impermeable boundary (Hyndman, 1988), and a major shear zone above the subducting Juan de Fuca plate (Calvert and Clowes, 1990). A similar band of reflectors has been observed in the subduction zone in southernmost Chile (Vera et al., 1989). Thus, understanding the nature and origin of these reflectors may be important on a global scale. A n ideal technique with which to tackle these questions is receiver function analysis. This method is sensitive to the S-velocity structure; it has been used to resolve boundaries to depths of 60 km (e.g. Owens et al., 1987); it provides site specific information and is sensitive to interface geometry (i.e. both dip angle and direction). The detailed S-velocity structure provided by this study will be a new and important source of information which when combined with the results of previous studies will provide additional constraints on the structure and physical properties of this subduction zone.  1.4  Receiver Function Analysis  It has long been recognised that teleseismic P-waveforms recorded at a 3-component seismic station contain valuable information on the local earth structure. The earliest studies of this type (e.g. Phinney, 1964) were conducted in the frequency domain, employing spectral ratios to constrain crustal structure. These studies were largely abandoned in favour of time domain analysis when the advantages of the latter were pointed out by Burdick and Langston (1977) and Langston (1977a). One important advantage is the ability to associate specific arrivals in the waveform with specific velocity-contrast interfaces in the earth. Langston (1979) proposed a technique which is simple, but very effective at suppressing the earthquake source function. The primary assumption is that for teleseisms, which are steeply incident (A > 30°), the vertical component of motion is dominated by source and path effects. By deconvolving this component of motion from the horizontal components of motion, locally generated S-waves contained in the P-coda are isolated. The amplitude and  Chapter 1. INTRODUCTION  16  arrival time of these phases are sensitive to S-wave velocities and layer thicknesses. The polarity indicates whether the impedance contrast is positive or negative. This method, which forms the basis of all current receiver function studies, was initially applied primarily to long period data to constrain earth structure in a wide variety of geologic settings, including the Cascadia subduction zone (as described in section 1.3.5). The advent of portable, digital broadband seismometers further increased interest in this technique. To take full advantage of the increased resolution capability of broadband waveforms, Owens (1984) developed a formal inversion routine to determine fine scale (1-2 km thick layers) earth structure. Assuming homogeneous, horizontal layers, this routine has been successfully used at many sites (e.g. Owens et ai, 1987). Zandt and Owens (1986) claim that this technique provides results comparable to those obtained from refraction studies. Gross lateral variations in the earth structure about the recording site have been examined (e.g. Owens, 1984) by analysing receiver functions corresponding to different back azimuths (e.g. NW, SE, SW) and developing acceptable earth models for each. Recently, receiver function analysis has been applied to more complex tectonic settings where the assumption of homogeneous, horizontal structure does not apply (Ammon, 1985; Owens et al, 1988; Lapp et al., 1990; Langston, 1989). By forward modelling data corresponding to 2 back azimuths, Owens et al. (1988) and Lapp et al. (1990) were able to place constraints on the geometry of the Juan de Fuca plate beneath western Washington state. Mathematically, the vertical, radial and transverse ground displacement resulting from a plane P-wave impinging at the base of an earth model can be represented by  D (t) z  =  S(t)*I(t)*E (t)  D (t)  =  S(t)*I(t)*E (t)  D (t)  =  S(t)*I(t)*E (t)  R  T  z  R  T  (1.1)  Chapter 1. INTRODUCTION  17  where S(t) is the effective source time function, I(t) is the instrument response (which is identical for all 3 components in this study - see section 2.1), and Ez(t),En(t) and Exit) represent the impulse response of the local earth structure and * represents the convolution operator. The effective source function includes any arrival having a slowness value similar to that of the direct P-wave (e.g. near-source and secondary phases such as pP and PcP). Langston (1979) suggested that Ez(t) may be approximated as a Dirac delta function:  E (t)~m  (i.2)  z  There is obviously some error introduced by this assumption.  However it is tolerable  (Langston, 1979; Owens, 1984), even in the presence of large velocity contrast interfaces (AV ~ 3 km/s) provided synthetic and observed data are treated in the same manner. P  In practice, this assumption removes P-reverberations associated with horizontal interfaces from the radial receiver functions. This is not true for P-reverberations associated with dipping layers (see section 3.2.4). Combining equations 1.1 and 1.2 it is apparent that the vertical component of ground motion is dominated by the parameters which need to be suppressed  D (t)~S(t)*I(t) z  (1.3)  Therefore by combining equations 1.1 and 1.3 one can estimate radial and transverse receiver functions by deconvolving Dz(t) from the radial and transverse components of ground motion, respectively. The following derivation is for the radial receiver function. Transverse receiver functions are calculated in exactly the same way (substitute the subscript'T' for the  Chapter 1. INTRODUCTION  18  subscript ' R ' ) . The deconvolution is performed in the frequency domain using spectral division  -  p E  R  m  n  - r n m ' m ^ )  (  L  AS 4  )  To stabilise this deconvolution the water-level method suggested by Helmberger and Wiggins (1971) is applied. This provides for a minimum allowable value for the denominator in equation 1.4. Finally a Gaussian spectral window, G(co), is applied in the frequency domain to remove high frequency noise. Thus, in practice the deconvolution is  <P(co) where ER(G}) is the estimate of ER(G}),  G(co)=e- ° t  2/4a2  and <)(OJ) = max(Dz((i))Dz(Gz), c • max[Dz(oi)Dz((ii)])  D (<&) is the complex conjugate of -Dz(co); O) is the angular frequency, ' c ' is the waterz  level parameter, expressed as a fraction of the maximum of the vertical power spectra, and a controls the width of the Gaussian filter. Transformation back to the time domain provides the estimate of the receiver functions. A well documented summary of potential errors introduced by noise in the data, the assumption that Ez(t) ~ 8(£), and the water-level stabilisation technique are provided in Owens (1984). In the time domain, Eji(t) may be thought of as the 'true' (or 'ideal') receiver function (i.e. a spike series) convolved with an averaging, or blurring function. Thus, equation 1.5  Chapter 1. INTRODUCTION  19  may be written as D (a)  A(co)  (1.6)  z  where A(co) is the averaging function given by  A(co) =  Z) (co)D ((D)G(co) z  z  <K<o)  (1.7)  A useful modification to the source equalisation procedure has recently been introduced by Ammon (1991) and is used in this study. While previous studies normalised both observed and synthetic receiver functions to unit amplitude, Ammon (1991) demonstrates that the true receiver function amplitude may be recovered by normalising the averaging function to unit amplitude in the time domain. In Chapter 3 the applications of absolute amplitudes are examined, and I demonstrate that there are important implications when attempting to resolve dipping structure. Finally, the time domain representation of the averaging function provides the resolution of the receiver function. For example any sidelobes resulting from the deconvolution are readily apparent in the averaging function. If significant sidelobes were present, the averaging function could be used with the synthetic transfer functions in the forward modelling stage to provide a better comparison with the observed data. This is discussed further in section 3.3.5.  1.5  Synthetic Receiver Functions  Synthetic seismograms are generated using a fast three-dimensional, ray-tracing scheme based on Langston (1977b). The earth model is parameterised in terms of a series of constant velocity, planar, dipping layers over a half-space. The P- and S-wave velocities, density, strike and dip angles, and thickness are specified for each layer in the model. Synthetic vertical,  Chapter 1. INTRODUCTION  20  radial and transverse seismograms are generated by specifying a back azimuth and ray parameter for the plane P-wave incident at the base of the model. A consecutive application of Snell's law in the local interface coordinate system allows rays to be traced to the receiver. Having determined path-lengths in each layer, the arrival time for each phase can be calculated. Amplitudes are calculated using wave potentials in the local coordinate system and applying the appropriate reflection or transmission coefficient at each interface. In addition to the direct P-arrival and Ps conversions, the synthetic seismograms may (at the modellers discretion) include the free-surface multiples associated with each interface. This ray-tracer is best suited to horizontal or moderately dipping structure (dip angles 5 < 30°). Steeply dipping boundaries may intersect, resulting in geologically unrealistic earth models. Synthetic radial and transverse receiver functions are then generated by treating the synthetic seismograms in the same way as the real data. For the synthetic data presented in Chapter 3, an a value (Gaussian pulse width) of 5.0, corresponding to an upper frequency limit of ~1 Hz, was generally used. When generating synthetic receiver functions to be compared with observed data (Chapter 4), an a value which provided an optimum match to the frequency content of the observed stacked receiver functions was chosen (see section 3.3.5).  Chapter 2  DATA ACQUISITION AND PROCESSING  2.1  Instrumentation  Three-component event-triggered seismographs were deployed at each site. The Guralp CMG-3 broadband seismometer system uses a horizontal boom with a 0.16 kg inertial mass supported by a triangular leaf spring for sensing vertical ground motion, and inverted pendulums for sensing horizontal (N-S and E - W ) motions. Each system consisted of 3 seismometers and associated electronics and a control/filter unit which distributed power and provided signal conditioning. Long-period temperature or pressure-induced mass drift was compensated by automatic mass recentring, a very useful feature for a field deployment. A ±15 V A . C . power supply capable of providing 0.5 A powered the seismometer system. The theoretical velocity pass-band of the seismometer system is fiat from 0.05 to 10.0 Hz and the dynamic range is 136 db. Instruments were calibrated prior to commencing, and several times during the field deployment. By injecting a known sinusoidal current into the calibration circuitry of the control/filter unit and recording the seismometer output, the phase response could be measured directly. The magnification was calculated knowing the input current, the calibration coil constant and the seismometer output. Measurements were made over the frequency range 0.02-10.0 Hz. In all cases the measured response compared favourably to the theoretical response calculated using the transfer functions provided by the manufacturer (Figure 2.1). Within measurement uncertainties each set of instruments had a matched response (this  21  Chapter 2. DATA ACQUISITION AND PROCESSING B r o a d b a n d Velocity  22  Response  ICP  u >  10  2  Solid line — Theoretical Error Bors - Meosured •  I  I  I  i  I I I I  10-'  10"  10  10°  1  Frequency (Hz)  Phase  Response  50  v  o  0L  -50  -100  -150 Solid line — Theoretical Error Bars -  Mecsured  -200 '  10-2  •  i  i  i  i  i i i  10  t  i  i  i  i  t i  11  10°  - 1  _i  i i ii ii 10'  Frequency (Hz)  Figure 2.1: Broadband velocity magnification and phase response of the Guralp seismometer systems. Dots and error bars represent measured values, solid lines represent theoretical values.  Chapter 2. DATA ACQUISITION AND PROCESSING  23  did not change during the course of this experiment), with the exception of very slight (23%) magnification differences between the horizontal and vertical components at A L B - B and L A S . This was compensated for during preliminary data processing (section 2.3). Teledyne Geotech MCR-600 Microcorders were used to record the output from the seismometers. These operated on float charged ±12 V internal batteries and therefore were not affected by A . C . power interruptions. Each microcorder contained an internal clock, a cassette tape drive, a keyboard and 5 microprocessors which allowed for programming, data monitoring and digital recording. Amplifiers were not required in this study and the only signal conditioning provided by the microcorders was an internal anti-aliasing low-pass filter with a corner frequency of 6.1 Hz and a roll-off of 24 db per octave. Recordings made at the test site PGC-B, demonstrated that most teleseisms had amplitudes of 0.1-0.5 V , very few were larger than - I V . Based on this, I chose to record over ± 5 V (although the Guralp systems output ±10 V) in order to improve resolution and avoid the introduction of bit noise by the recorder. During the 2 year recording period only one potentially useful teleseism had an amplitude > 5.0 V and was therefore clipped. The 3 data channels were digitised to 12 bit words at a sampling rate of 20 Hz, quite adequate for teleseisms having dominant frequencies of 0.2—2.0 Hz. Channels were sampled sequentially with the 1/60 s time skew being compensated for during preliminary data processing (see section 2.3). A short term average (STA), long term average (LTA) algorithm was used for event detection. The parameters used throughout the study were STA = 3.2 s, LTA = 102.0 s and the trigger threshold was set at a signal to noise (SNR) ratio of 4 to 1 (12 db). Only the vertical component was used for triggering purposes. Upon event detection, 30 s of pre-trigger noise and 143 s of post-trigger signal were recorded. Date, time and header information including station code and all recording parameters were recorded with the data on cassette tape. Each cassette had a capacity of 216,000 data samples, which at 20 Hz 3-channel recording represented 66 minutes or approximately 20-24 triggers. Although most teleseisms have the potential to  Chapter 2. DATA ACQUISITION AND PROCESSING  24  cause three triggers (P, S and surface waves) this was seldom the case. The majority of the teleseisms triggered on the P-wave only.  2.2  Station Description and Operation  Sites were chosen based on the availability of A . C . power, shelter, an acceptable seismic noise level and persons to maintain the equipment. Faults and sedimentary formations (i.e. the Nanaimo group of eastern Vancouver Island) were avoided where possible in order to reduce potential complexities in the receiver functions (Ammon, 1985; Owens and Crosson, 1988). The stations A L B - B and L A S are located on outcrops of the Upper Triassic Karmutsen formation (Figure 2.2), the thickest (up to 6000 m) basaltic sequence of Wrangellia (Sutherland Brown, 1966). A L B - B is located in the most complex geological setting of the three stations (Figure 2.2). There are numerous faults nearby, including the Beaufort Range fault 7 km to the N . The sedimentary Nanaimo formation may extend beneath this site and Jurassic-age Vancouver Island Intrusions composed of quartz diorite and granodiorite plutons and batholiths are located within 3 km to the N N W and SSW. E G M is located on an outcrop of the Cretaceous-Tertiary Coast Plutonic complex. There are no mapped faults in the vicinity of this station, however plutons and batholiths of quartz diorite and diorite are found only a few km to the N , E and W of this site. At each site, homes with bedrock exposed in the basement were found. A concrete slab approximately 1 x 1 x 0.1 m was poured directly onto the bedrock (after fractured rock was chipped away) and 3 small glass plates, used as seismometer stands, were set into the concrete and levelled. A typical station setup is shown in Figure 2.3. The seismometers, covered with 1 cm thick microfoam encased in aluminum foil, were positioned on the glass plates and a 5 cm thick foam box was placed over the concrete slab. The insulation reduced air currents and temperature variations, thereby decreasing the number of seismometer mass  B.C. Mainland  Vancouver Island  LAS  ALB-B  EGM  25 kr  Coast Plutonic Complex: (Late Jurassic-Early Tertiary)  I — I Nanaimo Group (Upper Cretaceous):  Karmutsen (Upper Triassic):  — sandstone, shale, conglomerate  basalt and pillow lava  I  I B o n a n z a (Lower Jurassic):  andesite, dacite, rhyolile  I—I Buttle Lake Formation (Permian):  — limestone  Vancouver Island Intrusions (Jurassic):  Sicker Group (Carboniferous):  granite, granodiorite, quartz diorite  meta-andesite, dacite  I gabbro, diorite, amphibolite  •  quartz diorite  diorite  Figure 2.2: Geology of the study area (after Roddick, Muller and Okulitch, 1979). Stations (labelled above) are denoted by a *. White areas represent bodies of water. Black lines represent faults; two of the longest in the region - the Beaufort Range fault (BRF) and the Cameron River fault (CRF) are labelled. The dashed line represents the inferred boundary between Wrangellia and the Coast Plutonic Complex.  Foam Box Encasing Seismometers Concrete Pier on Bedrock  Microcorder Battery Charger (Top) and Batteries (Bottom)  Figure 2.3: A typical station set-up (ALB-B). Seismometers and electronics are inside the tinfoil/foam enclosure sitting on bedrock, the recording system is to the right.  Chapter 2. DATA ACQUISITION AND PROCESSING  Station PGC-B ALB-B LAS EGM  Lat. +48.65 +49.28 +49.46 +49.75  Long. -123.45 -124.92 -124.21 -123.93  Start June 7 1987 Nov. 5 1987 Nov. 14 1987 Aug. 6 1987  End Nov. 12 1987 Oct. 12 1989 July 24 1989 Oct. 5 1989  27  Efficiency 92.4% 99.7% 78.4% 80.0%  Table 2.1: Station locations and. recording periods  recentres. This reduced the frequency of tape changes as the mass movement associated with each recentre caused a microcorder trigger. At L A S , A . C . power was not available and the seismograph was powered by two +12 V batteries. These powered the microcorder directly and provided +15 V to the Guralp system via a D.C. to D.C. converter. The batteries were recharged on average, every 3—4 days using a gas powered generator. At each station, homeowners maintained the equipment, ensuring that the recorder and clock were operational and checking the status of the data tape 2-3 times each week. At A L B - B and E G M the equipment operated at ambient temperatures of ~10-20°C, at L A S the temperature range was ~0-20°C. Tapes needed to be replaced on average about every 4—6 weeks. This depended primarily on the number of false microcorder triggers. These represented about 50% of the total triggers and were generally either temperature induced mass recentres or mass movement resulting from A . C . power outages. A more sophisticated triggering algorithm could easily reduce the false trigger rate. A l l equipment operated exceptionally well during the course of this study. Table 2.1 provides the location, recording period and overall efficiency (number of hours normal operation / total number of hours of deployment) for each array station and the test site PGC-B. The stations were visited every 3-6 months in order to re-set the microcorder clocks and calibrate the instruments. The clocks drifted between 1-3 s per month depending on  Chapter 2. DATA ACQUISITION AND PROCESSING  28  ambient temperature fluctuations. This was not a problem as absolute time was only used for identification of events and not required in the analyses of the data.  2.3 Preliminary Data Processing Preliminary processing of the raw cassette tapes involved first demultiplexing the 3-channel data and compensating for the 1/60 s time skew which resulted from sequential sampling. A uniform sample time was necessary prior to rotating the horizontal data into their radial and transverse orientation. Linear interpolation was applied to the horizontal components to align their sample times to that of the vertical component. Data were compensated for any slight differences between microcorder channel response as well as for any calibration differences between components at a given station. At A L B - B and L A S a 2—3% difference in magnification existed between components. The vertical component magnification was used as a reference and the appropriate multiplicative factor was applied to the horizontal signals. Finally, the data were combined with header information including date, time, station coordinates and relevant earthquake parameters and stored on 9-track tape.  2.4 The Data The method of triggered recording provided a very effective means of collecting data in this study. Examples of high and low quality events are shown in Figure 2.4a and 2.4b respectively. Note that the microcorder was able to trigger on very emergent events as shown in Figure 2.4b. Low SNR events such as this yielded receiver functions dominated by noise (see section 2.5) which were not used in this analysis. Continuous recording would permit low SNR earthquakes to be collected, however the resulting receiver functions would also be of low quality. It would be necessary to stack many of these before they could be  -20  0  20  40  -20  Time (s).  Time  HIGH SNR RECEIVER FUNCTIONS  -20  0 Time (s)  0  20  20  40  (s)  LOW SNR RECEIVER FUNCTIONS  40  -20  0 Time  20 (s)  Figure 2.4: Examples of recorded earthquakes and corresponding receiver functions: (a) a high SNR trigger and (b) a low SNR trigger (the arrowhead indicates the interpreted P-arrival). Both earthquakes were recorded at L A S . Receiver functions have been normalised to the vertical component of the direct P-wave (see scale on lower trace).  40  ' Chapter 2. DATA ACQUISITION AND PROCESSING  Station Code PGC-B ALB-B LAS EGM  Recording Period (Months) 5 23 20 26  P 25 98 71 83  Teleseisms S Surface 5 10 10 36 8 27 9 38  30  Regional  Local  10 23 12 23  4 12 8 14  Table 2.2: Summary of events recorded  modelled with confidence. However, even small (m^ -5.5) earthquakes are infrequent in the areas in which data are lacking (Europe, eastern North America and the mid-Atlantic Ridge) and it is doubtful that continuous recording would have provided a better data distribution over the 2 year recording period. In the case of a permanent station, continuous recording over a number of years would be a definite advantage. The ability to trigger on earthquakes depended on the background noise level. During the winter months (November-April) large amplitude 5-6 s microseisms were often present and generally only earthquakes of rm, > 6 were recorded. During the summer months earthquakes of mfc = 5-5.5 were occasionally recorded. Table 2.2 summarises the number and types of events recorded at each station. A total of 145 teleseisms covering a wide distance range (A = 30° to 100°) and most back azimuths (BAZ) were recorded during the course of this study. Figure 2.5 illustrates the data distribution for the station A L B - B , note that only the N E (BAZ = 10° to 80°) and SSW (BAZ = 170° to 230°) directions are poorly represented in this data set. Parameters for all teleseisms recorded in this study are given in Table A. 1 of Appendix A . Although only teleseismic P-waveforms are analysed in this study, Table 2.2 illustrates the wealth of data, ranging from high frequency local events to long-period surface waves, provided by the broadband sensors. The frequency content of teleseisms recorded in this study is typically 0.2-0.5 Hz. Examples of what are considered a 'low frequency' event (-0.1-0.3 Hz), an 'average frequency'  Chapter 2. DATA ACQUISITION AND PROCESSING  31  ALB-B DATA DISTRIBUTION Azimuthal Equidistant Projection  N 0°  180° Figure 2.5: Data distribution for A L B - B . Dots represent recorded earthquakes and the star denotes the array location. Back azimuth is indicated on the outermost circle. The two inner circles represent epicentral distances of 30° and 100° respectively. The stations L A S and E G M have similar data distributions.  Chapter 2. DATA ACQUISITION AND PROCESSING  32  event (-0.2-0.5 Hz), and a 'high frequency' event (-0.2-2.0 Hz) are illustrated in Figure 2.6a-c respectively. The limitation on resolution of the earth structure imposed by this frequency band is examined in the following chapter.  2.5  Receiver Function Estimation  For each teleseism lying within the distance range 30° < A < 100°, radial and transverse receiver functions were calculated using an algorithm provided by T. Owens (based on the method of Langston, 1979). After removing the D.C. offset from the data, horizontal components were rotated, based on source locations obtained from the Preliminary Determination of Epicenters (PDE's), into radial and transverse components. Prior to deconvolution a 3 s Hanning taper was applied to the ends of each trace. Numerous deconvolutions were performed for each event in order to examine the stability of the deconvolution and to choose the optimum parameters which provided the highest SNR receiver functions. A range of waterlevel parameters from 0.00001-0.1 were applied and trace window lengths ranging from 50 s to 178 s were considered. For the majority of the data, a waterlevel parameter of 0.001 was used (this represents a trough filler of 0.1% of the maximum vertical component spectral power). The length of the data trace did not affect the deconvolution for most events. However, in some cases a shorter window length could remove secondary phases such as PP from the signal and improve the SNR of the deconvolution. In general, the data were not filtered prior to processing. This is due to the fact that the spectra of the noise (typically 0.17-0.5 Hz) lies within the signal spectral window (-0.1-1.0 Hz). In some cases filtering was necessary due to long period drift (-.03-05 Hz) of the seismometers. For these events a 0.07 Hz high-pass filter was applied. A gaussian width (a value) of 5.0 was used for most deconvolutions, this removed high frequency noise (> 1.0 Hz) from the receiver functions. For those few events containing signals of 1-2 Hz  Chapter 2. DATA ACQUISITION AND PROCESSING  33  Event #80 Power  Spectra  a) "Low" Frequency (0.1-0.3 Hz) Event #80  iff  o-  9  i  10-  I  I  '  1  :  Ml  1  10"' 10° Frequency (Hz)  2  Event #70 Power  1 1 1 1 1 !  10'  Spectra  b) "Average" Frequency (0.2-0.5 Hz) Event #70  20S  '  ' ' I'M  10~  I 2  I  I I I llll  10"' 10° F r e q u e n c y (Hz)  Event".#83 Power  10'  Spectra  c) "High" Frequency (0.2-2.0 Hz) Event #83  20 S 10  10"'  10°  F r e q u e n c y (Hz)  Figure 2.6: Examples of vertical power spectra for teleseismic P-waveforms. Vertical-component seismograms are shown to the left of the spectra for: a) a "low frequency" event; b) an "average frequency" event; and c) a "high frequency" event. Dotted lines on spectral plots denote the waterlevel ('c') used in the deconvolutions (not shown). A l l events were recorded at A L B - B .  Chapter 2. DATA ACQUISITION AND PROCESSING  Vertical  0  20s  Radial  0  34  Radial Receiver  20s  0  20s  Figure 2.7: A n example of the source equalisation procedure for two earthquakes in the same epicentral region but at different focal depths. The shallow event has a more complex effective source function (i.e. note the near-source reverberation pP near 10 s on the vertical component). The radial receiver functions (right) are nearly identical, illustrating the effectiveness of this procedure in suppressing source effects. Both earthquakes were recorded at ALB-B. an a value of 7 was used (removing noise of f > 2.0 Hz). The source equalisation method is an effective means of suppressing source and transmission effects. This is illustrated in Figure 2.7 where two different looking events (from the same approximate source region but significantly different depths), have very similar receiver functions after applying source equalisation. This provides confidence that receiver functions are dominated by locally generated phases and not the earthquake source-time function or near-source phases. Examples of what are considered good, average and poor quality radial receiver functions are. given in Figure 2.8. To suppress spurious arrivals and improve the SNR, receiver functions are stacked within a range of 10° in both B A Z and A (this is discussed further in section 3.2.2). Prior to stacking, each receiver function was assigned a weight based on the energy ratio of the direct arrival to  -20  0 Time (s)  20  40 - 2 0  0  20  40  Time (s)  Figure 2.8: Examples of what are considered "good", "average" and "poor" quality receiver functions. These examples represent different source regions and therefore the traces should not be the same. A l l receiver functions have been normalised to the vertical component of the direct P-wave and have been plotted at the same scale (lower left). Events were recorded at A L B - B .  Chapter 2. DATA ACQUISITION AND PROCESSING  36  the noise. The window chosen for the direct arrival was centred at T = 0.0 s and was based on the full-width half-maximum of the pulse; the noise window extended from -30.0 s to -5.0 s. For each stack the mean and ±1 standard deviation are determined. A n example of a stack of 8 events is given in Figure 2.9. The standard deviation is not shown in this diagram. The similarity of these deconvolutions is remarkable,,and again provides confidence that the receiver functions are dominated by the near-receiver structure.  Chapter 2. DATA ACQUISITION AND PROCESSING  37  E X A M P L E OF STACKING SUITE A L B R A D I A L BAZ=230°, A=80°  -10  0  10  TIME  20  30  (s)  Figure 2.9: Example of a stack of radial receiver functions. Dashed vertical lines mark 1 s time intervals. The mean of the stack is the bottom trace. The weight assigned to each trace is based on the SNR.  Chapter 3  R E C E I V E R FUNCTION ANALYSIS O F DIPPING S T R U C T U R E  3.1  Introduction  Prior to this research only a few broadband receiver function studies have been conducted in a dipping layer environment. Two papers (Owens etal., 1988; Lapp etal., 1990) demonstrate applications of this technique to constrain the geometry of the subducting Juan de Fuca plate beneath Washington state. Another (Owens and Crosson, 1988) documents the effects of shallow structure on receiver functions and the variation in amplitude and arrival time of P-to-S conversions as a function of back azimuth and epicentral distance. Subsequent to these studies, Ammon (1991) has introduced a modification to the deconvolution routine which provides for the absolute amplitudes of receiver functions. In previous studies only the relative amplitudes of receiver functions were modelled. I demonstrate the applications of absolute amplitudes for resolving dipping, and shallow structure. In examining the use of absolute, rather than relative amplitudes I determined that the latter may, under certain circumstances, result in inaccurate earth models. Two examples are documented. The use of absolute amplitudes, and the goals of resolving deep (-70-80 km depth), dipping structure necessitate the re-examination of stacking bounds. Also, prior to formally inverting receiver functions to determine the fine-scale earth structure in a dipping layer environment, one must be confident in their ability to accurately model each arrival contained in the waveform. As such, the stability and lateral sampling range of Ps phases and reverberations in the presence of dipping layers are examined.  38  Chapter 3. RECEIVER FUNCTION ANALYSIS OF DIPPING STRUCTURE 39  In addition, several points pertaining to receiver function analysis are examined: • The minimum detectable A V . S  • The ability to differentiate velocity gradients from I  ST  order velocity discontinuities.  • The resolution of thin layers. • The effects of very shallow structure (< 2 km) on receiver functions. • The sensitivity of Ps conversions to density and P-velocity contrasts. Finally broadband and simulated short-period receiver functions are compared. This may have implications for future studies as short-period instruments are less expensive and often easier to operate and maintain than broadband instruments. The following discussion applies to the general case of dipping layers with emphasis placed on parameters relevant to this study (e.g. data distribution, noise level and frequency content). A l l synthetic seismograms are generated using a ray-tracing algorithm (written by T. Owens) based on the theory (see section 1.5) developed by Langston (1977b). In calculating synthetic receiver functions, a Gaussian pulse width of a = 5 (corresponding to frequencies of approximately < 1 Hz) was generally used.  3.2  Receiver Function Analysis and Dipping Layers  Dipping layers manifest themselves in receiver functions in three ways. To illustrate this, consider a simple 2-layer over a half-space model with an east dipping boundary (8 = 15°) at 20 km depth and a horizontal boundary at 40 km depth (Figure 3.1a). The S-velocity contrast is 0.87 km/s for both interfaces. Figure 3.1b illustrates the radial and transverse receiver functions generated using this model. Clearly there are significant differences between Ps phases generated at the dipping boundary (phase D) and those generated at the horizontal  Chapter 3. RECEIVER FUNCTION ANALYSIS OF DIPPING STRUCTURE 40  boundary (phase H) as a function of back azimuth. The three dipping layer effects on receiver functions are: 1. Ps amplitudes and arrival times vary as a function of both back azimuth (Figure 3.1b) and A (not shown). Waves travelling updip (i.e. approaching from B A Z = 90° in this example) generate the largest and latest arriving Ps phases; waves travelling downdip generate the smallest and earliest arriving Ps phases. The arrival time variation as a function of B A Z is demonstrated more clearly in section 3.2.2. These variations in amplitude and arrival time provide constraints on the geometry (dip angle and direction) of dipping interfaces (Owens and Crosson, 1988; Owens et al., 1988). 2. Dipping layers deflect P and S-waves from the R - Z plane, thus introducing a transverse component of ground motion (Langston, 1977a). The transverse amplitude is zero for arrivals from the updip or downdip directions and largest for arrivals travelling along the strike direction of the boundary (0° and 180°). The polarity of the transverse Ps is antisymmetric about the interface dip direction. Some long-period studies (e.g. Langston, 1979) have successfully modelled transverse polarities to constrain the geometry of dipping boundaries. However, they have had less success at matching transverse amplitudes. Broadband studies have generally been restricted to a detailed analysis of radial data and a qualitative description of transverse data (e.g. Lapp et al., 1990). 3. Refraction of the P-wave at a dipping interface causes the amplitude of the radial component of the direct P-wave (P ) to vary as a function of back azimuth. However, r  the vertical component of the direct P-wave (P ) is a relatively stable quantity. Prez  vious studies have normalised receiver functions to unit amplitude, thereby neglecting this information. As demonstrated in Figure 3.1b, P is smallest for updip arrivals r  (i.e. waves travelling updip) and largest for downdip arrivals. Also, a dipping layer  Chapter 3. RECEIVER FUNCTION ANALYSIS OF DIPPING STRUCTURE  41  Seismic Station  w 20 km  P  km/s km/s g/cm 5.0  2.89  3  2.53  -  20 km i  6.5  3.76  2.80  8.0  4.63  3.30  Strike = 0° Dip = 15°  a  i  i i | i r  j i i i j i i r  Radial  Transverse  D A  H TV TV  A  TV.  A  TV-  .A.  TV. TVTV. TV_  I  I  0  I  I  I  I  I  1  4 Time (s)  I  BAZ 000 045 090 135 180  TV  225 270 .315  I L  8  H  D  J  0  J  I L  I L  4  8  Time (s)  Figure 3.1: (a) The earth model used in this example, (b) Azimuthal variation in radial and transverse receiver functions resulting from a dipping interface (see text). P represents the direct P arrival; D and H denote Ps conversions from an eastward dipping and a horizontal interface respectively. A l l receiver functions were generated using a ray parameter of 0.068 s/km (A = 45°).  Chapter 3. RECEIVER FUNCTION ANALYSIS OF DIPPING STRUCTURE 42  introduces a transverse component of motion for the direct P-arrival. The azimuthal variation of this transverse component of motion is similar to that described above for Ps phases. However, if there exist several layers having different dip directions, a very complex azimuthal pattern may result for both the radial and transverse direct P-wave.  3.2.1  Modelling Absolute Amplitudes of Receiver Functions  Previous broadband studies have normalised radial and transverse receiver functions to unit amplitude, thereby neglecting the information provided by the ratio P fP - Ammon (1991) r  z  describes a modification to the deconvolution procedure to preserve this ratio, which he notes is sensitive to the near surface velocity structure. I demonstrate that the use of absolute amplitudes is more robust and recommend that future studies apply this technique. First the effects of shallow structure on normalised and absolute amplitude radial receiver functions are examined. Consider the two models shown in Figure 3.2a. The only difference between these models is the 1.3 km thick low-velocity ( V = 1.73 km/s) surface layer. s  Figure 3.2b compares the normalised radial receiver functions for the two earth models. A l l multiples for the thin layer are included in this calculation. The near surface low-velocity layer broadens the direct P-arrival (i.e. the Ps and multiples generated at this boundary arrive shortly after the direct P-wave and are not resolvable) and causes an increase in the apparent amplitude of the Ps phases (Ps/P). This is due to the refraction of P- and S-waves towards the normal. This causes P to decrease in amplitude and Ps to increase. If one were not aware r  of the near-surface layer when modelling, the large 'apparent' Ps amplitude would lead to an overestimate of A V for the boundaries at 10 and 30 km depth. In Figure 3.2c the absolute S  amplitudes of the receiver functions are compared for the two earth models. In this case the shallow low-velocity layer decreases the ratio P / P , but does not alter the amplitude of the r  z  Ps phases associated with the boundaries at 10 and 30 km depth. Thus, when modelling the  Chapter 3.RECEIVER FUNCTION ANALYSIS OF DIPPING STRUCTURE S h a l l o w Structure  N o S h a l l o w Structure V  (km/s)  s  43  V  Z (km)  (km/s)  s  z (km)  —1.73—-  -1.3-  2.89  10.0 a  3.76 •30.04.62  1.2  T  I  I  I  I  I  I  I  I  I  I  I  I  I  I  I  I  I  I  I  I  I  I  ,  I  I  I  I  |  I  I  I  I  I  I  I  I  I  1  I  I  I  I  Solid: No Shallow Structure Dash: Shallow Structure  .9  b  I  I  9; .6  g£  .3  3 i i i i i i i i i i i i i i i i i ' -1 0 1 2  i  i  i  ii  i  i  i  3  i  i  4  i i i i i i 5  i i i i i i i i i ii 6 7 8  i  Time (s) .6 i  i  i  i  I i i  i  i  | i  i  i i | i  i  i  i  |  i  i  i  .4 N  c  *  i  |  i  i  i  i  |  i  i  i  i  j  i  i  i i | i  i  i  i_  Solid: No Shallow Structure JDash: Shallow Structure  a  < o -.2 i  -1  i  i  i  1 0  i  i  i  i  1  I  i  i  i  i  2  I  i  i  i  i  1  i  i  3  i  4  i  I  i  i  i  5  i  1  i  i  i  6  i  I  i  i  i  7  i  I  i  i  i  i  8  Time (s)  Figure 3.2: Potential errors introduced by normalising to unit amplitude in the presence of shallow structure, (a) Earth models used in this example, (b) Radial receiver functions normalised to unit amplitude, (c) Absolute amplitude radial receiver functions. If the shallow velocity structure is not resolved or is incorrectly modelled, normalising to unit amplitude may alter Ps/P ratios and result in an incorrect earth model (see text).  Chapter 3. RECEIVER FUNCTION ANALYSIS OF DIPPING STRUCTURE  absolute amplitudes of receiver functions, failing to identify or inaccurately modelling the shallow velocity structure will not affect the interpretation of Ps phases from deeper layers. This example uses an extreme case, but nonetheless illustrates an important point, in receiver functions normalised to unit amplitude the ratio of Ps/P is sensitive not only to the S-velocity contrast of the generating boundary, but also to a certain extent, on the shallow velocity structure (i.e. less than ~3 km depth). In contrast the ratio Ps/P is insensitive z  (Figure 3.2c) to the shallow structure and provides the 'true' amplitude of the Ps phase. This allows for an unambiguous determination of the S-velocity contrast associated with the generating boundary. In addition, as pointed out by Ammon (1991) the use of absolute amplitudes provides information on the shallow velocity structure. Figure 3.2c demonstrates that the low-velocity (V = 1.73) near surface layer may be recognised by the low P / P ratio of 0.16, compared s  r  z  to the value of 0.27 for the model having a surface S-velocity of 2.89 km/s. I have also determined that normalising receiver functions to unit amplitude in a dipping layer environment may result in inaccurate earth models. Consider two earth models. One is a two-layer reference model having a boundary (AV = 0.4 km/s) dipping 15° to the east. The S  other is identical, but in addition to the above, has a shallower boundary (AV = 0.25 km/s) S  dipping 30° to the south. This boundary has a small enough AV that it may not be resolved S  by my data set (see section 3.3.1) and must be considered a potential 'missed boundary'. Figure 3.3a illustrates the normalised amplitude of the Ps phase (i.e. Ps/P) generated at the deeper east dipping boundary as a function of back azimuth for both the reference model and the more 'complex' model. In this case the largest Ps/P value does not correspond to the true dip direction of the deeper boundary (BAZ = 90°) but is offset by approximately 20°. In addition the amplitude ratio is larger than expected from the reference model. If one were not aware of the shallow dipping boundary (which would probably be the case given the small A V associated with this interface) and modelled these data, both the dip direction S  44  Figure 3.3: Azimuthal variation of a Ps phase generated at an eastward dipping boundary. The reference model (Ref.) contains only a single eastward dipping interface; the more complex model (Comp.) is similar but also contains a shallower, southward dipping interface (see text), (a) Normalising to unit amplitude may result in an incorrect estimate for the dip direction and A V for the deeper boundary if the shallow boundary is not imaged, (b) The use of 'absolute' amplitudes eliminates this potential for error. S  Chapter 3. RECEIVER FUNCTION ANALYSIS OF DIPPING STRUCTURE  46  and A V g for the deeper boundary would be incorrectly estimated. Figure 3.3b demonstrates that these potential modelling inaccuracies may be avoided by modelling absolute amplitudes. In this case a potential 'missed boundary' has no effect on a Ps phase generated at a deeper boundary. Thus the use of absolute amplitudes helps to avoid potential inaccuracies in earth models for which shallow or dipping structure may be a factor.  3.2.2  Stacking Procedure  The amplitude and arrival time of Ps phases generated at dipping interfaces varies as a function of both back azimuth and A. The amplitude variation depends primarily upon the dip angle, whereas the arrival time variation depends upon both the dip angle and the depth of the boundary. In general, Ps originating at greater depths will experience larger arrival time variations for a given change in either back azimuth or A . In this study a primary objective is to map the descent of the Juan de Fuca plate beneath southwestern British Columbia. Across this region, the plate is believed to be at depths varying from 30 to 80 km with a dip angle ranging between 15° and 30°. Thus, it is important to ensure that the stacking procedure employed to improve the SNR of the receiver functions does not diminish or significandy alter a Ps conversion from a deep, dipping boundary. Throughout this thesis one of the parameters used to describe the data is A. However, when stacking it is important to consider the slowness, or ray parameter (p). For a given A, the ray parameter will depend upon the focal depth of the earthquake. For example at A = 80° earthquakes of focal depths 33 and 600 km will have ray parameters which differ by about 0.003 s/km. In practice, data are stacked within a range of p values, however to be consistent with the rest of this thesis, the stacking bounds are described in terms of both A and p.  Chapter 3. RECEIVER FUNCTION ANALYSIS OF DIPPING STRUCTURE 47  Layer No. 1 2 3 4  Vp km/s 6.50 8.00 7.00 8.00  Vs km/s 3.76 4.62 4.05 4.62  P g/cm 2.80 3.30 2.94 3.30  3  Thickness km 30 25. 5. oo  Dip Angle  Dip Dir.  00. 00. 20. 20.  00. 00. 90. 90.  Table 3.1: Reference model used for Ps stacking tests  Figure 3.4 illustrates the amplitude and arrival time variations as a function of back azimuth and ray parameter (p) for a Ps phase generated at an eastward dipping boundary (dip angle = 20°) at 60 km depth (Table 3.1). Using this model, I examine the stacking bounds of 20° in back azimuth and 10° in A (for the A range 45° - 60°) or 15° in A (for the A range 80°— 100°) as suggested by Owens (1984). These bounds were for use in a predominantly horizontally layered media and for stacking receiver functions which have been normalised to unit amplitude. When using absolute amplitudes, more significant amplitude variations occur (especially as a function of A), and thus a tighter range of stacking bounds may be required. The most rapid variation in both the amplitude and arrival time of a Ps phase generated at a dipping interface occurs along the strike direction of the boundary. It is noted that arrival time variations play a large role in the potential error introduced by stacking. For example for A = 45° the Ps (generated by the boundary at 60 km depth) arrival time may vary by as much as 0.21 s over a B A Z range of 20° (Figure 3.4); for A = 80° the corresponding time shift is 0.13 s. Thus, as pointed out by Owens (1984), tighter bounds should be applied to both B A Z and A when stacking events of A < 60°. In this study data have been stacked, with few exceptions, over a range of < 10° in B A Z and < 10° in A (or 0.005 s/km in p). In fact the average bounds for stacking in this study are 6° in B A Z and < 5° in A (or < 0.003 s/km in p). These bounds could be relaxed for more distant events (A > 60°). However, for reasons  Chapter 3. RECEIVER FUNCTION ANALYSIS OF DIPPING STRUCTURE  Ps ~\  i  i  i  |  i  Ps  vs B A Z i  i  i  I  i  i  i  i  I  i  r~  48  v s Ray Parameter  ~i—i—i—i—I—i—i—i—i—I—i—i—i—i—i—i—r  .15  o  r  i  i i  i  0 7.5  d\$  .05  .05  i  i i i_i  100  i  i  i  r~]  i i i i i i  200 Back Azimuth i  i  i  i  I  i  i  l  i _  i  i  i  |  i  v  I  I  I  I  I  I  I  I  .05 .06 Ray Parameter (s/km)  .04  300  1  1  1  1  1  1  1  1  I  -  I  L  1  1  1  1  1  1  .07  1  1  1  1  1  1  1  1  1  1  1  1  1  1  1  1  1  1  1  1  — E a, l  CL,  i  6.5  -  5.5 <—>0  i  < i  100  i  i  l  i  I  200 Back Azimuth  i  '  i  300  '  .04  .05 .06 Ray Parameter (s/km)  .07  Figure 3.4: Amplitude and arrival time variation as a function of back azimuth and ray parameter for a Ps phase generated at a boundary at 60 km depth and dipping 20° to the east (see Table 3.1). Note that the most rapid azimuthal variations occur near the interface strike direction^ Variations as a function of ray parameter (or A - see text) occur most rapidly for large values of the ray parameter (small A) and updip arrivals.  Chapter 3. RECEIVER FUNCTION ANALYSIS OF DIPPING STRUCTURE  49  outlined later in this section; data are always stacked over the smallest possible range. In Figure 3.5 a comparison is made between stacking over a range of 20° in B A Z and 10° in A and the average range of 6° in B A Z and 5° in A used in this study. In each case the top two traces represent synthetic receiver functions at the extremal bounds of both the B A Z and A range, the third trace is the receiver function corresponding to the middle of the stacking range. The top two traces are stacked to form the lower-most trace which ideally would be identical in appearance to the one immediately above it. Note that the direct P-wave and the Ps phase generated at the horizontal boundary are not affected by stacking. However, the Ps generated at the dipping boundary at 60 km depth has been attenuated by 38% and 6% respectively for the two stacking ranges considered. This example, with only 2 events stacked from opposite bounds in both B A Z and A in the region where amplitudes and arrival times vary most rapidly, demonstrates the potential hazards in stacking data when one is interested in Ps phases generated at deep dipping boundaries. In order to avoid attenuating such phases, the B A Z and A range used in stacking should be < 10° for both parameters (Ap < .005 s/km). Stacking over a small range of B A Z and A is also advantageous as it allows the identification of phases in receiver functions that vary rapidly in either amplitude or arrival time as a function of only slight changes in either B A Z or A. Such phases undoubtedly represent reverberations or scattered energy. This is demonstrated in Figure 3.6 where the two traces represent stacks of 8 radial receiver functions at A = 80° (solid) and 4 radial receiver functions at A = 90° (dash). The back azimuth for both stacks is 230°. Over this slight A range of 10° (Ap = .005 s/km), the significant amplitude and arrival time variation of the two phases labelled " Reverb ?" cannot be explained by Ps conversions generated at a planar dipping interface. These arrivals must therefore represent either reverberations or scattered energy and should not be quantitatively modelled. Another advantage of stacking within tight bounds is that it is easier to correlate arrivals from trace to trace.  Chapter 3. RECEIVER FUNCTION ANALYSIS OF DIPPING STRUCTURE  Stack Over 20° (BAZ); 10° (A)  0  4  Time (s)  8  50  Stack Over 6° (BAZ); 5° (A)  0  4  8  Time (s)  Figure 3.5: Examples of stacking receiver functions generated using the model of Table 3.1: over 20° in back azimuth and 10° in A; and 6° in back azimuth and 5° in A. In both cases the lowermost trace represents the average of the top two traces and should be identical to the trace immediately above it. Note that for Ps generated at a horizontal boundary (H) stacking over the larger bounds produces negligible error. However in this example (see text), the amplitude of the Ps phase generated at the dipping interface at 60 km depth is reduced by 38% for the larger stacking bounds; and only reduced by 6% for the smaller stacking bounds.  Chapter 3. RECEIVER FUNCTION ANALYSIS OF DIPPING STRUCTURE  51  Figure 3.6: Stacked radial receiver functions (data recorded at A L B - B ) having a back azimuth of 230° but slightly different A's (80° and 90°). The two phases labelled 'Reverb ?' exhibit rapid amplitude and arrival time variations and cannot be modelled as Ps conversions generated at a planar dipping interface. Stacking over narrow B A Z or A ranges allows the identification of such erratic phases which represent either reverberations or scattered energy.  Chapter 3. RECEIVER FUNCTION ANALYSIS OF DIPPING STRUCTURE 52  3.2.3  Lateral Sampling Range of P and Reverberations s  In many previous receiver function studies (e.g. Owens et al., 1987), lateral variations in earth structure were identified by developing different earth models for arrivals from various back azimuths (e.g. N W , SE, SW). For dipping layers, an earth model which satisfies the observed data from all back azimuths is sought. To ascertain the sampled region, the lateral range of both Ps phases and reverberations are examined. The reference model used in the following calculations is that given in Table 3.1 and arrivals from A = 80° (p = .046 s/km) are considered. Although the exact lateral extent of sampling of Ps will be more complex than that illustrated by this simple model, experiments with more complicated models indicate that for Ps, the ranges determined here are valid first-order estimates. A horizontal boundary (Table 3.1 but all dip angles set to 0°) is considered first and the lateral extent of sampling on the 60 km depth interface by Ps phases and reverberations are examined. Figure 3.7 illustrates a simplified version of the model given in Table 3.1. Although a 4-layer earth model is used in the following calculations, only the boundary at 60 km depth is shown. Figure 3.7a illustrates the circular region (centred on the recording site) sampled by Ps (solid) and reverberations (dotted). Ps phases originate -12 km from the station whereas reverberations are generated -50 km from the station. For dipping layers (Table 3.1), the lateral extent of sampling has shifted to the updip side of the station (Figure 3.7b). Due to refraction of waves at a dipping boundary, arrivals approaching from the dip direction contain Ps phases generated very close to the recording site (2 km downdip). However, Ps arrivals from the opposite direction were generated approximately 18 km updip of the recording station. Reverberations show a similar pattern, but sample over larger distances. Reverberations approaching from the dip direction were generated -30 km downdip of the station whereas those from the opposite back azimuth were generated -60 km updip of the recording station.  Chapter 3. RECEIVER FUNCTION ANALYSIS OF DIPPING STRUCTURE  53  Figure 3.7: Examples of the lateral extent of sampling provided by Ps conversions and reverberations (in both plan view - top; and cross-section - bottom) for a horizontal boundary (a), and a dipping boundary (b). The cross-section represents a simplified version of the 4-layer earth model used. For both cases the boundary is at 60 km depth and arrivals from A = 80° are considered. Arrivals from a smaller A will sample a larger area (see text). In the case of a horizontal boundary a circular area about the recording site (black square) is sampled; for a dipping boundary the area sampled is offset to the updip side of the station.  Chapter 3. RECEIVER FUNCTION ANALYSIS OF DIPPING STRUCTURE  54  Thus, in the case of a boundary dipping 20° at 60 km depth the Ps phases observed in receiver functions representing the updip and downdip directions were generated at points on the boundary approximately 20 km from one another (i.e. 4—5 seismic wavelengths). The planar boundary assumption is likely valid over this distance range. However, in the case of reverberations, which may be generated at points separated by 100 km or more (20-25 seismic wavelengths), it is likely that the assumption of a planar, dipping interface will begin to break down. As a result of this, it is doubtful that reverberations from deep dipping interfaces can successfully be modelled. This example considers a specific case of a boundary dipping 20° at 60 km depth and A = 80°. For a smaller A, the sampling range of both Ps and reverberations will be greater and for a given A the lateral sampling range of all phases will be smaller for shallower boundaries and larger for deeper boundaries. In more general terms, for a horizontal boundary and a A range of 45° — 80° the total lateral sampling range of Ps will be -0.4-0.7 times the depth of the generating boundary while reverberations will sample a total lateral distance of 2-3 times the depth of the interface. For a boundary dipping - 2 0 ° , Ps and reverberations will sample a maximum lateral range of 0.3-0.5 times and 1.5-3 times the depth of the generating interface respectively. When modelling, one should consider the depth and dip angle of the boundary, estimate the lateral sampling range of Ps and reverberations, and judge whether the planar boundary assumption is justifiable.  3.2.4  Stability of Reverberations  Shear-wave reverberations form an integral component of receiver functions. A n incident P-wave has the potential to generate numerous reverberatory phases (Figure 3.8a) which arrive at times approximately 3-4 times greater than that of Ps. The top trace in Figure 3.8b  Chapter 3. RECEIVER FUNCTION ANALYSIS OF DIPPING STRUCTURE 55  Converted Ray Phase Diagram a  Seismic Station V = 6.5 km/s V = 3.8 km/s p = 2.8 g/cm p  s  3  Moho  30 km  V = 8.0 km/s V = 4.6 km/s ,3 p = 3.3 g/cm p s  After Owens, 1984  Figure 3.8: (a) Simplified ray diagram illustrating the reverberations and conversions which may be generated by a P-wave incident at a boundary, (b) Corresponding receiver functions for the model shown assuming a horizontal boundary (top trace) and a boundary dipping 15° to the east (lower 3 traces). Each Ps phase typically generates 2-3 reverberations of comparable amplitude. In the case of a dipping layer, rapid amplitude and arrival time variations occur as a function of back azimuth (lower 3 traces).  Chapter 3. RECEIVER FUNCTION ANALYSIS OF DIPPING STRUCTURE 56  illustrates Ps and reverberations contained in a radial receiver function generated using the very simple horizontal layered earth model of Figure 3.8a. Note that 2 of the reverberations, PpPmS and PpSmS, have an amplitude comparable to that of Ps. For a horizontal boundary the amplitude and arrival time of Ps and all reverberations are independent of back azimuth. In addition, any P-wave reverberations (i.e. PpPmP) arrive at the surface at the same angle of incidence as the direct P-wave and are therefore removed from the receiver functions by source equalisation (Ammon, 1985). To illustrate the effect of a dipping layer on reverberations, consider the simple model shown in Figure 3.8a but with the Moho dipping 15° to the east. In this case P-wave reverberations arrive at the surface at a different angle of incidence than the direct P-wave and therefore these phases are not completely removed by source equalisation; the amount of P-energy removed by this process is extremely dependent upon back azimuth. Figure 3.8b (lower 3 traces) demonstrates the large variations in both the amplitude and arrival time of the reverberations which occur as a function of back azimuth. Note that the phase PpPmP which is observed updip (positive polarity), and downdip (negative polarity), is not present in the horizontal layered case. Also note that PpPmS and PpSmS exhibit rapid variations in amplitude, arrival time and polarity as a function of back azimuth. Significant variations in reverberation amplitude and arrival time also occur as a function of A. Note that Ps amplitudes always decrease as A decreases (Figure 3.4). However the amplitude of reverberations may remain constant or even increase as A decreases. This may be useful to discriminate between reverberations and Ps conversions. Although the variations in Ps amplitude and arrival time as a function of back azimuth and A provide constraints on the geometry of dipping interfaces, modelling the predicted variations for reverberations is extremely difficult. To illustrate the futility in attempting to model reverberations in a dipping layer environment consider 2 earth models; R (given in Table 3.1), and S - which is identical to R with the exception of an extremely minor  Chapter 3. RECEIVER FUNCTION ANALYSIS OF DIPPING STRUCTURE 57  BAZ--= 270°  Phase Ps PpPmP PpPmS PpSmP PpSmS  T(s) Amp T(s) Amp T(s) Amp T(s) Amp T(s) Amp  R 5.9 .044 13.7 -.023 19.1 .039  S 5.9 .044 21.7 -.019 24.1 .052  -  -  -  23.9 .038  29.2 .017  B A Z == 180° B A Z = 90° R S R S 6.5 6.5 7.1 7.1 .101 .101 .153 .153 13.5 23.4 15.8 15.8 -.003 -.037 .063 .052 19.9 25.9 21.7 21.7 .090 .084 .022 .017 18.7 33.0 21.9 21.9 -.008 -.009 .008 .003 25.5 31.6 27.8 27.8 -.048 -.049 -.070 -.067  Table 3.2: Arrival times and amplitudes of Ps and reverberations  (AV = 0.08 km/s) boundary at 10 km depth dipping 20° to the SE. Table 3.2 compares the S  arrival time and amplitudes (for 3 back azimuths) of Ps and reverberations associated with the eastward dipping boundary at 60 km depth (see Table 3.1) for the models R and S. Note that the Ps amplitudes and arrival times are not affected in any way by the addition of the minor A V boundary. However, significant amplitude variations and arrival time differences S  of up to 14 s occur for the reverberations (Table 3.2). Only those reverberations contained in receiver functions having B A Z = 90° (i.e. waves travelling updip) are unchanged. Given the drastic changes in the arrival times of reverberations by the addition of a dipping boundary with such a minor S-velocity contrast, it is clear that reverberations cannot be used to constrain earth structure. As reverberations are an important constituent of receiver functions, it is concluded that it would be unwise to apply inverse theory to receiver functions in the presence of dipping layers. Forward modelling of only those phases whose amplitude and arrival time variations as a function of both back azimuth and A are indicative of Ps is recommended. Reverberations and scattered energy are best identified by their erratic nature (as illustrated previously in Figure 3.6). These arrivals do pose a potential problem as they  Chapter 3. RECEIVER FUNCTION ANALYSIS OF DIPPING STRUCTURE  58  may arrive at the same time as a Ps phase thereby altering its appearance. Thus, in a complex tectonic setting, a large data set and stacking over narrow azimuthal and A ranges is required.  3.3 Resolution Capability of Receiver Functions 3.3.1 Minimum Detectable AVs While previous studies in which receiver functions were formally inverted (e.g. Owens et al., 1984) resulted in earth models sometimes containing very minor S-velocity contrasts between layers (e.g. -0.1 km/s), these boundaries produced extremely subtle features in the waveform. This detail cannot be obtained when forward modelling and a valid question is: what is the minimum A V at a boundary which will generate an observable Ps phase in a receiver S  function? This obviously depends upon the noise level of the receiver functions. In this study the average noise level associated with the stacked receiver functions is approximately +0.04 P . The highest quality receiver functions have noise levels of approximately ±0.02 P . z  z  In addition, the amplitude of the Ps phase is dependent upon the angle of incidence at a boundary. This is dependent on A (Ammon, 1991), and in the case of dipping layers, the back azimuth. Therefore, the data distribution will play a role in determining the minimum detectable A V . S  Horizontal Structure Table 3.3 provides Ps amplitudes for S-velocity contrasts of 0.10 to 0.50 km/s for epicentral distances of 45° and 80°. Bold-faced values denote those Ps amplitudes which are larger than the average noise level of 0.04 P in this study. The highest quality receiver functions in this 2  study could potentially resolve boundaries having a A V > 0.1 km/s (if A - 45°) or 0.2 km/s S  (if A - 80°). However, considering the typical noise level of -0.04 P , it is apparent from 2  Table 3.3 that the minimum AVs which may be resolved lies between 0.2 and 0.4 km/s. In  Chapter 3. RECEIVER FUNCTION ANALYSIS OF DIPPING STRUCTURE  Epicentral Distance 45° 80°  0.10 0.018 0.011  AVs\'km/s) 0.20 0.037 0.022  0.25  0.30  0.35  0.40  0.45  0.50  0.046  0.055  0.064  0.074  0.083  0.092  0.027  0.032  0.038  0.043  0.049  0.054  Table 3.3: Ps amplitude as a function of S-velocity contrast and A  this study, approximately one half of the data set consists of arrivals from A = 30° - 60° and one half consists arrivals from A = 70° — 100°. Therefore, it should be possible to resolve boundaries having an S-velocity contrast of > 0.3 km/s.  Dipping Structure  As demonstrated in Figure 3.4, the amplitude of Ps phases generated at a dipping boundary varies as a function of back azimuth. Ps amplitudes are largest for updip arrivals and smallest for downdip arrivals. Ps phases in arrivals from the strike direction are comparable to those generated at a horizontal boundary having the same A V . A data set lacking events from S  A < 60° and having a gap of > 180° in back azimuth could 'miss' very significant boundaries (i.e. AV ~0.5 km/s) if the data gap coincided with the dip direction of the boundary. S  The data distribution in this study (see section 2.4) consists of numerous receiver functions from A = 30° - 90° at 2 back azimuths, namely 130° and 300°. In addition, each station has good coverage for back azimuths of 130° - 5°, typically for A = 80° - 90°. The largest azimuthal gap is between back azimuths of 5° - 100° (ALB, LAS) and 5° - 130° (EGM). Thus, the azimuthal gaps in this study are small enough (< 125°) that dipping boundaries of AVs ^ 0.3 km/s may be resolved.  59  Chapter 3. RECEIVER FUNCTION ANALYSIS OF DIPPING STRUCTURE  3.3.2  60  Transition Zones vs 1 Order Velocity Discontinuities st  To examine the ability to distinguish I * order velocity discontinuities from transition zones, s  experiments were performed using synthetic data and various earth models.  Figure 3.9  illustrates Ps conversions contained in radial receiver functions dominated by 3 frequency pass-bands: a - 7; a = 5, and a = 3 (for f < 2, 1, and 0.5 Hz respectively). Each of the 3 series of traces compares a Ps conversion at a I * order velocity discontinuity (top trace) s  with those generated at a linear transition zone from 1 to 5 km thick (lower 5 traces). The S-velocity contrast is 0.58 km/s. The numbers beside each Ps represent the amplitude of the conversion relative to that at a I  st  order discontinuity. For an a value of 5.0 it is difficult  to distinguish a discontinuity from a transition zone up to 2 km thick. Similarly, for an a value of 3.0 it is difficult to distinguish a discontinuity from a transition zone up to ~3 km thick. M y data set, dominated by receiver functions of a = 3-5 therefore cannot distinguish between sharp boundaries and transition zones up to 2-3 km thick. Broad transition zones (thicker than - 3 km) may be identified by comparing receiver functions having different a values. For example a transition zone 5 km thick may be identified by a relative doubling of the Ps amplitude for a - 3 compared to a = 7 (Figure 3.9).  3:3.3  Resolution of Thin Layers  The ability to discriminate thin layers represents a dramatic improvement of broadband receiver function studies over previous long-period studies which could only provide details on gross crustal properties. Using synthetic data the 'thinness' of thin layers which may be resolved with this technique is examined. Figure 3.10 illustrates the ability of the three frequency pass-bands represented by a= 7, 5, and 3 to resolve a 'thin' layer (ranging from 1-5 km thick). Note the top trace was calculated using a 10 km thick layer - the amplitude of this Ps phase is the reference value against which the others are compared. The model  0  4  Time (s)  0  4  Time (s)  0  4  ^  Time (s)  ^  Figure 3.9: Resolution of 1-5 km thick transition zones provided by the 3 frequency pass-bands represented by a = 7, a = 5, and a = 3 (see text). The numbers beside each Ps phase represent the amplitude relative to a Ps phase generated at a I * order velocity discontinuity (top trace in each case). s  Figure 3.10: Resolution of a thin layer (1-5 km thickness) provided by the 3 frequency pass-bands represented by a = 7, 5, and 3. The numbers beside each Ps phase are the amplitude relative to a Ps phase generated at a 10 km thick layer (top trace).  as to  Chapter 3. RECEIVER FUNCTION ANALYSIS OF DIPPING STRUCTURE  63  has a thin layer of V = 3.47 km/s embedded in a region of V = 4.05 km/s. s  s  The highest frequency receiver functions (a = 7) successfully resolve layers as thin as 1-2 km, for a = 5 layers as thin as 2-3 km may be resolved and for the lower frequency events (a = 3) layers 4-5 km thick may be resolved without a significant loss of information (<10% Ps amplitude decrease). Attempting to resolve a thinner layer than is justified by the frequency content of the data would result in an underestimate for the A V of the lowS  velocity zone. It should be noted that the above synthetic receiver functions were generated (as were all synthetic data) using a Gaussian pulse convolved with a spike series. In the case of real data, the ability to resolve thin layers may be reduced if the averaging function is not a simple Gaussian or delta-like pulse. In this study the averaging functions were simple (section 3.3.5) and these experiments clearly indicate that my data may accurately resolve 3-5 km thick low-velocity zones.  3.3.4  Very Shallow Structure  Horizontal Layers  The effects of shallow structure on broadband teleseismic P-waveforms have been documented by Owens and Crosson (1988). They demonstrate that high-velocity contrast dipping boundaries near depths of 1.8-2.5 km produce either double-peaked receiver functions or an apparent delay in the arrival time of the radial component of the direct P-wave at some back azimuths. Double-peaked receiver functions are not observed in this data set, however there are some modifications to the direct P arrival which are attributed to very shallow structure (0.5-1.0 km depth). In Figure 3.11 dotted traces represent radial receiver functions generated using a simple 2- layer reference model (i.e. Figure 3.8a); the superimposed solid traces represent those receiver functions generated by the reference model with an added shallow (0.5-2.0 km),  Chapter 3. RECEIVER FUNCTION ANALYSIS OF DIPPING STRUCTURE  Horizontal Shallow Structure Effects  0  2  4  Time (s) Figure 3.11: Radial receiver functions generated using an earth model (described in text) having a 0.5-2.0 km thick horizontal, low-velocity surface layer are compared to those generated using a reference model (Figure 3.8a). The ray parameter used in these calculations (p = .046 s/km) corresponds to a A of 80°. In all cases the Ps phase near 3.5 s is unaffected by the shallow structure. The Ps phase generated at the near-surface interface cannot be resolved from the radial component of the direct P-wave; however reverberations are observed for a layer thickness > 1 km.  64  Chapter 3. RECEIVER FUNCTION ANALYSIS OF DIPPING STRUCTURE 65  horizontal, low-velocity layer ( V = 2.60 km/s). In all cases a - 5. s  In all cases, the Ps conversion from the shallow boundary arrives within 0.2 s of the direct P-arrival and therefore cannot be resolved. The result is a slight asymmetry in the direct P-arrival for all back azimuths. Receiver functions for the station L A S exhibit this pattern (section 4.6.3). For a shallow layer thickness less than ~1 km the reverberations associated with the near-surface interface arrive at times less than 0.4 s and thus cannot be resolved from the direct P-arrival at T = 0.0 s. For a layer thickness > 1 km, some reverberations can be resolved as separate phases (e.g. those arrivals between 1.0-1.5 s in Figure 3.11). The presence of such large reverberations is a noteworthy point. When modelling early arrivals in receiver functions ( < approximately 2 s) one must be certain that these are Ps phases, and not reverberations from a shallow boundary (<2 km depth).  Dipping Layers Very shallow, dipping layers may be recognised, as are other dipping layers, by azimuthal variations in both the radial and transverse receiver functions. The dotted traces in Figure 3.12 represent receiver functions generated using the simple 2-layer model of Figure 3.8a. These are compared with receiver functions (solid lines) generated using a model which is similar but has an east dipping (20° dip) 0.5 km thick surface layer with V = 2.60 km/s. For all s  synthetics a = 5. To illustrate the azimuthal variations in receiver functions, arrivals from four back azimuths are presented. The most dramatic effect of very shallow dipping structure is the presence of a large transverse component of motion early in the waveform (Figure 3.12). The transverse amplitude is largest for arrivals from the strike direction of the dipping boundary, and zero for updip or downdip arrivals. The polarity is antisymmetric about the dip direction. For the case of 0.5 km thick layer, Ps and reverberations cannot be resolved from the  Dipping Shallow Structure Effects Radial  Transverse Solid: Shallow Structure Dash: No Shallow Structure  BAZ 000  090  180  270  0  2  Time (s)  J  0  I  L  2  Time (s)  Figure 3.12: Radial and transverse receiver functions generated using an earth model (described in text) having a 0.5 km thick, low-velocity surface dipping 20° to the east are compared to those generated using a reference model (Figure 3.8a). The shallow dipping layer introduces a significant arrival early in the transverse receiver function (along the interface strike direction) and causes a subtle, azimuth-dependent variation in the apparent radial direct P arrival.  Chapter 3. RECEIVER FUNCTION ANALYSIS OF DIPPING STRUCTURE  67  direct P-arrival on the radial receiver functions, resulting in an asymmetry at T = 0.0 s. For arrivals travelling along strike (BAZ = 0° or 180°) or updip (i.e. approaching from B A Z =. 90°), more energy appears to arrive at T > 0.0 s in the direct P-wave. For waves travelling downdip (BAZ = 21QFirc), more energy appears to arrive at T < 0.0 s than T > 0.0 s in the direct arrival. This is due to a negative polarity reverberation arriving near 0.2 s. In section 4.5.4 evidence for a shallow, dipping interface at A L B - B is presented. Note that the Ps conversion from the Moho at 30 km depth is not affected by the presence of the shallow structure (Figure 3.12). For layers thicker than about 2 km, Ps will be resolvable and result in double-peaked receiver functions (see Owens and Crosson, 1988). These calculations demonstrate that very shallow (depths <2 km) structure may be constrained by modelling the subtle, but predictable modifications to the direct P-wave in receiver functions and that reverberations associated with very shallow structure may complicate the first few seconds of receiver functions.  3.3.5  Resolution and Modelling  The frequency dependent response of receiver functions illustrated above, and by other studies (e.g. Owens and Zandt, 1985; Ammon, 1985), does not pose a problem in modelling but rather provides an opportunity to identify and examine features such as thin layers and broad transition zones by considering different frequency bands of data.  The above examples  indicate the importance of matching the frequency content of synthetic receiver functions to that of the observed data. This is best achieved by matching the half-width of the direct P-wave at T = 0.0 s on the synthetic radial receiver functions (by varying the a value) to those of the observed radial receiver functions. As pointed out in the previous section, very shallow structure may introduce an asymmetry into this phase and it is therefore best to emphasise the match of the half of the pulse prior to T = 0.0 s.  Chapter 3. RECEIVER FUNCTION ANALYSIS OF DIPPING STRUCTURE 68  Layer No. 1 2  Vp km/s 6.50 7.50  Vs km/s 3.76 4.34  P g/cm 2.80 3.11  3  Th. km 30 oo  Dip Angle  Dip Dir.  00. 00.  000. 000.  Table 3.4: Reference model used for Ps amplitude tests  As pointed out in section 1.4 the averaging function associated with each deconvolution may be calculated using equation 1.7. Performing this calculation for numerous deconvolutions has demonstrated that the averaging function is simply the pulse at T = 0.0 s in the radial receiver function. Averaging functions calculated in this study are simple Gaussian pulses and have no significant side-lobes. However, if these were present (e.g. a large trough immediately prior to T = 0.0 s), the averaging function could be calculated and used to generate synthetic receiver functions which would be more appropriate for modelling the data. In this study, the averaging function was calculated for each deconvolution and normalised to unit amplitude in the time domain to preserve the absolute amplitude of the receiver function. However, it was not necessary to use the averaging function with synthetic data in the forward modelling stage.  3.3.6  Ps Sensitivity to AVs, A V p and p  To examine the sensitivity of Ps amplitudes to the P-velocity, S-velocity and density contrasts at a boundary, a simple reference model (Table 3.4) was chosen and the contrast of one parameter was increased by 20% while the remaining parameters were held fixed. It was determined that increasing the S-velocity contrast by 20% resulted in a comparable increase in the amplitude of Ps generated at the boundary. However, increasing either the P-velocity or density contrasts by 20% (for a fixed AVs) resulted in a negligible (< 1%) change in the Ps amplitude. This indicates that boundaries dominated by density or P-velocity contrasts  Chapter 3. RECEIVER FUNCTION ANALYSIS OF DIPPING STRUCTURE  69  will not be detected using this technique.  3.3.7 Broadband vs Short-period Receiver Functions The velocity response of the broadband Guralp seismometers, flat from approximately 0.05 to 10 Hz, encompasses the response of the typical Willmore MK-II short-period systems (flat from 1 to 10 Hz). Thus, by filtering the data a short-period waveform can be simulated and used to determine 'short-period' receiver functions. Event No. 11 recorded at PGCB is chosen as an example based on its high SNR and broad amplitude spectrum. Figure 3.13 compares the vertical broadband recording (Figure 3.13a) with a simulated short-period recording (Figure 3.13b) obtained by filtering the broadband data. Figures 3.13c and 3.13d illustrate the corresponding amplitude spectra. Note that the broadband recording is dominated by frequencies of 0.08 to 0.8 Hz, whereas the simulated short-period recording is dominated by frequencies of approximately 0.4 to 2.0 Hz. The two radial receiver functions (Figure 3.13e) generated using the broadband and simulated short-period data (in both cases the deconvolution parameters are a = 5 and c = 0.00001) are nearly identical. Repeating this process for numerous events recorded at PGC-B indicates that in general, short-period waveforms yield very similar receiver functions, albeit with a slightly higher noise level, to those obtained using broadband data. This indicates that at this location a short-period system would provide useful data for a receiver function study. This short-period and broadband comparison demonstrates that in some cases at least, short-period data may be useful for receiver function analysis. However, it is important to note (as demonstrated in section 3.3.2) that in contrast to broadband data, short-period data cannot resolve large scale earth structure (e.g. a 5-10 km thick transition zone). Thus, broadband data have the distinct advantage of being able to provide information on both gross and fine-scale (1-2 km) earth structure. In addition, as demonstrated by Ellis and Basham  Broadband Vertical  Broadband Amplitude Spectrum  free)  (Hi)  Figure 3.13: Comparison of the broadband vertical component for Event No. 11 recorded at PGC-B (a); and the simulated short-period waveform (b) obtained by filtering (a) (note the different amplitude scales). The corresponding spectra (c and d) illustrate the different frequency content of these waveforms, yet the corresponding receiver functions (e) are nearly identical.  Chapter 3. RECEIVER FUNCTION ANALYSIS OF DIPPING STRUCTURE 71  (1968), scattering of short-period waves in a sedimentary environment may restrict the use of such data. However, longer period broadband data collected at the same site might may be useful. 3.4  Summary  In this chapter several important aspects of receiver function analysis as applied to a dipping layer environment have been examined and documented. These are summarised as follows: 1. It is important to model the absolute rather than the relative amplitudes of receiver functions. This allows extraction of information on the near-surface velocity structure and on dipping layers. More importantly, potential inaccuracies in the earth model caused by undetected boundaries altering the relative amplitude ratios (Ps/P) are avoided. 2. In a dipping layer environment, only events clustered within 10° in BAZ and 10° in A should be stacked. This is especially important when research objectives include deep, dipping structure. Further, stacking over a narrow range of BAZ and A facilitates the identification of reverberations and scattered energy contained in receiver functions. 3. Ps phases and reverberations sample a total lateral distance about the recording site of approximately 0.4-0.7 times, and 2-3 times, respectively, the depth of the interface. Thus, reverberations are more likely to encounter lateral velocity variations or nonplanar structure and invalidate the modelling assumptions. In addition, it was demonstrated that the presence of an undetectable (e.g. AV = 0.08 km/s) dipping boundary S  can cause significant changes in the amplitude and arrival time of reverberations generated at deeper interfaces. Thus in a dipping layer environment, reverberations cannot be readily used to provide constraints on earth structure. As reverberations are an important constituent of receiver functions, inversion of receiver functions is not justified  Chapter 3. RECEIVER FUNCTION ANALYSIS OF DIPPING STRUCTURE  72  in this environment. It is concluded therefore that only those phases exhibiting the amplitude and arrival time variations indicative of Ps be modelled. 4. The data distribution and typical signal-to-noise ratio associated with this data set allows for the resolution of boundaries having an S-velocity contrast > 0.3 km/s. It is also shown that layers of thickness 3-5 km may be resolved without difficulty and that very shallow (0.5-2.0 km) structure may be identified by systematic modifications to the direct P-wave in receiver functions. 5. By filtering broadband recordings to simulate short-period waveforms, it was demonstrated that receiver functions obtained from short-period data for PGC-B are very similar to those obtained from broadband data. However, broadband data still provide the flexibility to explore both thefine-scaleand large-scale features of the lithosphere. At those sites where scattering of short-period waves is severe, the longer period data contained in broadband recordings could still provide information on the sub-surface structure.  Chapter 4 MODELLING RECEIVER FUNCTIONS  4.1 Introduction This chapter describes the forward modelling procedure and outlines a simple technique which was derived to facilitate the direct comparison of reflection data and teleseismic receiver functions. The S-velocity models derived from the interpretation of receiver functions at each array site, A L B - B , L A S and E G M , are then presented and discussed.  4.2 Forward Modelling Throughout this study, receiver functions were interpreted using trial-and-error forward modelling. Although many recent broadband studies (e.g. Owens et al., 1987) have formally inverted receiver functions to determine fine-scale earth structure (1-2 km thick layers), it was demonstrated in Chapter 3 that in the presence of dipping layers it may be very difficult to model the amplitude and arrival time of reverberations and therefore these phases should not be quantitatively used to determine earth structure. As such, only those phases whose amplitude and arrival time variations with back azimuth and A are consistent with that expected of a Ps phase are modelled. Synthetic seismograms were generated using the fast ray-tracing scheme outlined in section 1.5. For the reasons outlined above, to model the observed data only the direct P and Ps arrivals were included in the synthetic seismograms. However, synthetics containing multiples were examined to better understand the potential effects of these phases. It was  73  Chapter 4. MODELLING RECEIVER FUNCTIONS  74  determined that although reverberations may occasionally interact with a Ps phase of interest thereby distorting it's appearance, this is a 'spurious' effect in that it does not occur over a wide range of B A Z or A. The starting model for each station was based on the refraction P-velocity model of Drew and Clowes (1990). Other models were considered (Spence et al., 1985; White and Savage, 1965; and Langston, 1981). However, this model is the most recent interpretation and incorporates both refraction and reflection data acquired in the last decade. Densities and S-velocities for each layer were based on the P-velocity. The P-velocity-density relationship used (Zelt, 1989) was a fourth order polynomial fit to the data of Ludwig et al. (1970). Unless otherwise noted, V /V = 1.73 (o = 0.25), based on local earthquake data (Rogers, p  s  1983). Synthetic radial and transverse receiver functions were generated by treating the synthetic seismograms in the same way as the real data. In the deconvolution process an a value (Gaussian pulse width) was chosen which provided an optimum match to that of the observed stacked receiver function as described in section 3.3.5. In modelling the data, emphasis was placed on fitting high quality stacked radial receiver functions. The S-velocity contrast, dip direction and dip angle of major boundaries were varied until a satisfactory fit to those phases identified as Ps was obtained. Dipping boundaries were identified by an azimuthal variation in the amplitude and arrival time of the radial Ps phase. The amplitude and polarity of the transverse component of motion were also considered. However, as in previous broadband studies, these data were often not useful. As S-velocities were varied to provide a fit to the amplitude of the Ps phases, P-velocities were changed accordingly. In order to incorporate refraction results into my models, the crustal average V in the interpretation of Drew and Clowes .(1,990) was maintained. Trial-and-error p  forward modelling demonstrates that AVs variations of + 0.1-0.2 km/s result in Ps amplitude variations which lie within the bounds of the overall noise level of the deconvolutions and  Chapter 4. MODELLING RECEIVER FUNCTIONS  75  the error bars associated with the stacked receiver functions.  4.3  Comparison of Reflection Data and Receiver Functions  To date, the complementary nature of seismic reflection and receiver function data sets has not been fully exploited. P-wave reflection data sets are sensitive to the P-velocity structure; whereas receiver functions are sensitive primarily to the S-velocity structure. In theory, by considering both data sets, V / V could be determined layer-by-layer throughout the earth p  s  model. This would require coincident data sets of very high quality, and one would have to be confident that each Ps conversion observed in a receiver function corresponds to an arrival observed in the reflection section (possibly through a comparison of P-wave reflection, and P-to-S conversion coefficients). This is difficult due to the quite different frequency bands. Reflection data are generally dominated by frequencies of 10-40 Hz, whereas teleseismic receiver functions are dominated by frequencies of 0.1-1.0 Hz. Therefore in those cases where reflections result from the constructive interference of P-waves interacting with finescale structure (e.g. the -100 m lamellae which have been suggested (Hyndman, 1988) as the origin for some deep crustal reflectors), receiver functions will likely only image the upper and lower bounds of the reflective zone. To compare these two data sets, note that teleseismic receiver functions provide the oneway near-vertical travel-time difference of S- and P-waves (T -T ) s  p  from boundaries to the  surface. Reflection sections provide the 2-way vertical travel-time of P-waves (2T ) from p  the surface to a reflecting interface. Thus, the appropriate T -T s  (Ts-T ) 2T ~2;„ y\.  W IV )-1 nr  P  relationship is:  (4.i)  p  s  p  p  S  To apply this technique an estimate of the average V /V from the surface to the boundary p  s  of interest is required. Using synthetic data and a 2-layer earth model this equation yields  Chapter 4. MODELLING RECEIVER FUNCTIONS  76  2T estimates within 3% of their true value for arrivals from A > 80°, which have an average p  crustal travel-path within ~ 14° of the vertical; and for Ps phases generated at horizontal, or gendy dipping structure, in which case T - T is relatively insensitive to back azimuth. For s  p  Ps phases generated by steeply dipping structure, significant T — T variations as a function s  p  of back azimuth may exist thereby resulting in a larger scatter in the 2T  p  estimates. For  arrivals from A < 80° the more non-vertical travel-path will result in larger errors and thus this equation is not recommended. This technique has a significant advantage over modelling receiver functions and then comparing the final earth model with a reflection depth section; knowledge of V  p  is not  required. In the following sections this technique is applied to receiver functions recorded at A L B and E G M .  4.4  Outline  Prior to describing the interpretation of the A L B - B data set, the organisation of this, and the following interpretation sections are first outlined. For each site, specific research goals are summarised. This is followed by a presentation and description of the data set. If reflection data are available, the technique described in section 4.3 is used to compare these data with the receiver functions. Next, synthetic receiver functions are generated using the refraction interpretation of Drew and Clowes (1990) and compared to the observed data. I then outline the analysis of the receiver functions and describe the S-velocity discontinuities interpreted in the crust and upper mantle beneath the recording site. Examples of the fit provided by the final model to a subset of the data are presented at this stage. Next, the final model and the fit provided to the entire data set are presented and discussed. Finally, uncertainties in the parameters are estimated and a comparison is made between my model and previous earth models.  Chapter 4. MODELLING RECEIVER FUNCTIONS  4.5  77  ALB-B Interpretation  The station A L B - B lies at a critical location in the earth models derived from reflection and refraction data. It is at the point in LITHOPROBE reflection line 84-01 where both the ' C (at 19-23 km depth) and the ' E ' (34-38 km depth) reflective zones fade. This does not necessarily indicate that the zones terminate at this location however. The absence of all deep reflections in this region may be caused by shallow geological or topographical complexities. It is noted that the reflection line crosses a major N W - S E trending anticline (represented by the Beaufort Range) near this point. Further, A L B - B is at the suture of the 'continental' and 'accretionary' type crustal structures (see Figure 1.4) in the refraction interpretations of Drew and Clowes (1990) and Spence et al. (1985). Few earthquakes occur at depth in this region (see Figure 1.3b). Therefore the position of the subducting Juan de Fuca plate is poorly constrained. Thus, using receiver function data the following questions are addressed: 1. Is there evidence for a continental Moho beneath A L B - B ? 2. Do the ' C and ' E ' reflective zones continue beneath this site? 3. What is the nature and geometry of the Juan de Fuca plate at this location? A potential problem at this site is the proximity of numerous faults (see Fig. 2.2 and the discussion in section 2.2). In a receiver function study at Berkeley, California, Ammon (1985) observed very large amplitude arrivals in transverse receiver functions. He could not explain these in terms of planar, dipping structure and suggested that they may result from waves interacting with the nearby San Andreas and Hayward faults.  4.5.1  The Data  Of the three recording sites, the largest and most complete data set was collected at A L B - B . A total of 94 teleseisms, providing 22 stacked and 5 single event receiver functions were  Chapter 4. MODELLING RECEIVER FUNCTIONS  78  used. Throughout this thesis the notation B A Z - A (e.g. 100-60) is used to represent the back azimuth and A of receiver functions. With the exception of the stacked data 230-80 and 280-100, all radial and transverse receiver functions are given in Figures 4.1—4.4. Two A-profiles  (i.e. stacked receiver functions over a A range of 30° - 90°) at B A Z ' s of 130°  (Fig. 4.1) and 300° (Fig. 4.2) provide valuable constraints on interface dip directions as they are separated by almost 180°. In addition, there is a good azimuthal distribution of data at A = 90° (Fig. 4.3). The largest azimuthal data gaps are from 5° - 80° and 170° - 230° (see polar plots of Figures 4.3 and 4.4). Figure 4.4 illustrates the data distribution at A < 65°; note that several are single event receiver functions, and with the exception of the data at B A Z = 300°, all have a low SNR. It is unfortunate that this data set contains so few arrivals from A < 60°, since these waveforms contain the largest amplitude Ps phases. In addition, arrivals from the N E quadrant, where the largest amplitude Ps phases generated at N E dipping boundaries would be observed, are lacking. Nonetheless, the data set is of high quality and adequate to allow the identification of any significant (AV > 0.3 km/s) S  planar interfaces beneath this site. The large transverse arrivals and the significant azimuthal variations observed in both the radial and transverse receiver functions indicate the presence of dipping layers beneath this site. In general, the largest arrivals are contained in the first 7 s of the receiver functions (Fig. 4.3). Only those phases which exhibit the amplitude and arrival time variations expected of Ps phases, are above the noise level, and are observed over a wide range of B A Z and A are modelled. There are several large amplitude phases observed in radial receiver functions from the southwest quadrant (denoted by unlabelled arrowheads at B A Z = 170° - 2 6 0 ° in Figure 4.3) which are not modelled. The phase near 2.5 s has a very large amplitude, but is observed only over the B A Z range 230° - 260°. Although a very steep dipping boundary (8 > 40°) having a small A V could generate such a phase, we would expect to observe other effects, such as S  large transverse arrivals near 2-2.5 s at B A Z = 130° and 300°, and a significant variation  A L B - B RECEIVER FUNCTIONS : B A Z = 1 3 0 °  Radial II  I I  I I I I  Number  J  in Stack  j  2  j  3  J  I  I I I I  Transverse  I  I I I I  I  6  I I  -5  I  f\fV A A / W / V V  40°  \J  II 0  I | I I i I  Y  50° 60°  A A/ I  | I li  S  B  V/  4  i i i i | i i i i | i i M  I I I I  I I i  75° 85°  I 5  i i i i  I 10  Time (s)  i i i i  I 15  i i i i  i i i i  20  -5  I 0  i i i i  I 5  i i i i  I 10  i i i i  I  i i i i  15  Time (s)  Figure 4.1: A L B - B receiver functions at a B A Z of 130°. Polar plot (far right) illustrates the B A Z and A for each trace (see triangles). A l l traces are plotted at the same scale. Phases discussed in the text are denoted by dotted lines.  20  A L B - B RECEIVER FUNCTIONS : B A Z = 3 0 0 °  Radial  i  -5  i  i  i  I 0  i  i  i  i  I  1i 5 10 Time (s) i  i  i  i  Transverse  i  i' i  I 15  i  i  i  i  I  20  1i -5  i  i  i  I 0  i  i  i  i  I  i  i  i  i  I  5 10 Time (s)  i  i  i  i  1i 15  i  Figure 4.2: A L B - B receiver functions at a B A Z of 300°. Polar plot (far right) illustrates the B A Z and A for each trace (see triangles). A l l traces are plotted at the same scale. Phases discussed in the text are denoted by dotted lines.  i i  20  A L B - B RECEIVER FUNCTIONS : A = 9 0 °  Radial  -5  Transverse  i i i 1i i i i I i i i i I i i i i 1 i i i i I i  0  5 10 Time (s)  15  20  -5  1  iiii1 iiii1 iiiiI iiii1 iii i  0  5 10 Time (s)  15  Figure 4.3: A L B - B receiver functions at a A of 90°. Polar plot (far right) illustrates the B A Z and A for each trace (see triangles). A l l traces are plotted at the same scale. Phases discussed in the text are denoted by dotted lines.  20  A L B - B RECEIVER FUNCTIONS : A < 65°  Radial  -5  0  5 10 Time (s)  Transverse  15  20  -5  0  5 10 Time (s)  15  Figure 4.4: A L B - B receiver functions at a A of < 65°. Polar plot (far right) illustrates the B A Z and A for each trace (see triangles). A l l traces are plotted at the same scale. Phases discussed in the text are denoted by dotted lines.  20  Chapter 4. MODELLING RECEIVER FUNCTIONS  83  in the amplitude of the direct radial P-wave. Such patterns are not observed. The second arrival near 4 s is observed over the B A Z range 170°-245°. The rapid variation in amplitude and arrival time as a function of A suggests that this arrival represents a reverberation or scattered energy.  Radial Data The radial phases modelled in this analysis are described below: 1. A subtie, azimuthal variation in the direct P-wave. This is not apparent in Figures 4.14.4, but is illustrated and discussed in section 4.5.4. This effect is attributed to a shallow, southwestward dipping boundary. 2. The phase (denoted 'S') which is observed at 1.1-1.2 s on all data from B A Z = 130° and 300° (Figs. 4.1, 4.2), and which has the largest amplitude in arrivals from the north (Fig. 4.3) and east (Fig. 4.4) but is not present in the radial receiver functions from the west and southwest (BAZ = 230° - 280°). 3. The pair of phases ('Cj-' and ' C s ' , having negative and positive polarity, respectively) observed near 2.3 and 3.3 s in data from B A Z = 130° and 300° (Figs. 4.1, 4.2) and in other arrivals from the north and northwest. 4. The prominent pair of phases ('Ey' and ' E 5 ' , having negative and positive polarity, respectively) observed near 3.9 and 4.7 s in data from B A Z = 300° (Fig. 4.2). These phases are observed in most receiver functions, but are largest in arrivals from the north (300° < B A Z < 5°, see Fig. 4.3). 5. The pair of phases ('F' and ' O M ' , having negative and positive polarity, respectively) observed near 5.4 and 6.0 s in receiver functions from the east and northwest. The amplitudes of these phases increase as the B A Z approaches north (Fig. 4.3).  Chapter 4. MODELLING RECEIVER FUNCTIONS  84  Transverse Data Although large transverse arrivals are present in most receiver functions, there are several points to consider. With the exception of phase labelled ' S ' (Figures 4.1^4.3), large amplitude transverse phases generally do not correlate with the Ps phases identified on the radial receiver functions. Transverse motion associated with Ps should be antisymmetric about the interface dip direction and exhibit a systematic amplitude variation (see Fig. 3.1).  The observed  transverse arrivals do not exhibit this pattern (compare Figs. 4.1 and 4.2). Further, as in previous receiver function studies (Ammon, 1985; Lapp etal., 1990), the amplitudes of these phases are much larger than can be expected from gently dipping boundaries (8 < 15°). Thus, no attempt is made to model these large transverse arrivals, but it is recognised that they are indicative of complex structure. Later in this chapter the fit provided to the transverse data by the final model is presented and discussed.  4.5.2  Comparison With L I T H O P R O B E Reflection Data  A L B - B is 12 km N W of LITHOPROBE reflection line 84-01 (Figure 4.5) near the point where almost all reflections fade. Using the technique outlined in section 4.3, the correspondence between the receiver functions and the reflection data is examined. Only those arrivals identified as Ps phases (and from A > 80°) are considered in this analysis. Using equation 4.1, 2T times are calculated for the following phases: ' S ' (T -T P  s  ' C ' (T -T B  s  p  = 1.15-1.40 s); ' C y ' and  p  = 1.9-2.3 s and 2.9 - 3.3 s, respectively); 'E ' and ' £ ' (T -T T  and 4.6 - 4.8 s, respectively); ' F ' (T -T s  p  B  s  = 5.3 - 5.5 s) and ' O M ' (T -T s  p  p  = 3.8-4.1 s  = 5.9 - 6.0 s).  Note that the time variation for each phase is largely due to T - T being a function of B A Z s  p  for Ps generated at a dipping boundary. Overall there is a remarkable correspondence between the most prominent Ps phases  CD SW  LINE DRAWING  A  NE  Figure 4.5: Line drawing summary of LITHOPROBE line 84-01 (after Clowes etal, 1987a). A L B - B is located 12 km to the N W of this line (see insert). Diamond symbols and vertical bars represent 2T values estimated from Ps-P times (using equation 4.1) and the scatter associated with these estimates respectively, for the major phases discussed in the text. Contours on insert (after Green et al, 1986) indicate the two-way travel-time (in seconds) to the top of the ' E ' reflective zone. The curvature of the dashed contours is speculative. P  Chapter 4. MODELLING RECEIVER FUNCTIONS  86  observed in the receiver functions and the major reflective zones imaged to the SE of A L B IS (Figure 4.5). The phase ' S ' may correlate with the top of the prominent band of N E dipping reflectors (near 2.5 s) about 10-15 km to the SW of the point marked ' A L B - B ' . The phases 'Cf' and 'Cjg' correlate with the top and bottom of a diffuse band of reflectors which appear to be an extension of the ' C reflective zone. The phases 'E^ and ' E B ' are slightly below the projection of the top and bottom respectively, of the reflective ' E ' zone. However, there is some evidence for an increase in the dip angle of the ' E ' reflectors near the point where continuity is lost (Yorath etal., 1985; Green etal., 1986). The 2T thickness estimated p  from the receiver functions is in good agreement with that observed on the reflection section. In addition, the large amplitude of 'E^' and 'Eg' require a significant A V . Calvert and S  Clowes (1990) estimated reflection coefficients of 11-22% for the ' E ' reflectors, this is also indicative of large V  p  or density contrasts. Finally, the receiver function phases ' F ' and  ' O M ' are imaging deeper than the reflection study at this location. However, ' F ' correlates with the projection of the top of the oceanic plate (reflector 'F') as imaged on the western end of this line. Based on offshore reflection data which have imaged the oceanic plate, the 2T difference between the top of the plate and the oceanic Moho is ~2 s (Hyndman et al., P  1990). Therefore, the phase ' O M ' is consistent with a Ps from the oceanic Moho. A n interesting conclusion from this analysis is that the most prominent phases in the data ('ET and 'EB') are clearly not associated with the subducting Juan de Fuca plate, but rather with the reflective zone located approximately 10 km above the top of the plate. Note that the contours representing the travel-time to the top of the ' E ' reflective zone (see inset - Figure 4.5) suggest a N N E dip direction; a similar direction is estimated from forward modelling the receiver functions. Finally, the polarity of the phases (negative for ' C y ' , 'ET and ' F ' ; and positive for 'CV, ' E B ' and ' O M ' ) indicate that the ' C and ' E ' reflective zones, as well as the oceanic crust are regions of low S-velocity.  Chapter 4. MODELLING RECEIVER FUNCTIONS  4.5.3  87  Initial Considerations  Prior to forward modelling the data, synthetics generated using the P-velocity model of Drew and Clowes (1990) are compared to the data.  Unfortunately, A L B - B lies on the  poorly defined boundary between what may be called 'accretionary' and 'continental' type earth structures (Figure 1.4b). This change in structure is required to explain a traveltime delay observed in refraction P-arrivals at mainland stations relative to Vancouver Island stations. The 'accretionary' structure contains a wedge of high-velocity material ( V = 7 . 1 p  7.2 km/s) extending from -17 km depth to the top of the Juan de Fuca plate. Directly above, and embedded in this high velocity region are two bands of low-velocity material (V  p  = 6.35 km/s) representing the ' C and ' E ' reflective zones. The major features of  the 'continental' structure are the broad mid to lower crustal low-velocity zone and a nearly horizontal Moho at 37 km depth beneath eastern Vancouver Island (Fig. 1.4). In Figure 4.6 several observed radial receiver functions (those data having the highest SNR, and covering as wide a range of B A Z and A as possible were chosen) are compared with synthetics generated using the 'accretionary' structure and the 'continental' structure. Although neither of these models provides a good fit to the observed data, the synthetic receiver functions generated using the 'accretionary' structure contain negative and positive polarity phases generated at the top and bottom of the ' E ' reflective zone which appear to be present, albeit with a slight time shift, in the two highest quality receiver functions in the data set, 300-50 and 320-95. The constant amplitude arrival ' C M ' corresponding to the continental Moho in the 'continental' type structure is not readily observed in the data, suggesting that this boundary is not present beneath A L B - B . This comparison, as well as the direct comparison between reflection data and receiver functions (section 4.5.2), suggests that the ' C and ' E ' reflectors do continue beneath A L B - B , and that the 'accretionary' structure employed in the Drew and Clowes (1990) interpretation  Chapter 4. MODELLING RECEIVER FUNCTIONS  88  ALB-B RADIAL RECEIVER FUNCTIONS 'Accretionary Structure'  'Continental Structure' Solid: Observed Dash: Synthetic  BAZ-A 100-65  130-85  230-80  300-50  320-95  5-90  Figure 4.6: A comparison of observed radial receiver functions and synthetics generated using the 'accretionary' and 'continental' type (see text) earth structure of Drew and Clowes (1990). Ps phases associated with major velocity discontinuities in these models (see text) are labelled.  Chapter 4. MODELLING RECEIVER FUNCTIONS  89  is most appropriate as a starting model. However, it should be noted that the choice of a starting model is not critical to the final model derived.  4.5.4  Crustal S-Velocity Structure  As demonstrated in all earth models for this region based on refraction interpretations (White and Savage, 1965; McMechan and Spence, 1983; Spence et al. 1985; Drew and Clowes, 1990), the crust beneath central and western Vancouver Island is not a typical continental type structure. There is nO readily visible continental Moho and relatively high velocity material (V  p  -  7 . 0 - 7 . 2 km/s) is required at shallow depths (-17 km). In the following  discussion 'crustal' refers to the region above the subducting Juan de Fuca plate, the top of which is estimated at a depth of 47 km. In the refraction interpretation of Drew and Clowes (1990), the upper crust consists of a smoothly varying V from the surface to a depth of -17 km. The near-surface V of p  p  -5.4 km/s increases rapidly to 6.4 km/s and then to 6.75 km/s at depths of 2 km and 17 km, respectively (Fig. 1.4b). In contrast, the receiver functions indicate significant S-velocity discontinuities in the upper crust at depths of -0.5 and 11 km. A very shallow, dipping interface is required by the subde effects (see section 3.3.4) on the direct arrival (T = 0.0 s) observed in the receiver functions. Figure 4.7a illustrates the azimuthal variation in the radial component of the direct P-wave. Note that starting at B A Z = 170° the direct P-arrival (T = 0.0 s) appears slightly asymmetric and has a maximum amplitude at T = 0.05 s. This effect increases towards the SW and is most pronounced over the B A Z range 230°-280°, where the direct arrival is clearly asymmetric about T- 0.0 s and has a maximum amplitude at T= 0.1 s. The effect decreases towards the N and disappears at B A Z -340°. Note also that at B A Z = 80° the direct P-arrival appears to be shifted slightly towards -0.05 s. This azimuthal variation can be explained by a Ps conversion generated  ALB-B 'Direct Radial P-wave'  Time (s) Figure 4.7: Shallow structure effects at A L B - B . (a) The azimuthal variation in the direct radial P-wave. Arrivals from the SW are asymmetric about T = 0.0 s and have a maximum amplitude (denoted by arrowheads) at T = 0.1 s. This is attributed to the presence of a large amplitude Ps phase generated at a shallow (0.5 km depth), SW dipping interface, (b) The fit provided to a sample receiver function (error bars indicate ±1 standard deviation on the mean of the stacked data) by a model with, and without the shallow dipping interface.  91  Chapter 4. MODELLING RECEIVER FUNCTIONS  at a shallow, high-velocity contrast boundary dipping towards the southwest. The apparent asymmetry in direct P-arrivals from the SW results from the arrival of a strong Ps conversion at T < 0.1 s. This effect becomes significant at a B A Z between 140° -170° and insignificant at a B A Z between 320° - 340°, thus indicating a dip direction between S50°W and S70°W. To constrain the dip angle and A V associated with this boundary it is noted that the effect is S  observed over a B A Z range of 170°; this suggests a significant A V (> 1 km/s) and a shallow S  dip angle. In addition, the lack of significant azimuthal variations in the amplitude of the direct P-wave (at exactly T = 0.0 s) constrains 5 to be < 20°. The best fit to the data over the entire range of B A Z and A is provided by a boundary at 0.5 km depth which dips 15° in the direction S65°W and has a AV of 1.2 km/s (or AV = 2.1 km/s). A surface layer P-velocity S  p  of -4.0 km/s yields V = 6.1 km/s at 0.5 km depth, which is reasonable given the estimated p  V  of 6.4 km/s near 2.0 km depth (Drew and Clowes, 1990). It is noted that similar low  V  values and large V contrasts have been observed at depths of up to 0.3 km at hard-rock  p  p  p  sites (see Lueschen et al., 1987; Ebel, 1989) and attributed to the weathering layer. The southwestward dip of this boundary is reasonable given that A L B - B lies on the southwest flank of a N W trending anticline represented by the Beaufort Range (see Figure 6.12 of Sutherland Brown, 1966). Figure 4.7b compares a typical observed receiver function from the southwest quadrant to synthetics generated by the final model with, and without the shallow, southwest dipping interface. Note that a shallow dipping boundary is required to reproduce the asymmetry about T = 0.0 s and the maximum amplitude at T = 0.1 s of the direct P-wave. The presence of a second significant S-velocity discontinuity in the upper crust beneath A L B - B is indicated by the large amplitude arrival ' S ' observed near 1.1-1.2 s in both the radial and transverse data (Fig. 4.3). As described earlier, phase ' S ' clearly increases in amplitude as the B A Z approaches N (see the lower 4 traces of Figure 4.3), and also has a large amplitude in arrivals from the east (BAZ = 80, 100, 130 - of Figure 4.4). In  Chapter 4. MODELLING RECEIVER FUNCTIONS  92  addition, large transverse arrivals are observed at B A Z = 300° (positive polarity) and B A Z = 130° (negative polarity). The polarity of the transverse arrival reverses between B A Z = 170°-230°, constraining the dip direction to be between N10°W and N50°E. Both the radial (Figure 4.8) and transverse data are best modelled by an interface at 11 km depth having a A V g = 0.6 km/s and dipping 25° in the direction N50°E. The relatively steep dip is required by the rapid azimuthal variation in the radial amplitude (see the lower 4 traces in Figure 4.3) and the large transverse amplitudes associated with this phase. To accommodate both a large velocity discontinuity ( A V  =0.6 km/s, or AV = 1 km/s) in the upper crust and maintain  S  P  the upper crustal average V of 6.4-6.7 km/s (Drew and Clowes, 1990), V has been set at p  p  6.1 km/s above the discontinuity, and 7.1 km/s below the boundary. In the mid to lower crust, two low-velocity zones, each about 5 km thick, are imaged. As described in section 4.5.2, these layers correlate with the ' C and ' E ' reflective zones imaged on LITHOPROBE reflection lines (Figure 4.5). The phases ' C y ' and ' C s ' are only observed in receiver functions over the B A Z range 300°-130°. Like phase ' S ' , this indicates a N E dipping zone. In addition, the polarity of these phases, negative for 'CV and positive for ' C B ' , is indicative of a low-velocity zone. These phases have a larger amplitude on SE receiver functions ( B A Z = 130°, see Figure 4.1) than on N W receiver functions (BAZ = 300°, see Figure 4.2). Thus the boundaries must have a dip direction east of N35°E (the bisector of these two data sets). A low-velocity zone dipping 15°+5° in the direction N50°E +20° provides the best fit to the data (Figure 4.8). The S-velocity contrast at the top of the boundary at 20 km is estimated at -0.46 km/s. To reproduce the broad positive peak 'CT observed on many receiver functions (see Figure 4.2), the lower boundary is modelled as two velocity jumps of +0.3 km/s and +0.35 km/s at depths of 23.1 and 26.6 km, respectively. The phases 'ET' and ' E B ' are the most prominent arrivals in the data set. With the exception of the B A Z range 170°-245°, these phases are observed in all receiver functions. They dominate receiver functions from B A Z = 300°, and exhibit an increasing amplitude  Chapter 4. MODELLING RECETVER FUNCTIONS  93  A L B - B RADIAL RECEIVER FUNCTIONS Solid: Observed Dash: Synthetic  i i  I  i  I  '  i 1i  0 Time (s)  5  I  I  Ii l l .  I I  0  j i i  I I  5  Time (s)  Figure 4.8: Synthetic radial receiver functions generated using the final model are compared with selected (see text) receiver functions. Phases discussed in text are indicated. Polar plots illustrate the data distribution  Chapter 4. MODELLING RECEIVER FUNCTIONS  94  for B A Z approaching N (lower 4 traces in Figure 4.3). The amplitudes of these phases are larger at B A Z = 300° compared to B A Z = 130°, indicating a zone dipping to the north of N35°E (the bisector of these two data sets). The positive and negative polarity of 'Ex' and 'Eg' respectively, indicates a low-velocity zone. Observations of large amplitude phases over a wide azimuthal range requires a shallow dip and a large AV . The best fit to these S  arrivals (e.g. see Figure 4.8) is provided by a very prominent low-velocity zone extending from 37^-1 km depth and having an S-velocity contrast of 1.0 km/s at both the top and bottom boundaries. This large AV is anomalous compared to the P-velocity contrast of S  0.8 km/s estimated for this feature (Drew and Clowes, 1990) and suggests a Poisson's ratio other than 0.25. This is examined further in the next section. The optimum dip direction is found to be N10°E ±20° and 5 is 7° ± 5°. This dip direction is in good agreement with that estimated for the top of the ' E ' reflective zone from the LITHOPROBE reflection data (see contours in Figure 4.5 inset). In this model both the ' C and ' E ' low-velocity zones have been thickened relative to the Drew and Clowes (1990) model (from - 2 km to ~5 km each). Thus, to preserve the average velocity of the mid to lower crust the V of this high-velocity p  region has been increased to 7.4 km/s from 7.1-7.2 km/s. It is noted that V estimates for p  this region range from 7.1 km/s (Drew and Clowes, 1990) to 7.7 km/s (Spence et al., 1985).  Estimation of Poisson's Ratio in The 'E' Zone The data indicate that a large A V (1.0 + 0.2 km/s) is associated with the top and bottom of S  the reflective ' E ' zone. Yet the refraction interpretation of Drew and Clowes (1990) suggests AVp = 0.8 km/s at these boundaries. Clearly these two estimates are incompatible with a=0.25, for which A V = 1.0 km/s would require a AVp = 1.7 km/s. Thus, Poisson's ratio S  appears to be anomalous in this region and warrants further investigation. Forward modelling of the phases 'ET and 'EQ' does not permit a unique determination  95  Chapter 4. MODELLING RECEPVER FUNCTIONS  of a in the ' E ' zone due to tradeoffs between layer thickness, V and a. Estimates of o are s  made by considering the A V estimate of 1.0 ± 0.2 km/s with the AV estimate of 0.8 km/s. S  P  It is noted that the latter value is poorly constrained and therefore bounds of +50% are considered. Additional constraints are provided by P-wave reflection coefficients. Calvert and Clowes (1990) have determined ' E ' zone reflection coefficients as large as 11-22% on the offshore lines 85-01 and 85-05 (Fig. 1.3a). They point out that these are upper bounds as they represent the most prominent reflections, and may be partially attributed to constructive interference. The AVp associated with a AV of 1.0 km/s may be reduced by either increasing a in the S  ' E ' zone or by decreasing o in the layers both above and below the ' E ' zone. The latter case (assuming a = 0.25 in the ' E ' zone) requires unrealistic values of c for the region above and below the ' E ' zone (i.e. a = 0.14 if AV  p  = 0.8 km/s) and is not considered further.  Table 4.1 summarises the possible combinations of Poisson's ratio in the ' E ' zone and reflection coefficients for a range of AVp from 0.4 to 1.2 km/s for the former case. The preferred Poisson's ratio for the ' E ' zone, based on the AVp estimate of 0.8 km/s (Drew and Clowes, 1990) is 0.34 with a corresponding reflection coefficient of 10%. The ±50% bounds considered for the AVp estimate permits a range of Poisson's ratio from 0.31-0.36, and reflection coefficients of 5-15%. Further, allowing for the uncertainty in the AV  S  estimate  of +0.2 km/s, a range of Poisson's ratio from 0.27 to 0.38 is permitted (Table 4.1). This is a relative Poisson's ratio in that a is assumed to be 0.25 in the region above and below the ' E ' zone. Thus it appears that within the ' E ' reflective zone V and V are low relative to p  s  the surrounding material, and a is high. It is noted that at Corvallis, Oregon (-500 km south of A L B - B ) , Poisson's ratio was estimated to be > 0.33 in a low-velocity zone ( V = 3.3 s  km/s) interpreted between 28-45 km depth (Langston, 1981). A similar estimate was made for a 10 km thick low.S-velocity channel (3.4 km/s) interpreted at the base of the crust in eastern Canada (Jordan and Frazer, 1975). Both of these studies were based on the analysis  Chapter 4. MODELLING RECEIVER FUNCTIONS  AVp A V km/s 0.4 0.8 0.8* 0.8 1.2 0.8 0.4 1.0 0.8* 1.0 1.2 1.0 0.4 1.2 0.8* 1.2 1.2 1.2  S  v /v  a  2.01 1.90 1.78 2.13 2.01 1.89 2.27 2.14 2.01  0.34 0.31 0.27 0.36 0.34 0.31 0.38 0.36 0.34  p  s  RC % 5.1 9.9 15.0 5.1 9.9 15.0 5.1 9.9 15.0  Table 4.1: ' E ' zone V /V , c, and reflection coefficients versus AVp and A V p  96  s  S  V and a in the layers above and below the ' E ' zone are fixed at 0.25 and 7.4 km/s, respectively. A (*) symbol denotes the A V estimate of Drew and Clowes (1990), other values represent ± 50% bounds on this estimate. R C are reflection coefficients calculated based on P-wave velocities and densities. p  p  of long-period teleseismic body waves.  4.5.5  Upper Mantle S-Velocity Structure  The only feature imaged in the upper mantle beneath A L B - B is the subducting Juan de Fuca plate (JdF). The phases ' F ' and ' O M ' are interpreted as converted phases from the top of the plate and the oceanic Moho, respectively. They are observed most clearly over the B A Z range 320°-130° (see Figure 4.8), and may be present over all B A Z ' s (e.g. the double-peaked nature of the arrival near 5 s in receiver functions from the SW quadrant is interpreted as resulting from interference of the phases ' E B ' and ' O M ' ) . The best fit to ' F ' and ' O M ' is provided by a low-velocity zone, corresponding to the Juan de Fuca plate, dipping 15° + 5° in the direction N30°E ±20°. The top of the plate is at 47 km depth and the oceanic Moho is at 53 km depth. The velocity structure is similar to that determined for the Juan de Fuca plate to the west of Vancouver Island (Waldron et al., 1990).  Chapter 4. MODELLING RECEIVER FUNCTIONS  4.5.6  97  A L B - B Summary  The S-velocity model derived in this study, as well as a comparison of the P-velocity model to that of Drew and Clowes (1990) is given in Figure 4.9. In addition, the final model is presented in tabular form in Appendix B (Table B . l ) . The primary differences between the V profile estimated from this analysis, and that of p  Drew and Clowes (1990) are; 1. A 0.5 km thick low-velocity surface layer, attributed to weathering has been resolved. 2. A significant S-velocity boundary has been imaged at 11 km depth. This boundary may correlate with a prominent reflector near 3 s. 3. The ' C and ' E ' low-velocity zones have been thickened to ~5 km each, and the Pvelocity of the 'high-velocity wedge' has been increased to 7.4 km/s from ~ 7.1 km/s. 4. The top of the Juan de Fuca plate has been imaged at 47 km. In the refraction interpretation of Drew and Clowes (1990) this value was set at 44 km, however it is noted that the presence of the subducting plate is not required by the refraction data. Synthetic receiver functions are compared to the ±1 standard deviation bounds on the mean of the stacked data (both radial and transverse) in Figures 4.10-4.13. Note that in contrast to previous figures, synthetic data are now represented by solid traces. Overall there is good agreement between observed and synthetic data. The exception is receiver functions over the B A Z range 170°—245° (Figure 4.12). The arrival near 5 s in these receiver functions, which appears to be a double phase, is believed to result from interaction of the 'Eg' phases and ' O M ' phases. Receiver functions over this B A Z range contain several large amplitude arrivals near 2.5^4.0 s which appear to be either reverberations or scattered energy. The reader is reminded that the emphasis was placed on matching the highest quality receiver functions. The final model provides a good fit to all arrivals over the B A Z range  Chapter 4. MODELLING RECEIVER FUNCTIONS  ALB-B  VELOCITY STRUCTURE  0  In  10  20  L  Solid: This Study Dots: Drew a n d Clowes ( 1 9 9 0 )  Dip 2 5 N50E  1  1 i  98  P 15 N50E  C  D I  30 Dip 7 N10E  40 -M CL  O  L_  50  J  d  F  Dip 15 N30E  60 h  V  70  80 3  5 Velocity  6  7  8  10  (km/s)  Figure 4.9: Final V model at A L B - B derived from this analysis (dashed line), and a comparison of the V structure (solid line) with the interpretation of Drew and Clowes (1990) (dotted line), a is 0.25 with the exception of the ' E ' low-velocity zone where it is 0.34 (see text). s  p  Chapter 4. MODELLING RECEIVER FUNCTIONS  0  5 T i m e (s)  99  0  5 T i m e (s)  Figure 4.10: Synthetic receiver functions at a B A Z of 130° are compared to the ±1 standard deviation bounds on the mean of the stacked data. A l l traces are plotted at the same scale.  Chapter 4. MODELLING RECEIVER FUNCTIONS  100  A L B - B B A Z = 300° Radial  0  Transverse  5  Time (s)  0  5  Time (s)  Figure 4.11: Synthetic receiver functions at a B A Z of 300° are compared to the i l standard deviation bounds on the mean of the stacked data.  Chapter 4. MODELLING RECEIVER FUNCTIONS  101  A L B - B A = 90° Radial i  i  n  i  i  i  i  i  r  "j  Transverse i  1  1  1  i—j  l  Solid: Synthetic  C  T  E OM  B  Dots: Observed Bounds  B  BAZ 130  170  230  245  260  S T 280  300  320  340  005  1  i  i  i  i  0  i  5 Time (s)  i  i  I  I  I  I  1  1  0  I  I  L  5 Time (s)  Figure 4.12: Synthetic receiver functions at A of 90° are compared to the ±1 standard deviation bounds on the mean of the stacked data.  Chapter 4. MODELLING RECEIVER FUNCTIONS  102  A L B - B A < 65° Radial  0  Transverse  5 T i m e (s)  0  5 T i m e (s)  Figure 4.13: Synthetic receiver functions at A of < 65° are compared to the ±1 standard deviation bounds on the mean of the stacked data. For single event receiver functions there is only one dotted line.  Chapter 4. MODELLING RECEIVER FUNCTIONS  300° - 5° (Figures 4.11, 4.12). Receiver functions over the B A Z range 80° - 130° are in overall agreement (Figures 4.10, 4.13), however there are some time shifts between observed and synthetic data which cannot be explained by this simplified model. For example over the B A Z range 80° —130° the arrival corresponding to the phase ' S ' in the synthetic receiver function arrives slightly early relative to the observed phase and would provide a better fit if the boundary were at 12 km depth rather than 11 km. At B A Z = 130°, the phases in the synthetic data corresponding to the ' E ' zone and JdF plate arrive slightly late suggesting that these boundaries should be at depths 1-3 km shallower than modelled. These time variations may represent lateral variations in the boundary (i.e. a non-planar interface) or lateral velocity variations beneath the station. Based on the scatter in arrival times, the estimated bounds for the depths of the major boundaries are: ' S ' : 10.5-12.0 km; the ' C and ' E ' zones: ±1.5 km; ' F ' (top of the JdF plate): 44-47 km; and ' O M ' : 50-53 km. Uncertainties in the estimated depths of boundaries also depends upon the velocity uncertainties. However, it is unlikely that refraction velocities (e.g. the overall crustal average velocity) are in error by more than a few per cent. In summary, this study has provided several new pieces of information on the earth structure beneath central Vancouver Island. (1) the ' C and ' E ' reflective zones are imaged beneath A L B - B at depths of 20-26 and 3 7 ^ 1 km, respectively. This study provides the first definitive evidence that these reflective zones are associated with regions of low S-velocity. Prior to this study they were interpreted as low-velocity zones based on their location within a wedge of high-velocity material (Drew and Clowes* 1990). The ' E ' zone is especially prominent, having a A V of 1.0 km/s at both the top and bottom boundaries. Further, the S  preferred Poisson's ratio in this reflective zone is estimated at 0.34, with bounds of 0.27-0.38 based on the uncertainties in AV and A V . (2) there is no evidence for a continental Moho p  S  beneath central Vancouver Island. (3) this study provides the first direct seismic evidence for the position and geometry of the subducting Juan de Fuca plate beneath central Vancouver  103  Chapter 4. MODELLING RECEIVER FUNCTIONS  104  Island. The depth estimate of 47 km to the top of the plate and dip direction of 15° in the direction N30°E are in good agreement with the estimated position of the plate from extrapolation of the reflection data on western Vancouver Island and offshore. This depth would indicate that the few earthquakes which occur in this region are contained within the oceanic crust.  4.6  LAS Interpretation  L A S lies at a key location along the LITHOPROBE 'Corridor' (Figure 1.3b). No deep reflection studies have been conducted in the Georgia Strait region, and therefore it is not known whether the deep reflective zones imaged in LITHOPROBE reflection data collected on the Sechelt peninsula (see Varsek etal., 1990) are associated with the ' C and ' E ' reflective zones beneath Vancouver Island. Another important question to be addressed is whether a continental Moho exists beneath Georgia Strait. The refraction interpretation of Drew and Clowes (1990) indicates the presence of a continental Moho developing to the east of central Vancouver Island, but it is unclear exactly where this occurs. Finally, based on the tension axes of the deep earthquakes, and the worldwide average depth of -100 km to the top of subducting plates at the volcanic belt (Isacks and Barazangi, 1977) it has been suggested that the JdF plate dips more steeply to the east of Vancouver Island (see Rogers, 1983). This study may help constrain the nature and geometry of the JdF plate in this region and reveal whether the concentration of seismicity - 60 - 70 km beneath Texada and Lasqueti Islands is occurring within the crust or the mantle of the subducting slab.  4.6.1  The Data  A total of 65 teleseisms recorded at L A S were analysed. These events provided 14 stacked, and 5 single event receiver functions. The data distribution is very similar to that of A L B - B ;  Chapter 4. MODELLING RECEIVER FUNCTIONS  105  with two A-profiles at B A Z ' s of 130° and 300° (Figure 4.14), arrivals from A = 90° covering a wide range of B A Z (Figure 4.15), and a number of events from A < 65° (Figure 4.16) sampling the N / N W and E/SE directions. The largest azimuthal gaps are 10° - 80° and 170° — 230° (see polar plots of Figures 4.15, 4.16). Like A L B - B , the receiver functions generally become incoherent at T - T 8  p  greater than 8-10 s. Therefore, only the first 7-8 s  of the waveforms are modelled. As illustrated in Figure 4.15, both the radial and transverse waveforms are often rapidly varying. The transverse arrivals are often as large as, or larger than the radial arrivals, indicating complex, laterally inhomogeneous structure. Note the large arrival near 2.6 s in all receiver functions at B A Z = 300° (Figure 4.14 - unlabelled, but bracketed by arrowheads). This arrival is associated with a large positive polarity transverse arrival (Figure 4.14) and, based on these data only, could be interpreted as a Ps conversion generated at a mid-crustal N E dipping interface. However, receiver functions over the range 300° < B A Z < 10° (Figures 4.15 and 4.16) reveal that the amplitude of this phase increases and at the same time it arrives earlier for B A Z ' s approaching N . This is in contrast to a Ps phase generated at a dipping interface which should increase in amplitude and arrive later for B A Z ' s approaching the dip direction. Despite many attempts to model this phase, the increasing amplitude and decreasing Ps-P time (e.g. 2.7 s at B A Z = 300° to 2.1 s at B A Z = 10°) cannot be satisfied by a Ps conversion generated at a planar dipping interface. Thus, it is concluded that this arrival represents scattered energy or a reverberation from a shallow boundary (3-7 km depth). It is noted that a prominent gravity high exists to the north and northeast of this station over the B A Z range in which this phase is observed. Dehler (1991) suggests that this is due to a shallow (~5 km depth), high-density intrusive body. The only feature common to the early portion (T< 3.5 s) of most receiver functions is a negative polarity arrival near 2.5-3.0 s (see Fig. 4.15).  Chapter 4. MODELLING RECEIVER FUNCTIONS  106  L A S RECEIVER FUNCTIONS: BAZ = 300°  Radial l I I I  -5  0  5  10  15  20  -5  Transverse Il I l l II I I I j l l I  0  Time (s)  5  10  Radial 1 1 11  j  1  i  15  20  15  20  Time (s)  L A S RECEIVER FUNCTIONS: B A Z = 130  M i l  l ] l l l  C  Transverse  iij iiiij ii  II  Number in Stack  CM  OM  A  r  40 10  50  1  85 i II i  IIII  -5  0  1 5  1 10  1i 1 1  Time (s)  11 11  15  1  11 1 1  20  -5  0  5  10  Time (s)  Figure 4.14: L A S receiver functions at B A Z ' s of 300° and 130°. A l l traces are plotted at the same scale. Phases discussed in the text are denoted by arrows.  Chapter 4. MODELLING RECEIVER FUNCTIONS  107  LAS RECEIVER FUNCTIONS: A = 90°  -5  0  5 10 Time (s)  15  20 .  -5  0  5 10 Time (s)  15  20  Figure 4.15: L A S receiver functions at A = 90°. A l l traces are plotted at the same scale. Phases discussed in the text are denoted by arrows.  108  Chapter 4. MODELLING RECEIVER FUNCTIONS  LAS  -5  0  5 10 Time (s)  RECEIVER  15  20  FUNCTIONS: A < 65°  -5  0  5 10 Time (s)  15  20  Figure 4.16: L A S receiver functions at A < 65°. A l l traces are plotted at the same scale. Phases discussed in the text are denoted by arrows.  Chapter 4. MODELLING RECEIVER FUNCTIONS  109  The phases modelled in this study are: 1. The large radial arrival near 4 s (denoted ' C M ' in Figures 4.14-4.16). The amplitude of this phase is relatively insensitive to B A Z and it is generally associated with little or no transverse energy. This is indicative of a horizontal boundary. Although this arrival is sometimes imaged as a pair of positive amplitude arrivals in the higher frequency receiver functions (e.g. 300-40, 300-50 and 10-90), low-pass filtering these data to match the frequency content of other receiver functions (e.g. 130-80, 230-80) reveals a single phase which encompasses the two observed peaks. This suggests the presence of two closely spaced S-velocity discontinuities. 2. A negative polarity phase CE?' Figures 4.14-4.16) observed near 5 s on most radial receiver functions. There is generally little or no transverse energy associated with this arrival. 3. A pair of phases ('F' and ' O M ' ) observed near 6-6.5 s in most radial receiver functions over the range 300° < B A Z < 130° (Figures 4.14, 4.16).  4.6.2  Initial Considerations  The major features of the starting model in the vicinity of L A S (Figure 1.4b) are (1) a lowvelocity zone extending from 19-37 km depth, (2) a continental Moho at 37-39 km and (3) the oceanic crust from 59-65 km depth (dipping 20°, approximately N40°E). The subducting plate is not required by the refraction data but is included in the model to be consistent with other geophysical data, primarily the seismicity at depth. The plate dip angle and direction are poorly constrained. In Figure 4.17 synthetics generated using the Drew and Clowes (1990) interpretation are compared with the observed data over a range of B A Z and A. The most prominent arrivals  Chapter 4. MODELLING RECEIVER FUNCTIONS  110  LAS RECEIVER FUNCTIONS Solid: Observed Dash: Synthetic  Transverse t i l l  BAZ-A 10-90  130-50  230-80  300-50  Figure 4.17: Synthetic receiver functions generated using the interpretation of Drew and Clowes (1990) are compared with observed data over a range of B A Z and A. Phases corresponding to major boundaries in the model (discussed in text) are labelled.  Chapter 4. MODELLING RECEIVER FUNCTIONS  111  predicted by this model are Ps conversions from the top of the subducting Juan de Fuca plate ('F') and the oceanic Moho ('OM'). These phases are somewhat larger than those observed, however both the radial and transverse components are in reasonable agreement with the observed data. Note that the transverse components of ' F ' and ' O M ' reverse polarity between a B A Z of 10° and 130°. In addition both the radial and transverse components of these phases are very small (as predicted) in the SW quadrant. These observations are consistent with a N E dipping low-velocity zone. The phase corresponding to the continental Moho ('CM') is similar to an observed arrival, albeit at a slightly earlier time, in the data. This indicates the presence of this boundary which was not observed beneath central Vancouver Island. Finally, a negative polarity arrival near 2 s ('LVZ') corresponding to the top of the low-velocity zone arrives about 1 s earlier than a negative phase observed in most of the data. This suggests the presence of a crustal low-velocity zone at a depth slightly greater than the one proposed by Drew and Clowes (1990) and Spence etal. (1985).  4.6.3  C r u s t a l S-Velocity Structure  The effects of very shallow structure are observed in the direct radial P-waves recorded at this site. Although it is not apparent in Figures 4.14-^1.16, this arrival has a maximum amplitude at 0.1 s for most receiver functions (the slight asymmetry about T = 0.0 s may be observed in these figures). In contrast to A L B - B , this effect is independent of azimuth and therefore indicative (as demonstrated in section 3.3.4) of very shallow horizontal structure. This effect is best modelled by a horizontal interface near 0.5 km depth and having an S-velocity contrast of -1 km/s. The phase ' C M ' near 4 s dominates the early portion of most receiver functions. The arrival time is consistent with a Ps phase generated at the continental Moho near 36 km depth. This phase is sometimes observed as a large single pulse (e.g. all receiver functions  Chapter 4. MODELLING RECEIVER FUNCTIONS  112  at B A Z = 130°, Fig. 4.14), but often appears as a smaller double phase (e.g. Fig. 4.14, 300—40, 300-50). It is interesting to note that the receiver functions which exhibit a double phase are dominated by slightly higher frequencies (e.g. -0.5 Hz) than those which contain only a single phase (-0.1-0.2 Hz). Note the higher resolution (e.g. see the width of the direct P-arrival at T = 0.0 s) provided by 300-50 (a = 5) relative to 130-50 (a = 3). If the raw data which are represented in the receiver function 300-50 are low-pass filtered (e.g. if a = 3 rather than a = 5 is used in the deconvolution procedure) the result is a single phase which is identical in appearance to that of 130-50. This suggests that the double-peak represents two S-velocity contrasts, each of -0.6 km/s, separated in depth by 5-7 km. Such a large S-velocity contrast (i.e. -1.2 km/s over 5-7 km) would require a low S-velocity in the lower crust. A negative arrival labelled ' C which precedes the Moho in many receiver functions (see Figure 4.18) may represent the top of a low-velocity zone which forms the base of the continental crust (as suggested in the refraction interpretation of Drew and Clowes, 1990). Thus, the top of the low-velocity zone (AV = -0.6 km/s) is interpreted at a depth of 26 km. S  This boundary is poorly constrained (see section 4.6.5) and may be as shallow as 20 km. The average crustal velocity in the final model (V = 6.54 km/s) matches that of Drew p  and Clowes (1990). In Figure 4.18, synthetics generated using this model are compared to a selection of high-quality data covering a wide range of B A Z and A.  4.6.4  Upper Mantle S-Velocity Structure  Two interesting features are imaged in the upper mantle beneath L A S . These are the subducting JdF plate, and a low-velocity zone -10-15 km above the top of the plate. The top of a significant low-velocity zone is denoted by the presence of a negative polarity arrival ('Ex) near 5 s in most receiver functions (Fig. 4.18). The amplitude of this phase is largest for arrivals over the range 300° < B A Z < 130°, indicating a N E dipping boundary, however the  Chapter 4. MODELLING RECEIVER FUNCTIONS  113  LAS RADIAL RECEIVER FUNCTIONS Solid: Observed Dash: Synthetic  Ii i i i Ii i i i I 0  5 Time (s)  10  1  i i Ii 0  i  i i Ii i ii 5  10  Time (s)  Figure 4.18: Synthetic receiver functions generated using the final model are compared with select (see text) data. Phases corresponding to major boundaries in the model (discussed in text) are labelled. Polar plots illustrate the data distribution.  Chapter 4. MODELLING RECEIVER FUNCTIONS  114  dip angle must be shallow as the arrival is observed at all B A Z ' s . This phase is interpreted as a Ps conversion from the top of a low-velocity zone (AV = -0.7 km/s) at 45 km depth S  beneath L A S . The interface dips 10°±5° in the direction N30°E ±20°. If projected to the SW, this boundary coincides with the ' E ' zone imaged beneath A L B - B . The dip direction, dip angle and A V at both locations are comparable. This will be discussed further in Chapter 5. S  As illustrated in Figure 4.17, synthetic receiver functions generated using the Drew and Clowes (1990) model contain the phases ' F ' and ' O M ' which originate at the top of the JdF plate and the oceanic Moho, respectively. Both the radial and transverse phases are in general in good agreement with observed arrivals near 6.5 and 7 s, indicating that the top of the plate is near 60 km depth and the oceanic Moho is near 65 km. These phases are largest in those arrivals from A < 65° and over the range 300° ^ B A Z < 130°. As described in section 4.6.2, transverse arrivals correlate with the radial arrivals and are also indicative of a N E dipping low-velocity zone. The optimum fit to the data is provided by a low-velocity zone dipping 22° + 5° in the direction N30°E ±20°. The S-velocity of the oceanic crust is estimated to be 0.5 km/s lower than the surrounding mantle, and the thickness of the plate is estimated at -5 km.  4.6.5  LAS Summary  The S-velocity model derived in this study, as well as a comparison of my P-velocity model and that of Drew and Clowes (1990) is given in Figure 4.19. The final model is also presented in tabular form in Appendix B . The only significant differences between the velocity structure derived from this analysis, and that of Drew and Clowes (1990) are; (1) the prominent lowvelocity zone in the upper mantle at 45-50 km, (2) the more pronounced low-velocity zone forming the base of the crust (near 30 km) and (3) the smaller velocity contrast associated with the subducting oceanic crust at 60 km depth. Although the crustal low-velocity zone  Chapter 4. MODELLING RECEP/ER FUNCTIONS  LAS V E L O C I T Y  0  115  STRUCTURE Solid: This Study Dots: Drew and Clowes (1990)  i  10 20 30 JCM  40 Dip 10 N30E  E  50 60  i J d F Dip 22 N30E  V  70 gO I  2  I  I  3  4  I  I  I  I  U  !  I  5  6  7  8  9  10  Velocity ( k m / s )  Figure 4.19: Final V model at LAS derived from this analysis (dashed line), and a comparison of the V structure (solid line) with the interpretation of Drew and Clowes (1990) (dotted line). s  p  Chapter 4. MODELLING RECEIVER FUNCTIONS  116  imaged in this study is more pronounced (i.e. V is lower) than in the refraction interpretation, s  it is thinner and therefore the travel-time is comparable. Synthetic receiver functions are compared to the +1 standard deviation bounds on the mean of the stacked radial and transverse data in Figures 4.20—4.23. Little weight is assigned to the receiver function 80-65 (top trace, Figure 4.23). It is a very poor quality deconvo^ lution and is dominated by long period data. The receiver function 130-85 (bottom trace, Figure 4.20) closely resembles the synthetic, however there is a time shift of ~ 0.5 s for all arrivals which cannot be explained. Slight time shifts are also observed for the phases ' F ' and ' O M ' at B A Z ' s of 325° and 340° (Fig. 4.23), however overall the synthetics provide a good fit to the observed data. As discussed in section 4.6.1 the first 2.5 s of many receiver functions are dominated by arrivals which exhibit rapid arrival time variations and are associated with large transverse arrivals. Such phases could not be modelled as a Ps generated at a planar interface and hence are not fit by the synthetics (e.g. see Figs. 4.22, 4.23). The only boundary (AV = 0.4 km/s) modelled in the upper crust is near 6 km depth S  and has a dip angle of 15° in the direction N50°E. Errors are not assigned to these values as the Ps generated at this interface is not observed in most of the data and therefore it is not a necessary feature in the model. This interface does however explain an arrival near 0.8 s observed in all receiver functions at B A Z = 130° (Fig. 4.20) and in some of the data at B A Z = 300° (Fig. 4.21). It is also noted that the refraction interpretation of White and Savage (1965) contains a similar boundary approximately 6 km beneath the Strait of Georgia. The top of the crustal low-velocity, zone is poorly constrained. Note that although the negative polarity radial arrival in the synthetics (labelled ' C ' ) provides a good fit to some of the data (e.g. BAZ = 300° - 325°; Figs. 4.20, 4.22), it arrives late relative to the observed phase over the B A Z range 170° - 280° (Fig. 4.22). In order to fit these arrivals, the top of this feature must be at 20-21 km depth rather than 26 km as in this model. Thus, although there is evidence for a low-velocity zone at the base of. the crust, the exact location of the  Chapter 4. MODELLING RECEIVER FUNCTIONS  117  Figure 4.20: Synthetic receiver functions at a B A Z of 130° are compared to the +1 standard deviation bounds on the mean of the stacked data.  Chapter 4. MODELLING RECEIVER FUNCTIONS  118  LAS BAZ = 300° Radial  i  I  i  i  ' i  i  0  Transverse  I  5 T i m e (s)  i  i  I  I  i  I  i  i  i  i  0  I  i  i  5 T i m e (s)  Figure 4.21: Synthetic receiver functions at a B A Z of 300° are compared to the +1 standard deviation bounds on the mean of the stacked data. Single event receiver functions are represented by a single dotted line.  Chapter 4. MODELLING RECEWER FUNCTIONS  119  L A S A = 90° Radial i  i  i  I  i  i  i  i  i  i  i  i  0  ri  i  1i  i  5 Time (s)  Transverse 1  I  i  I  i  j  i  i  i  i  I  i  i  i  I  i  i  "  Solid: Synthetic  I  1  i  i  0  i  5 Time (s)  Figure 4.22: Synthetic receiver functions at A of 90° are compared to the ±1 standard deviation bounds on the mean of the stacked data. Single event receiver functions are represented by a single dotted line.  Chapter 4. MODELLING RECEIVER FUNCTIONS  LAS A <  120  65°  Radial l—|—i—i—i—i  Transverse |  i  i  I  j  i  |—i  i  i  i  |  i  r  Solid: Synthetic  I iI JI II  0  5  Time (s)  I  I  1  lIl J lI Il I 0  5  Time (s)  Figure 4.23: Synthetic receiver functions at A of < 65° are compared to the +1 standard deviation bounds on the mean of the stacked data. Single event receiver functions are represented by a single dotted line.  Chapter 4. MODELLING RECEP/ER FUNCTIONS  121  top of this feature is poorly constrained. The base of the crustal low-velocity zone and the continental Moho ('CM') at 31 and 36 km depth respectively, reproduce the observed double-peak in higher frequency receiver functions (e.g. 300-40, 300-50) and the broad pulse of the lower frequency data (e.g. 245-90, 260-90) observed near 3.5-4.0 s. The phase 'ET,  associated with the top of the low-velocity zone at 45 km depth also provides a good  fit to most of the observed data. In contrast to A L B - B however, a large positive arrival associated with the base of this zone is not observed. Thus, rather than a first order velocity discontinuity at the base of this zone, a velocity gradient over 7 km has been used. The phases ' F ' and ' O M ' , generated at the top of the JdF plate (60 km) and the oceanic Moho (65 km) respectively, provide a good fit to most receiver functions. They are observed only over the range 300° <BAZ< 130°. The transverse arrivals are also in agreement with the data; they are of low amplitude in the SW quadrant (Fig. 4.22) and have large amplitude at B A Z ' s of 130° and 300°. Note the polarity reversal between these two B A Z ' s (Figs. 4.20, 4.21). With the exception of the top of the crustal low-velocity zone (at 20-26 km depth), the depth uncertainty associated with each interface is estimated at +1.5 km. In summary this analysis indicates that a continental Moho exists at a depth of 36 km beneath the Strait of Georgia. In addition, the data suggest that a low-velocity zone forms the base of the crust. The upper boundary of a prominent low-velocity zone (AV = -0.7 km/s), S  which correlates with the ' E ' zone beneath central Vancouver Island, has been imaged at a depth of 45 km. Finally, this study provides the first direct seismic constraints on the geometry of the subducting JdF plate beneath the Strait of Georgia. The oceanic crust is at a depth of 60-65 km and dips 22° in the direction N30°E.  Chapter 4. MODELLING RECEIVER FUNCTIONS  122  EGM Interpretation  4.7  Located at the northeast end of the array, E G M lies within the Coast Belt (Figure 1.2). This is in contrast to L A S and A L B - B which are located within the Wrangellia Terrane of the Insular Belt. Thus, this station may provide information on the interrelationships between these two major tectonic units. A primary goal of this research is to examine the deep structure beneath this site. Few sub-crustal earthquakes occur in this region and therefore the position of the Juan de Fuca plate is poorly constrained. In addition, LITHOPROBE reflection line 88-16 located 30 km to the W N W of E G M , displays a spectacular suite of reflections from mid crustal to upper mande depths (Figure 1.3b). Clowes (1990) speculates that some of these reflective bands correlate with the ' C and ' E ' zones beneath Vancouver Island. Analysis of receiver functions may provide an answer to this question.  4.7.1  The Data  A total of 79 teleseisms recorded at E G M were used in this analysis. These provide a total of 17 stacked and 3 single event receiver functions. The data distribution is comparable to the other two stations; A-profiles at B A Z ' s of 130° (Figure 4.24) and 300° (Figure 4.25), and arrivals from A — 90° covering a wide range of B A Z (Figure 4.26). Unfortunately this station lacks arrivals from A < 50° and has azimuthal data gaps of 5° - 130° (slighdy larger than the other sites) and 170° - 230° (see polar plot of Figure 4.26).  There are several  interesting aspects of this data set. • Relative to the other stations, a significant amount of energy is present at T > 5 s (e.g. compare Figure 4.26 with Figs. 4.3 and 4.15). This suggests the presence of significant S-velocity contrasts at greater depths at this site.  EGM RECEIVER FUNCTIONS: BAZ = 130°  Radial II  II  II  II  | II  Transverse  I I | I I I I | I I II  i  i  i  i  |  i  i  i  i  |  i  i  M  |  i  i  i  i  |  i  i  i  i  Number in Stack  A  5  A  4  A  A 50 60: 75  f I I I I  -5  II 0  85 I I  I I I I I I I' 5 10 Time (s)  I I I  II 15  I I I  I I I I  20  -5  II 0  ! I I  IIII III 5 10 Time (s)  II 15  I I I  I II  Figure 4.24: E G M receiver functions at a B A Z of 130°. Polar plot (far right) illustrates the B A Z and A for each trace (see triangles). A l l traces are plotted at the same scale. Phases discussed in the text are denoted by dotted lines.  20  EGM RECEIVER FUNCTIONS: BAZ = 300  c  Transverse  Radial l I I M I I I i I I  li  i I i i i i I I I i i i I i i ii  Number in Stack  I  -5  I  I  I  0  I  I I I  I  I  I  I  I  I  5 10 Time (s)  I  I  I. I  I  15  I  !  i  I I  20  i  i  i  i  0  i  i  i  i  i  i  i  i  i  i  5 10 Time (s)  i  i  i  i  i  i  i  i i  15  Figure 4.25: E G M receiver functions at a B A Z of 300°. Polar plot (far right) illustrates the B A Z and A for each trace (see triangles). A l l traces are plotted at the same scale. Phases discussed in the text are denoted by dotted lines.  20  9  EGM  RECEIVER  FUNCTIONS: A  Radial  -5  0  5 10 Time (s)  = 90°  Transverse  15  20  -5  0  5 10 Time (s)  15  Figure 4.26: E G M receiver functions at a A of 90°. Polar plot (far right) illustrates the B A Z and A for each trace (see triangles). A l l traces are plotted at the same scale. Phases discussed in the text are denoted by dotted lines.  20  Chapter 4. MODELLING RECEIVER FUNCTIONS  126  • Receiver functions from the north (lower 2 traces, Fig. 4.26) are dominated by extremely large amplitude arrivals (both radial and transverse) and have a "ringing appearance". • The 2 large phases near 8-9 s display a very unusual pattern. Over the B A Z range 5°-300° (from N to NW) their amplitude decreases, however over the range 300°-245° (NW to SW) their amplitude increases. Such behaviour cannot be explained by Ps phases generated at planar boundaries. • The transverse receiver functions at B A Z = 300° (Figure 4.25) are dominated by very large arrivals between 6-12 s and have a "ringing appearance". Some of the phases correlate with radial arrivals but most do not. The data suggest very complex structure in the vicinity of this station. No attempt is made to model arrivals late in the waveform (> 8 s) where large transverse arrivals, interpreted as scattered energy dominate many receiver functions. Only those arrivals which are observed over a wide range of both B A Z and A and exhibit the pattern of Ps phases are modelled. Nonetheless, the final model derived at this site must be considered less reliable than the other sites due to the presence of significant scattered energy. The arrivals which are modelled in this study are: 1. A positive polarity arrival (A) near 0.8-1.0 s on most receiver functions (Figs. 4.244.26). The largest amplitudes are observed over the B A Z range 170°-245°, suggestive of a SW dipping interface. 2. The negative and positive polarity arrivals ( ' B ^ ' and ' B g ' , respectively) near 1.5 and 2 s on most receiver functions (Figs. 4.24-4.26). 3. The negative polarity arrival ('C') near 3 s (Figs. 4.24-4.26).  Chapter 4. MODELLING RECEIVER FUNCTIONS  127  4. Two sets of negative/positive pairs, collectively called ' E ' , which arrive near 5-7 s and are observed over most B A Z ' s (Fig. 4.26). They increase in amplitude and arrive later for B A Z approaching N (consistent with N or N E dipping interfaces) and are clearly observed in all receiver functions at B A Z = 300° (Fig. 4.25). 5. The negative and positive polarity phases ('F' and ' O M ' ) which are best observed near 7-8 s over the B A Z range 300° - 5° (Figs. 4.25, 4.26). Again, at this station transverse amplitudes are problematic. Generally their arrival times do not correlate with phases interpreted as Ps on the radial receiver functions. In addition, the azimuthal variation in amplitude and polarity which is predicted by planar dipping interfaces is not observed. For example if the very large amplitude arrivals observed in the transverse receiver functions at B A Z = 300° (Figure 4.25) are associated with Ps phases generated at dipping interfaces, similar large transverse arrivals of the opposite polarity should be observed at the opposite B A Z (130°). This is clearly not the case (see Fig. 4.24). Therefore, only radial waveforms are actively modelled. The fit provided to the transverse data by synthetics generated using the final model is presented later in this chapter.  4.7.2  Comparison With L I T H O P R O B E Reflection Data  E G M is 27 km SE of LITHOPROBE reflection line 88-16 (Figure 4.27). Using the technique outlined in section 4.3, the correspondence between the receiver functions and the reflection data is examined.  Only those arrivals identified as Ps phases (and from A > 80°) are  considered in this analysis. Using equation 4.1, 2T times are calculated for the following p  phases: ' A ' (2T = 2.3±0.3 s); \ B ' and \ B ' (2T = 3.7+0.2 s and 5.0+0.2 s, respectively); p  r  f i  p  ' C (2T = 8.6 + 0.3 s); the first pair of ' E ' phases \ E 1 ' and 'Elg' (2T = 12.4 + 0.8 s P  T  P  and 13.910.6 s, respectively) and the second pair of ' E ' phases E2 ' and 'E2B' (2T = l  T  15.3 + 0.9 s and 16.6+ 1.1 s, respectively).  P  Figure 4.27: LITHOPROBE line 88-16 coherency filtered seismic reflection section. Dots and vertical bars represent 2T values estimated from Ps-P times and the scatter associated with these estimates respectively, for the major phases discussed in the text (after Clowes, 1990). ' C M ? ' denotes the possible location of the continental Moho. E G M is located 27 km to the SE of the point indicated (see insert). P  _ oo  Chapter 4. MODELLING RECEIVER FUNCTIONS  129  The phases 'B^ and BB' may be associated with the reflective band imaged between C  4-5 s two-way travel-time (TWTT) and phase ' C may coincide with the top of the thickest reflective zone between 8 and 12 s TWTT (Fig. 4.27). Although the ' E ' phases are not associated with any specific feature, it is interesting to note that they lie within a thick band of N E dipping reflectors. If these Ps phases are associated with these reflective regions, the polarity of these phases (negative for 'BT\ ' C , 'EIT and 'El?; 'EIB'  and  'E2B')  and positive for ' £ # ' ,  would suggest that the ' B ' and ' E ' reflective zones, and the thickest zone  of reflectors between 8-12 s TWTT are regions of low S-velocity. However, with a distance of 27 km between these two data sets, one must be cautious in attaching much significance to this comparison.  4.7.3  Initial Considerations  The starting model, based on the refraction interpretation of Drew and Clowes (1990), is very similar to that used at L A S (Fig. 1.4). The major features are a low-velocity zone at the base of the crust, a continental Moho near 39 km and a low-velocity zone associated with the JdF plate near 70 km depth and dipping approximately 25° in the direction N40°E. The latter feature is not required by the refraction data upon which this starting model is based. Synthetics generated using this model are compared to observed data over a range of B A Z and A in Figure 4.28. None of the phases contained in the synthetics ( ' L V Z ' - generated at the top of the crustal low-velocity zone, ' C M ' - the continental Moho, ' F ' and ' O M ' corresponding to the top of the JdF plate and the oceanic Moho, respectively) provide a good overall fit to the observed data. The phases which provide the best agreement are those associated with the subducting plate (e.g. both the radial and transverse components at B A Z = 5° and the radial component at B A Z = 300°). It is interesting to note that the "ringing" observed on the transverse receiver functions at B A Z = 300° begins at the predicted arrival  Chapter 4. MODELLING RECEIVER FUNCTIONS  EGM  130  RECEIVER FUNCTIONS Solid: Observed Dash: Synthetic  Radial  0  5 Time (s)  Transverse  10  0  5 Time (s)  10  Figure 4.28: Synthetic receiver functions generated using the interpretation of Drew and Clowes (1990) are compared with observed data over a range of B A Z and A. Phases corresponding to major boundaries in the model (discussed in text) are labelled.  Chapter 4. MODELLING RECEIVER FUNCTIONS  131  time of Ps from the top of the oceanic crust. Overall, the poor agreement between the synthetics and the data indicates that the earth structure is far more complex in this area than suggested by the refraction data. It is noted however that the velocity structure below 18 km on the mainland is the most poorly constrained portion of the refraction model (Drew and Clowes, 1990).  4.7.4 Crustal S-Velocity Structure There is no evidence for the presence of very shallow structure at this station. The first arrival in the receiver functions, phase A , is interpreted as a Ps conversion generated at a boundary near 6.5 km depth and having a A V of 0.5 km/s. The amplitude of this arrival is largest g  over the B A Z range 170° - 230°, where the data are best fit (Figure 4.29) by an interface dipping 15° in the direction S10°E. However, a south dipping interface cannot explain the large amplitude arrival near 1 s in 5-90. A horizontal boundary with A V -0.5 km/s will S  adequately satisfy this phase, but conversely cannot explain the increasing amplitudes over the B A Z range 170° - 230°. This apparent conflict in dip angle may represent real lateral variations of this interface. The phases contained in the receiver functions 5-90 and 170-80 are generated approximately 3-5 km apart, thus this interface may have topography of this scale length associated with it. The phases ' B y ' and 'B#', of negative and positive polarity respectively, indicate the presence of an upper crustal low-velocity zone. They are observed over the entire B A Z and A range and are best fit by a low-velocity zone (AV = -0.4 km/s) extending from 10.5-15.5 km S  depth (Fig. 4.29). The negative polarity phase ' C requires a significant low-velocity zone (AV =-0.65 km/s) in the lower crust. This phase is observed in most receiver functions at S  B A Z = 130° and 300°, and with some slight time shifts, over the entire B A Z range. This arrival is best modelled by a horizontal boundary.  Chapter 4. MODELLING RECEPVER FUNCTIONS  132  E G M RADIAL RECEIVER FUNCTIONS Solid: Observed Dash: Synthetic  0  5 Time (s)  10  0  5  10  Time (s)  Figure 4.29: Synthetic receiver functions generated using the final model are compared with select (see text) receiver functions. Labelled phases are discussed in the text.  133  Chapter 4. MODELLING RECEIVER FUNCTIONS  In contrast to L A S , the continental Moho is enigmatic at this site. The time window in which Ps phases generated at this boundary (at 35-40 km depth) are expected is 3.54 s, but only over the range 300° < B A Z < 5° are large radial arrivals observed at this time. However, it is noteworthy that the arrivals at B A Z = 300° are "double-peaked" and very closely resemble some continental Moho arrivals observed at L A S (e.g. compare E G M receiver functions 300-35, 300-50 and 300-60 of Fig. 4.29 with L A S receiver functions 30050, 325-60 and 340-50 of Figure 4.18). Based on this similarity, the final model includes 2 discontinuities (AV = 0.3 and 0.5 km/s) at depths of 30 and 36 km beneath E G M . These may S  represent the base of the low-velocity zone associated with phase ' C ' , and the continental Moho respectively. However, these features are observed only at B A Z = 300° and therefore are weakly constrained.  4.7.5  Upper Mantle S-Velocity Structure  Significant S-velocity discontinuities in the upper mantle are suggested in general terms by the energy content of the receiver functions at T — T > 5 s, and specifically by the large s  p  amplitude arrivals (collectively labelled ' E ' ) which are observed over a wide range of B A Z and A. The first pair of phases (negative and positive polarity near 5 and 5.5 s, respectively) are modelled as Ps conversions from a low-velocity zone (AV = -0.7 km/s) at 44-47 km S  depth. These arrivals are larger at B A Z = 300° relative to B A Z = 130° indicative of a more northerly dipping zone. A dip direction of N10°E ±20° and 5 = 15° ± 5 ° provides the best fit to these observations (Figure 4.29). The second set of arrivals near 6.0 and 6.5 s are very similar to the first set and are observed over the same azimuthal and distance range. They are modelled as Ps conversions from a low-velocity zone (V = -0.8 km/s) dipping 15° ± 5° s  in the direction N30°E ±20°. The deepest feature in the model provides a fit to some observed arrivals, but cannot be  Chapter 4. MODELLING RECEIVER FUNCTIONS  134  considered a requirement of this data set. As illustrated in Figure 4.28, the model of Drew and Clowes (1990) provided reasonable agreement between observed and synthetic data for Ps phases associated with the subducting plate (extrapolated to greater depths). This feature has been maintained in the model to explain some of the large amplitude arrivals (both radial and transverse) near 7-8 s (Figure 4.29). There are some time shifts, as discussed in the next section, however 5 = 30° in the direction N30°E provides the best overall fit. The top of the plate is modelled at a depth of 70 km with an S-velocity contrast of -0.7 km/s.  4.7.6  E G M Summary  The S-velocity model for E G M , as well as a comparison of my P-velocity model and that of Drew and Clowes (1990) is given in Figure 4.30, and presented in tabular form in Appendix B . Relative to A L B - B and L A S , the earth model developed at this site is subject to larger uncertainties for several reasons. Beneath the B.C. mainland the refraction interpretation of Drew and Clowes (1990) is poorly constrained at depths >18 km. In addition, there are few deep earthquakes and hence V /V is also poorly constrained. Thus, the uncertainty p  s  in V and V /V in the vicinity of this station results in larger depth uncertainties associp  p  s  ated with the interpreted interfaces. Also, the presence of large transverse arrivals (e.g. at B A Z = 300°) which are clearly not associated with Ps phases indicates that scattered energy or reverberations are significant at this site. However, the largest transverse phases generally arrive after the time window considered in this study (T = 0.0 to 8 s). In addition, only the radial phases which exhibit the amplitude and arrival time variations expected of Ps conversions and are observed over a wide range of both B A Z and A are modelled. The correlation between the major boundaries imaged beneath this site and those observed at A L B - B and L A S , supports the interpretation of the modelled phases as Ps conversions and not scattered energy.  Chapter 4. MODELLING RECEIVER FUNCTIONS  VELOCITY S T R U C T U R E  EGM 0  i  .  Solid: This Study Dots: Drew and Clowes (1990)  L A _ D f p 15 'S10W  10  135  20  c  30 E sz  40 •|f  CL  J  1  Q  i i i  50  Dip 15 N10E  Dip 15 - - - - : N30E F  60  9  70  Dip 3 0 N30E .  80 3  4  !  5 Velocity  6  7  8  10  (km/s)  Figure 4.30: Final V model at E G M derived from this analysis (dashed line), and a comparison of the V structure (solid line) with the interpretation of Drew and Clowes (1990) (dotted line). It is noted that the velocity structure below 18 km depth in the refraction model is poorly constrained (Drew and Clowes, 1990). s  p  Chapter 4. MODELLING RECEIVER FUNCTIONS  136  The main differences between the final and starting models are; (1) a significant boundary (A V  s  = 0.45 km/s) imaged at 6.5 km depth, (2) a low-velocity zone extending from 10-  15 km depth, (3) the more pronounced low-velocity zone in the lower crust and (4) the 2 low-velocity zones imaged in the upper mantle. The mid-crustal low-velocity zone interpreted in this study which is not present in the refraction interpretation may be a local feature. Ps phases associated with this boundary (arrivals 'BT' and 'BE' in Figures 4.24—4.26) originate within approximately a 3 km radius of E G M . Synthetic receiver functions are compared to the ± 1 standard deviation bounds on the mean of the stacked radial and transverse data in Figures 4.31-4.33.  Synthetics generated  using the final model provide a good fit to radial data over the entire A range at BAZ= 130° and B A Z = 300° (Figures 4.31, 4.32), and a reasonable fit to data over the entire B A Z range (Figure 4.33). The exceptions are receiver function 170-80 (Figure 4.33) which is a very noisy trace (note the large error bounds on most phases), and the 3 receiver functions in the range 320° < B A Z < 5° (Figure 4.33). In the latter case the synthetics match the overall appearance of the observed data, however the amplitudes are low and the arrivals appear to be time shifted. It is noted that the extremely large amplitudes (both radial and transverse) and "ringing" appearance of these 3 receiver functions cannot be modelled with any reasonable earth structure. As expected, the observed transverse data are not well matched by the synthetics. However, as pointed out earlier they do not exhibit the pattern that one expects of transverse phases associated with Ps. It is noted that the synthetic transverse arrivals associated with the ' E ' , ' F ' and ' O M ' phases provide a very good fit for all data at BAZ= 130° (Figure 4.31) and are at least similar in nature to observed transverse phases throughout the data set (Fig. 4.33). As at the other stations there are some Ps arrival time variations as a function of B A Z which cannot be reproduced by this simple model. The arrival ' A ' in the synthetic receiver  Chapter 4. MODELLING RECEIVER FUNCTIONS  137  EGM RECEIVER FUNCTIONS: BAZ = 130° Radial  i  I 0  i  i  i  i  I i 5  Time (s)  Transverse  i  i  i  I 10  I  i  I 0  i  i  i  i  I 5  i  i  ii 10  Time (s)  Figure 4.31: Synthetic receiver functions at a B A Z of 130° are compared to the ±1 standard deviation bounds on the mean of the stacked data.  Chapter 4. MODELLING RECEIVER FUNCTIONS  138  E G M R E C E I V E R F U N C T I O N S : B A Z = 300°  Radial  0  5 Time (s)  Transverse  10  0  5 Time (s)  10  Figure 4.32: Synthetic receiver functions at a B A Z of 300° are compared to the +1 standard deviation bounds on the mean of the stacked data. Single event receiver functions are represented by a single dotted line.  Chapter 4. MODELLING RECEIVER FUNCTIONS  139  EGM RECEIVER FUNCTIONS: A = 90° Radial  0  5 T i m e (s)  Transverse  10  0  5  10  T i m e (s)  Figure 4.33: Synthetic receiver functions at A of 90° are compared to the +1 standard deviation bounds on the mean of the stacked data. Single event receiver functions are represented by a single dotted line.  Chapter 4. MODELLING RECEIVER FUNCTIONS  140  functions provides a good match to the observed radial data over the B A Z range 170°-280°, but is early for some data at BAZ= 130° and 300° (Figs. 4.31, 4.32). A depth of 8.5 km rather than 6.5 km for interface ' A ' would provide good agreement however. Similarly, the phases ' B y ' , ' B g ' , ' C , and ' E ' are associated with slight time shifts which would correspond to uncertainties of +1-2 km in the depth estimates. Some features in the final model are very poorly constrained. These include the velocity jumps of +0.3 and +0.5 km/s at 30 and 36 km depth. This structure is very similar to the earth model developed 40 km to the SW at L A S , and provides a good match to the observed data at BAZ= 300°, however it provides a poor fit to much of the data set. Replacing these boundaries with a gradient would be a viable alternative. The subducting JdF plate is included in the model but is very weakly constrained.  Chapter 5 INTERPRETATION AND DISCUSSION  The S-velocity models derived from analysis of receiver functions provide answers to many of the questions presented in section 1.3. In this chapter the structure common to each site is first identified. The S-velocity information is then incorporated with seismicity, seismic reflection, seismic refraction, magnetotelluric and potential field results prior to interpreting the regional structure. Finally, ideas for future studies which may provide for a better understanding of this complex tectonic environment are outlined. 5.1 Regional S-Velocity Structure The low-velocity zones labelled ' C , ' E ' and 'JdF' are common to each of the models (Figure 5.1). Within uncertainties, the S-velocity contrasts associated with these features are comparable and the estimated dip angle and dip direction are consistent in that they may be projected from station to station across the study area. In addition, there is a strong correlation between the interpreted low-velocity zones and the results of previous studies including seismic reflection and seismicity (see section 5.1.1). The most prominent low-velocity zone (AV = 0.7-1.0 km/s) interpreted at each site is S  the ' E ' zone. At ALB-B it extends from 37-41 km depth and dips 7° + 5° in the direction N10°E ± 20°. At LAS the top of the ' E ' low-velocity zone lies at 45 km depth. It dips 10° + 5° in the direction N30°E ±20°. One complication is that although a negative polarity Ps conversion from the top of this zone is imaged, the corresponding positive polarity arrival is only observed occasionally. This could indicate the presence of a gradational lower 141  Chapter 5. INTERPRETATION AND DISCUSSION  A L B - B  4 7 km •  - • L A S  142  EGM  37k m  O r — 10  20  -i Dip 15 S10W  Dip 25 •N50E  1  o 10  20  Dip  N50E  6  ^  xi  30  30  40  Dip 7 -N10E  M  M  E  * J  a, <u Q  50  10 N30E Dip  JdF  Dip 15  ~N30E  Dip 15 N10E  60  Dip  22  80  El  Dip 15 N30E  ridF  70  E2  I  I  3.0  L  4.5  V  s  km/s  _!  6.0  I  I  3.0  I  4.5  V  S  _l  6.0  I  1  I  3.0  70  dF U U  4.5  V  60  j  Dip 30 K30E I  40  50  N30E  J  B  !_  6.0  80  S  Figure 5.1: S-velocity versus depth profiles for the three array sites. The ' C \ ' E ' and 'JdF' low-velocity zones are observed at each station. The continental Moho ('M'), develops between central Vancouver Island and the Strait of Georgia.  Chapter 5. INTERPRETATION AND DISCUSSION  boundary or a scattered arrival or reverberation masking this phase. At E G M , negative and positive polarity arrivals corresponding to the top and bottom of the ' E ' zone respectively, are observed. Here 2 pairs of negative/positive arrivals suggests the presence of two low-velocity zones (AVg =0.8 km/s), E l and E2, at depths of 44-^7 km and 53-57 km, respectively. Both zones have a similar A V s , dip direction, dip angle and coincide with a zone of reflectivity (Fig. 4.27). E2 coincides with a projection of the N E dipping ' E ' zone as imaged at L A S , whereas the shallower low-velocity zone at E G M lies at the same depth as the ' E ' zone at LAS. The horizontal low-velocity zones interpreted near 26-31 km depth at both L A S and E G M may represent an extension of the ' C reflective zone which lies 20-26 km beneath A L B - B . This is based on their similar A V (0.5-0.6 km/s) and their possible correlation (at S  A L B - B and EGM) with a reflective zone. There is no evidence for a continental Moho beneath central Vancouver Island, however Ps conversions generated at this boundary 36 km beneath the Strait of Georgia are clearly observed at L A S . There is some evidence for this interface, albeit with significant lateral variations, beneath the B.C. mainland. At each site the dip direction determined for the low-velocity zone 'JdF', interpreted as the oceanic crust, is N30°E. The dip angle increases from 15° ± 5 ° at A L B - B , to 22° ± 5 ° at L A S , to -30° at E G M . The oceanic crust may be projected from A L B - B (47-53 km depth) to L A S (60-65 km depth) using a dip angle of 15°. This indicates that the increase in the dip angle from - 1 5 ° to -22° occurs beneath, or very near to L A S . At E G M , the oceanic crust is interpreted at 70-75 km depth with a dip angle of -30°. If projected to the volcanic arc, the top of the plate would lie at a depth of 104 km. This compares favourably to the worldwide average depth of -100 km (Isacks and Barazangi, 1977). However, the estimated plate depth at E G M is anomalous with respect to A L B - B and L A S . The oceanic crust at E G M is -5 km shallow relative to its projected position from L A S (based on a dip angle  143  Chapter 5. INTERPRETATION AND DISCUSSION  144  of 22°). This suggests a decrease in the dip angle between L A S and E G M , yet the data indicate a steeper dip at E G M . It is noted that at this location the refraction P-velocity model is weakly constrained at depths greater than -18 km (Drew and Clowes, 1990). In addition, there are very few deep earthquakes in this region and therefore V /V is poorly constrained. p  s  Thus, an uncertainty of ±5 km in the depth of the oceanic crust is possible, and the apparent shallow depth at E G M may not be real.  5.1.1  Comparison With Previous Results  Prior to comparing the ' C , ' E ' , and 'JdF' low-velocity zones with the results of previous studies, those features which are imaged at only 1 or 2 stations are briefly examined. At A L B - B , the well defined boundary (AV = 0.6 km/s) labelled ' S ' near 10 km depth may S  be associated with the uppermost band of prominent reflectors on LITHOPROBE line 84-01 (Fig. 4.5). This has been interpreted as the contact between the Buttle Lake Limestone and Karmutsen Formations (Yorath et al., 1985). This reflector is only imaged over a distance of 10-20 km near the N E end of Line 84-01 (Yorath et al, 1985). At E G M , a Ps phase from the boundary labelled ' A ' ( A V = 0.6 km/s) near 6.5 km S  depth is observed on almost every receiver function. As described in section 4.7.4 there is some evidence for lateral variations (dip angle or direction) in this interface over a distance range of approximately 3-5 km. Based bn the interpretation of gravity data, Dehler (1991) concludes that the westernmost 20 km of the intrusive Coast Plutonic Complex (CPC) is a thin surface 'sliver' of low density material overlying the higher density Wrangellia Terrane. Approximately 10 km to the N E of E G M the C P C thickens rapidly. Her gravity model suggests that the lower boundary of the CPC 'sliver' is dipping 7.5° to the southwest at 1.5-2 km depth near E G M . Approximately 20 km to the SW of E G M , this boundary is near 5 km depth. While there are differences in the depth estimates from these two techniques,  Chapter 5. INTERPRETATION AND DISCUSSION  145  the presence of this boundary in both interpretations is noteworthy. Finally, a low-velocity zone ( A V s = 0.4 km/s) between 10.5-15.5 km depth at E G M coincides with a zone of reflectivity between 3.5-5 s on LITHOPROBE line 88-16 (Fig. 4.27). There are numerous N / N W trending, eastward dipping shear zones up to 2 km wide near the surface in this region (Monger, 1990). Thus, the low velocity zone imaged in this study may represent an extension at depth of one such shear zone interpreted along the west side of the Sechelt Peninsula (e.g. the listric fault to the west of E G M , Figure 5.2). There is a good correlation between the ' C ' , ' E ' , and 'JdF' low-velocity zones imaged in this study and other geophysical data including reflection and seismicity. At A L B - B and E G M the ' C ' and ' E ' low-velocity zones appear to coincide with prominent reflective bands (Figure 5.2). No reflection data are available for comparison at L A S . The low-velocity zone 'JdF', correlates with the seismicity at depth. The implications of this will be discussed in section 5.2.1. Receiver function analysis does not provide 'absolute' velocities, thus it is not possible to make a comparison with the P-velocity interpretations of Drew and Clowes (1990) or Spence et al. (1985). However, specific features in these interpretations may be examined. Both refraction interpretations include a low-velocity zone in the lower crust. The negative polarity arrival ' C in receiver functions determined at L A S and E G M denotes the top of such a feature in the depth range 20-26 km. Also the refraction interpretation includes a continental Moho near 37-39 km depth originating to the east of central Vancouver Island. At L A S there is clear evidence for this boundary at 36 km depth. Thus, the results of this study are in agreement with the previous refraction interpretations. However, boundaries have been imaged at greater depth, including the subducting oceanic crust to a depth of 70 km, and low-velocity zones to depths near 60 km in the upper mantle. In a previous S-velocity study using long-period body waves at VIC, Langston (1981) interpreted a high-velocity contrast (AV = 1.3 km/s) interface near 45-50 km depth. This S  Figure 5.2: Tectonic interpretation based on multidisciplinary studies (after Clowes, 1990). To the east of Georgia Strait the location of the oceanic crust is poorly constrained (stripe pattern). Based on a worldwide average, the plate is assumed to be near 100 km depth beneath the volcanic arc (Isacks and Barazangi, 1977). The location of the low S-velocity zones determined in this study are denoted by hatched areas. ' W R ' and ' m ' represent Wrangellia and metamorphic rocks, respectively.  Chapter 5. INTERPRETATION AND DISCUSSION  147  is comparable to my depth estimate of 50-53 km for the oceanic Moho beneath central Vancouver Island (Figure 5.1). At this site the A V between the oceanic crust and mantle S  is estimated to be ~1 km/s. The only other S-velocity study in this area (Wickens, 1977) is based on surface waves sampling a 350 km long travel path from P H C - V I C . His interpretation suggests the presence of low-velocity zones above discontinuities near 30 and 50 km depth (AV = 0.3 and 0.5 km/s, respectively). At A L B - B discontinuities representing the base S  of -5 km thick low-velocity zones are interpreted at 41 and 53 km depth (AV -1.0 km/s S  for both). Again, the long-period waves he analysed would not resolve thin low-velocity zones. Overall, the results of this study are consistent with previous S-wave results, but have provided more detailed models. Low-velocity zones as thin as 4-5 km have been imaged and the geometry (dip angle and direction) of interfaces has been determined beneath my 3 sites.  5.2 Results and Implications 5.2.1 Juan de Fuca Plate Subduction Geometry The results of this analysis provide the first definitive evidence for the location and subduction geometry of the JdF plate beneath central Vancouver Island and the Strait of Georgia. The top of the oceanic crust and the Moho are estimated at 44-47 km and 50-53 km depth respectively beneath A L B - B ; 58.5-61.5 and 63.5-66.5 km depth beneath L A S ; and 67.5-71.5 km and 73.5-77.5 km depth beneath E G M . Superimposing these estimates on a cross-section of seismicity along the LITHOPROBE corridor (Fig. 5.3) indicates that the majority of the deep earthquakes occur within the subducting oceanic crust. This supports the hypothesis (Rogers et al., 1990) that the Benioff zone is thermally controlled with the deep seismicity being confined to the relatively cold oceanic crust. The dip direction for the plate, estimated at N30°E ± 20° at both A L B - B and L A S is somewhat north of the direction expected based  Chapter 5. INTERPRETATION AND DISCUSSION  CQ•  CO  <  m < Vancouver  70  0ISTRNCE  '0  60  Island  148  CD LU  Gaorgia Strait  T  100  170  140  160  K O  ZOO  720  HO  HO  760  300  'KM)  Figure 5.3: The location and dip angle of the subducting oceanic crust estimated at each site is compared with earthquake hypocentres located within a 30 km wide band centred on the LITHOPROBE Corridor (cross-section A - B of Figure 1.3a). The dashed lines denote the inferred location of the upper plate based on hypocentres (after Rogers et al., 1990). The depth uncertainty in the hypocentres is ±3 km. The size of the dots is related to the magnitude (M).  Chapter 5. INTERPRETATION AND DISCUSSION  149  on the plate interaction vectors (N56°E). However, this is consistent in style with the model proposed by Rogers (1983) that the plate is folded or arched upwards in the vicinity of Puget Sound. There is growing evidence for this hypothesis based on seismicity patterns (Crosson and Owens, 1987; Weaver and Baker, 1988) and receiver function studies conducted in Washington State (Owens et al., 1988; Lapp et al., 1990) which indicate that the JdF plate has a southeasterly dip to the south of Puget Sound. A contour map of the subducted JdF Moho beneath southwestern B.C. and Washington state is given in Figure 5.4. This map incorporates the results of the seismicity and receiver function studies described above, and seismic reflection results as summarised by Hyndman et al. (1990). Finally, it has been suggested that both the concentration of seismicity beneath Lasqueti and Texada Islands (Figure 5.3) and the increasing dip angle of the plate results from the gabbro to eclogite phase change occurring within the subducting oceanic crust (Rogers, 1983). The results of this study support this hypothesis in two ways. First, the dip angle of the JdF plate increases from 15+5° to 22 ± 5 ° beneath the Strait of Georgia where the plate is at 60-65 km depth and where the deep seismicity is concentrated. Based on hypocentres, Weaver and Baker (1988) note a similar, albeit poorly constrained, increase in the dip angle of the JdF plate at depths greater than 50 km in the Puget Sound region. Secondly, the A V  S  associated with the oceanic crust is estimated to be 0.95 km/s at A L B - B but only 0.52 km/s at L A S , and 0.72 km/s at E G M . Thus, relative to A L B - B where the oceanic crust is at 4753 km depth, the S-velocity contrast associated with the subducting oceanic crust is 0.43 km/s lower at L A S (60-65 km depth), and 0,23 km/s lower at E G M (70-75 km). The S-velocities of gabbro and eclogite samples (measured at room temperature and a pressure of 10 kbar) are 3.89 km/s and 4.58 km/s, respectively (Manghnani et ai, 1974). Compensating V for s  temperature (assuming a plate temperature of T ~ 400°C) and using average temperature derivatives (after Goodwin and McCarthy, 1990) of -2.8 x IQr kms /°C for gabbro and 4  _1  -4.3X10" kms-V°C for eclogite (Birch,1943; Kern and Richter, 1981), yields S-velocities of 4  Chapter 5. INTERPRETATION AND DISCUSSION  126°  125"  124°  123°  150  122°  121°  Figure 5.4: Contour depths in kilometres to the top of the subducting Juan de Fuca plate. Contours are based on results of receiver function studies, reflection data in the Vancouver Island region (after Hyndman etal., 1990), and seismicity in Washington State (after Crosson and Owens, 1987). Dots represent the locations of receiver function studies and arrows indicate dip direction estimates. Results from stations 1 and 7 are given by Owens et al. (1988) and Lapp et al. (1990), respectively. Solid triangles represent Quaternary Cascade stratavolcanoes, which show a similar 'arch-like' pattern. Based on a worldwide average, the top of the plate should be near 100 km depth at the volcanic belt (Isacks and Barazangi, 1977). The thick arrow gives the Juan de Fuca/America interaction direction.  Chapter 5. INTERPRETATION AND DISCUSSION  151  3.78 km/s and 4.41 km/s, respectively. Thus, the decrease in A V which should accompany S  this phase change is -0.63 km/s. At L A S this is within our measurement uncertainties. In other subduction zones, a concentration of seismicity within the subducting oceanic crust, and an increase in the dip angle are believed to result from the gabbro to eclogite phase change (Pennington, 1983). This location marks the transition from a buoyant to a non-buoyant slab (Rogers, 1988), and therefore the point at which strong mechanical coupling terminates (Ruff and Kanamori, 1980).  5.2.2  The ' C ' and <E' Reflective Zones  The ' C and ' E ' reflective zones and intervening high-velocity wedge have been interpreted in several ways. Clowes et al. (1987a, b) speculate that beneath Vancouver Island the ' E ' reflective zone is the zone of active decoupling and the sequence above this represents underplated materials from the downgoing plate. Fuis and Clowes (1991) compare the geology, potential field data and deep structure of the continental margins of Alaska, Vancouver Island and California. The similarities between the Alaska and Vancouver Island margins are striking. In Alaska, a sequence of landward dipping low-velocity (V = 5.7-6.7 km/s) and high-velocity p  (Vp = 6.8-7.7 km/s) layers are imaged to depths of 40-^-50 km. The low-velocity zones contain numerous reflectors, similar in nature to the 'C and ' E ' zones beneath Vancouver Island. Based on a surface outcrop of the shallowest layer, this sequence is interpreted as an underplated terrane, likely a fragment of the Kula plate and its sedimentary overburden. The results of this study, indicating that the reflective zones beneath Vancouver Island are associated with regions of low S-velocity, further support this comparison. Like Alaska, the 4-5 km thick low-velocity zones would represent the oceanic crust and sediments, the thicker high-velocity region would represent oceanic mantle. This study indicates that the ' C and ' E ' zones have a thickness (4—5 km) similar to that of the actively subducting plate (6 km).  Chapter 5. INTERPRETATION AND DISCUSSION  152  In addition, the A V estimates of 0.5 km/s for the ' C zone and 0.7-1.0 km/s for the ' E ' S  zone are within the range of that estimated for the oceanic crust (0.5-1.0 km/s). In an alternative interpretation, Hyndman (1988) proposes that the reflective ' E ' zone represents saline fluids trapped at a metamorphic facies boundary. In this model the ' E ' zone has 1—4% porosity based on the assumption of randomly oriented ellipsoidal pore spaces with an effective aspect ratio of 0.03-0.1. Based on high reflection coefficients (up to 20%), Calvert and Clowes (1990) interpret this feature as a major shear zone above the Juan de Fuca plate. Recently, analysis of reflections observed within the accretionary wedge sediments suggests that they are associated with low seismic velocities and high pore fluid pressures (Calvert and Clowes, 1991). They suggest that the ' E ' reflective zone may represent a region in which fluids, originating by dehydration of the subducting Juan de Fuca plate, are migrating to the landward base of the accretionary complex.  ' E ' Reflective Zone Interpretation This study has provided several new pieces of information regarding the ' E ' zone. It is associated with low S-velocities ( A V = 1.0 ± 0.2 km/s), a high Poisson's ratio of -0.34 S  (section 4.5.4), and it extends into the continental upper mantle beneath the Strait of Georgia and the mainland. With this new information previous interpretations can be examined. Recall that this region is interpreted as a low P-velocity zone with AV = -0.8 km/s (Drew and p  Clowes, 1990). This value is poorly constrained, therefore bounds of ±50% are considered. Thus, the interpretation for this feature should explain the low P- and S-velocities, the high a, and the low electrical resisitivity of -30 Q-m (Kurtz et al., 1986). Theoretical studies (O'Connell and Budiansky, 1974) and in-situ measurements (Moos and Zoback, 1983) demonstrate that fluid saturated fractures decrease V and V and increase p  s  Chapter 5. INTERPRETATION AND DISCUSSION  Pore Aspect Ratio 0.01 0.003 0.001  R  Porosity (%) AVp : a AVp : a  153  AVp : a  30 Q-m 0.4 : 0.36 .0.8* : 0.34* 1.2 : 0.31 1.5 0.6 0.3  0.6 : 3.4 • 1.2 : 2.8 0.3 : 0.8 0.5 : 0.6 0.1 : 0.4 0.2 : 0.3  1.8 : 1.9 0.7 : 0.4 0.2 : 0.2  Table 5.1: Porosity vs aspect ratio for resistivity, A V (km/s) and a p  The (*) denotes the AVp estimate of Drew and Clowes (1990) and its corresponding a value. Other combinations (see Table 4.1) represent ±50% bounds on this AVp. Bold-faced numbers represent porosities (in %) which for a given aspect ratio provide the best (i.e. within 0.5%) simultaneous agreement for the three parameters. The electrical resistivity estimate of 30 Q-m is from Kurtz et al. (1986). a. Following Hyndman (1988), a theoretical model for the effect of porosity on seismic velocity as presented by Schmeling (1985) is considered. Randomly oriented, water-saturated ellipsoidal pore spaces are assumed. Pore shape is described by the aspect ratio, the ratio of the minor to the two major axes. In this analysis, a combination of aspect ratio and porosity which will simultaneously satisfy the AVp, a and electrical resistivity estimates of 0.8±0.4 km/s, 0.27-0.38, and 30 Q.-m respectively, is sought. Figure 5.5 illustrates the relationships between o and porosity, and A V  p  and porosity.  Note that an aspect ratio of >0.1 (Figure 5.5a) requires o < 0.25, in obvious disagreement with the determined value between 0.27-0.38. In addition, for the AVp range of 0.4-1.2 km/s, a pore aspect ratio of 0.03 requires porosities of < 3.5% (Figure 5.5b). However, the a estimate of 0.27-0.38 requires porosities of >4.5% (Figure 5.5a). Thus, a crack aspect ratio >0.03 cannot simultaneously satisfy the a and AVp estimates. However, thin cracks (aspect ratios of 0.001-0.01) provide simultaneous agreement to these two parameters. Table 5.1 provides the porosity values which satisfy AVp, a (assuming AV  S  = 1.0 km/s) and electrical resistivity data for aspect ratios of 0.001-0.01. Note that  numbers separated by colons in the AVp : a columns represent independent estimates of  Figure 5.5: (a) a as a function of porosity for the ellipsoidal pore model assuming random orientation of pores (after Hyndman and Shearer, 1989). Numbers on the curves indicate the aspect ratio of the pore spaces; 1 being spherical pores, 0 being infinitely thin cracks. The zero-porosity o" = 0.25. (b) AVp-porosity relation for the ellipsoidal pore model with various aspect ratios (after Hyndman and Klemperer, 1989). The zero-porosity velocity is 7.4 km/s.  Chapter 5. INTERPRETATION AND DISCUSSION  155  porosity (Figure 5.5) for valid AVp, a combinations (Table 4.1). The porosity-electrical resistivity relationship (not shown) is based on Schmeling (1986).  A n exponent of 1.5  (Hyndman, 1988) is used for Archie's law relation and partially interconnected ellipsoids are assumed. Pore aspect ratios of 0.001-0.003 and porosities of 0.1-1.9% satisfy the a, AVp and electrical resistivity estimates simultaneously. Allowing for the uncertainties in the A V estimate of ±0.2 km/s, and the a range of 0.27-0.38 (see Table 4.1), similar results to S  those given in Table 5.1 are'obtained. Thus, thin (aspect ratio 0.001-0.01), fluid-saturated fractures, and porosities of 0.1-1.9% will simultaneously satisfy the A V , o~ and electrical p  resistivity estimates for the ' E ' reflectors. Previous estimates for these parameters (Hyndman, 1988) are 0.03-0.1 and 1-4% respectively. Thus, for the ' E ' zone, the pore spaces are in the form of thinner cracks than previously believed and bounds on the porosity are lower than previous estimates. These results thereby support the interpretation of this band of deep reflectors beneath Vancouver Island as dominated by thin, fluid-saturated cracks representing a major shear zone above the subducting Juan de Fuca plate (Calvert and Clowes, 1990). Beneath L A S the top of this low-velocity zone is interpreted at 45 km. As pointed out earlier, the phase which would be expected from the bottom of this zone is generally absent. This may be the result of changes occuring within the zone in the upper mantle. At E G M the ' E ' zone appears as two low-velocity layers near depths of 44—47 and 53-57 km. It is interesting to note that in contrast to the relatively thin ' E ' reflective zone beneath Vancouver Island, a diffuse band of reflectivity extends from -30-60 km depth beneath the mainland. Thus, the shear may be distributed over a larger region within the mantle. Fluids may be trapped by structural or pressure and temperature controlled boundaries at two depth levels within this broad zone, resulting in two distinct low-velocity zones.  Chapter 5. INTERPRETATION AND DISCUSSION  156  ' C Reflective Zone Interpretation The similar A V of 0.5 km/s and thickness of ~5 km interpreted for the ' C low-velocity S  zone at each site suggests that this is one feature extending across the study area. If this interpretation is correct, it's location above the continental crust of L A S and E G M argues against the interpretation of this feature as underplated materials. It is noted that this lowvelocity zone coincides with a band of prominent reflectors at A L B - B and possibly at E G M . In addition, it is approximately isothermal at -300° C, and it is near the depth at which shallow seismicity beneath the mainland terminates (see Figs. 1.3b, 5.3). Thus, numerous geophysical data indicate that a significant change occurs near 20-26 km depth in this region. The ' C ' zone may represent a change in lithology, e.g. the base of Wrangellia as suggested by Dehler (1991), or it could represent the brittle-ductile transition zone. The latter is generally associated with a temperature of 350° + 50° and the abrupt termination of seismicity. In addition, this transition zone is often overlain by an impermeable cap (Jiracek et al., 1983), which may effectively trap fluids thereby explaining the observed low P and S-velocities. In the Cascade Range of Washington, Oregon and California, Stanley et al. (1990) invoke this interpretation to explain mid crustal velocities of -6.5 km/s coincident with an electrically conductive (2-20 Q-m) zone.  5.3  Summary and Conclusions  Receiver function analysis has proven to be a powerful tool to explore the deep structure of the Cascadia subduction zone. The major conclusions of this dissertation are summarised below.  Juan de Fuca Plate Ps conversions provide the first definitive evidence for the depth and subduction geometry  Chapter 5. INTERPRETATION AND DISCUSSION  157  of the Juan de Fuca plate to the east of central Vancouver Island. The oceanic crust is at 4753 km beneath central Vancouver Island, 60-65 km beneath Georgia Strait and 70-75 km beneath E G M on the Sechelt Peninsula. There is a good correlation with the seismicity at depth, indicating that these deep earthquakes are confined to the relatively cold oceanic crust. The dip direction of the plate at each site is N30°E ±20°. This is a more northward dip direction than the Juan de Fuca/America plate interaction vector (N56°E) would suggest and therefore provides new evidence that the Juan de Fuca plate is arched upwards in the Puget Sound region (to the south of this study area). Finally, these data suggest that the dip angle increases from 15° + 5° to 22° ± 5° and the AV$ associated with the plate decreases from 0.95 km/s to 0.5 km/s near L A S , where a concentration of seismicity occurs near 60 km depth. These observations may be explained by a phase change from gabbro to more dense eclogite occurring within the oceanic crust. The increasing dip angle at this point may denote the limit of strong mechanical coupling in this subduction zone.  The 'E' Reflective Zone The ' E ' reflective zone is associated with a low S-velocity (AV = 1.0 ± 0.2 km/s), S  and a high a (0.27-0.38) relative to the layers both above and below it. It extends into the continental mantle beneath Georgia Strait and coincides with the deep reflective band imaged beneath the mainland. Combining the results of this study with electrical resistivity and refraction P-velocity estimates, it was demonstrated that these observations are consistent with the ' E ' zone being a region dominated by thin, fluid-saturated cracks.  The 'C Reflective Zone The reflective ' C zone imaged beneath Vancouver Island may extend beneath the Strait of Georgia and B.C. mainland as a horizontal low-velocity zone near the base of the continental crust. This low-velocity zone lies just below the terminating point of shallow seismicity, is approximately isothermal at -300° C, and may coincide with a reflective zone at E G M . This zone may represent the base of Wrangellia and thus be a lithological boundary, or it may be  Chapter 5. INTERPRETATION AND DISCUSSION  158  associated with the brittle-ductile transition zone in the continental crust.  The Continental Moho There is no evidence for a continental Moho beneath A L B - B , but Ps conversions are interpreted from this feature at a depth of 36 km beneath the Strait of Georgia. There is some suggestion for the presence of this boundary, albeit with significant lateral variations, beneath E G M .  Analysis Procedures Prior to applying this technique to a dipping layer environment the applications and limitations of this method were examined. A recent modification to this technique which provides absolute amplitudes of receiver functions was considered. It is demonstrated that this procedure is more robust than the previous method of modelling normalised amplitudes. I present examples which demonstrate that the latter method may result in inaccurate earth models in the presence of dipping interfaces or shallow, high-velocity contrast boundaries. Additionally, in a dipping layer environment it is concluded that strict stacking bounds (< 10° in B A Z and A) should be applied and that the amplitude and arrival time of reverberations may be very difficult to predict, and therefore these phases should not be modelled. As reverberations represent an integral component of receiver functions, formal inversion of these waveforms is not justified in this environment.  5.4  Future Studies  One of the most interesting results of this study is the low S-velocities and high a which have been determined for the ' E ' reflective zone. These results are consistent with a region dominated by thin, fluid-saturated cracks, and thus are supportive of the recent interpretation of this feature as a major shear zone above the subducting Juan de Fuca plate. However, a number of questions remain. Does this feature extend beneath northern Vancouver Island,  Chapter 5. INTERPRETATION AND DISCUSSION  159  where the Explorer plate is being over-ridden by the America plate? Does it extend along the length of the Juan de Fuca plate (i.e. beneath Washington and Oregon)? To fully understand the history and seismic potential of the Cascadia subduction zone, the deep structure of the Juan de Fuca, Explorer, and Gorda segments must be compared. A combination of passive and active source experiments will provide answers to these questions. Receiver function data recorded in Washington state (Owens et al., 1988; Lapp et ai, 1990) which have been interpreted to place constraints on the subduction geometry should be re-examined for evidence of a low-velocity zone above the oceanic crust. At one site, it was demonstrated that short-period data provides receiver functions comparable to those generated using broadband data. This warrants further investigation, and therefore teleseismic data gathered at any 3-component, digitally recorded seismic station (short-period or broadband) in the Cascadia subduction zone should be examined. The earth structure beneath the seismic station C O R at Corvallis, Oregon was interpreted using long-period receiver function data (Langston, 1977a). This would be an ideal site to re-examine the earth structure with short-period or broadband data. If an extension of the ' E ' low-velocity zone exists above the Explorer plate (beneath northern Vancouver Island), it could be identified using receiver function studies. Stations should be located on north-central Vancouver Island where refraction data provide constraints on the P-velocity structure (Drew and Clowes, 1990; McMechan and Spence, 1983). As demonstrated in this study, coincident reflection data would also provide useful constraints. To further constrain a, AV and A V within the p  S  ' E ' reflective zone, 3-component recordings of local earthquakes (e.g. those which occur at depths of 60-65 km beneath the Strait of Georgia) should be examined for reflections or P-to-S conversions. If reflection data could be collected on Texada or Lasqueti Islands, this would fill a critical gap which currently exists. Recently, as a part of the Pacific to Arizona Crustal Experiment of the U.S. Geological Survey, seismic refraction shots were recorded into 3-component reflection spreads. This  Chapter 5. INTERPRETATION AND DISCUSSION  160  provided constraints on P-velocity and o" which were successfully used to determine the nature and composition of the lower crust (Goodwin and McCarthy, 1990). Thus, 3-component seismic data should be considered for the British Columbia mainland region where few deep earthquakes occur, and therefore a in the lower crust is poorly constrained. This would provide for a more accurate interpretation of receiver functions, and constraints on the composition of the reflective zones extending to upper mande depths. Receiver function data could be collected to the east of E G M , where Dehler (1991) interprets the main body of the Coast Belt, to examine the differences in the structure between the Insular and Coast belts. Finally, the data set presented in this study could be used to estimate scattering Q (e.g. Langston, 1989) at each of the recording sites. The phases which have been identified as scattered energy or reverberations (e.g. the arrival near 2-3 s observed over the range 300° <BAZ< 10° at L A S , or the large transverse arrivals at 7-12 s at B A Z = 300° in the E G M data) could be examined in more detail. A n array of 3-component stations could be deployed to provide important slowness and azimuth of approach information for these arrivals.  References  Ammon, C.J., Time domain teleseismic P waveform modelling and the crust and upper mantle structure beneath Berkeley, California, M.S. thesis, 97 pp., State University of New York, Binghamton, N.Y., 1985. Ammon, C.J., The isolation of receiver effects from teleseismic P waveforms, Bull. Seismol.  Soc. Am., in press, 1991. Birch, F., Elasticity of igneous rocks at high temperatures and pressures, Bull. Geol. Soc. Am., 54, 263-286, 1943. Burdick, L.J., and C A . Langston, Modeling crustal structure through the use of converted phases in teleseismic body-wave forms, Bull. Seismol. Soc. Am., 67, 677-691, 1977. Calvert, A.J., and R . M . Clowes, Deep, high-amplitude reflections from a major shear zone above the subducting Juan de Fuca plate, Geology, 18, 1091-1094, 1990. Calvert, A.J., and R . M . Clowes, Seismic evidence for the migration of fluids within the accretionary complex of western Canada, Can. J. Earth Sci., in press, 1991. Clowes, R . M . , LITHOPROBE - Multidisciplinary studies of continental evolution: A n example from Western Canada, Proc. Pacific Rim Congress '90 , Gold Coast, Queensland, Australia, 6-12 May 1990, Australasian Inst, of Min. and Metall., Parkville, Victoria, Australia, ///: 83-97, 1990. Clowes, R . M . , M.T. Brandon, A . G . Green, C.J. Yorath, A. Sutherland Brown, E.R. Kanasewich, and C. Spencer, LJTHOPROBE-southern Vancouver Island: Cenozoic subduction complex imaged by deep seismic reflections, Can. J. Earth Sci., 24, 31-51, 1987a. Clowes, R . M . , C.J. Yorath, and R.D. Hyndman, Reflection mapping across the convergent margin of western Canada, Geophys. J. R. Astron. Soc, 89, 79-84, 19876. Coney, P.J., D.L. Jones, and J.W.H. Monger, Cordilleran suspect terranes, Nature, 288, 329333, 1980. Crosson, R.S, and T.J. Owens, Slab geometry of the Cascadia subduction zone beneath Washington from earthquake hypocenters and teleseismic converted waves, Geophys. Res. Lett., 14, 824-827, 1987. 161  References  162  Davis, E.E., and R.D. Hyndman, The structure and tectonic history of the northern Cascadia subduction zone: recent accretion and deformation, Geol. Soc. Am. Bull, 101, 14651480, 1989. Davis, E.E., R . D . Hyndman, and H . Villinger, Thermal structure of the northern Cascadia accretionary continental margin, / . Geophys. Res., in press, 1991. Dehler, S.A., Integrated geophysical modelling of the northern Cascadia subduction zone, Ph. D. thesis, 151 pp., Univ. of British Columbia, Vancouver, B.C., Canada, 1991.  Drew, J.J., and R . M . Clowes, A re-interpretation of the seismic structure across the active subduction zone of western Canada, in Studies of laterally heterogeneous structure using seismic refraction and reflection data, edited by A . G . Green, Geological Survey of Canada, Paper 89-13, 115-132, 1990. Ebel, J.E., The effect of crustal scattering on observed high-frequency earthquake seismograms, Geophys. J. Int., 98, 329-341, 1989. Ellis, R . M . , and P.W. Basham, Crustal characteristics from short-period P waves, Bull. Seismol. Soc. Am., 58, 1681-1700, 1968. Fuis, G.S., and R . M . Clowes, Comparison of deep structure along transects of the western North American continental margin, in preparation, 1991. Gabrielse, H . , and C.J. Yorath, D N A G #4. The Cordilleran Orogen in Canada, Geoscience Canada, 16, 67-83, 1989. Goodwin, E.B., and J. McCarthy, Composition of the lower crust in west central Arizona from three-component seismic data, / . Geophys. Res., 95, 20,097-20,109, 1990. Green, A . G . , R . M . Clowes, C.J. Yorath, C. Spencer, E.R. Kanasewich, M.T. Brandon, and A. Sutherland Brown, Seismic reflection imaging of the subducting Juan de Fuca plate, Nature, 319, 210-213, 1986. Heaton, T.H., The calm before the quake?, Nature, 343, 511-512, 1990. Helmberger, D., and R.A. Wiggins, Upper mantle structure of midwestern United States, / . Geophys. Res., 76, 3229-3245, 1971. Hyndman, R.D., Dipping seismic reflectors, electrically conductive zones, and trapped water in the crust over a subducting plate, / . Geophys. Res., 93, 13,391-13,405, 1988.  References  163  Hyndman, R.D., and S.L. Klemperer, Lower-crustal porosity from electrical measurements and inferences about composition from seismic velocities, Geophys. Res. Lett., 16, 255-258, 1989. Hyndman, R.D., and P . M . Shearer, Water in the lower continental crust: modelling magnetotelluric and seismic reflection results, Geophys. J., 98, 343-365, 1989. Hyndman, R.D., C.J. Yorath, R . M . Clowes, and E.E. Davis, The northern Cascadia subduction zone at Vancouver Island: seismic structure and tectonic history, Can. J. Earth Sci., 27, 313-329, 1990. Isacks, B . , and M . Barazangi, Geometry of Benioff zones: lateral segmentation and downwards bending of the subducted lithosphere, in Island Arcs, Deep Sea Trenches and Back-Arc Basins, American Geophysical Union, Maurice Ewing Series, 1, edited by M . Talwani and W.C. Pitman, 243-258, 1977. Jiracek, G.R., E.P. Gustafson, and P.S. Mitchell, Magnetotelluric results opposing magma origin of crustal conductor in the Rio Grande rift, Tectonophysics, 94, 299-326, 1983. Jones, D.L., A . Cox, P. Coney, and M . Beck, The growth of western North America, Sci. Am., 247(5), 70-84, 1982. Jordan, T.H., and L . N . Frazer, Crustal and upper mantle structure from S phases, J. Geophys. Res., 80, 1504-1518, 1975. p  Kern, H., and A . Richter, Temperature derivatives of compressional and shear wave velocities in crustal and upper mantle rocks at 6 kbar confining pressure, / . Geophys., 49, 47-56, 1981. Kurtz, R.D., J.M. Delaurier, and J.C. Gupta, A magnetotelluric sounding across Vancouver Island detects the subducting Juan de Fuca plate, Nature, 321, 596-599, 1986. Kurtz, R.D., J.M. Delaurier, and J.C. Gupta, The electrical conductivity distribution beneath Vancouver Island: a region of active plate subduction, / . Geophys. Res., 95, 10,92910,946, 1990. Langston, C.A., Corvallis, Oregon, crustal and upper mantle receiver structure from teleseismic P and S waves, Bull. Seismol. Soc. Am., 67, 713-724, 1977'a. Langston, C.A., The effect of planar dipping structure on source and receiver responses for constant ray parameter, Bull. Seismol. Soc. Am., 67, 1029—1050, 1977b.  References  164  Langston, C.A., Structure under Mount Rainier, Washington, inferred from teleseismic body waves, / . Geophys. Res., 84, 4749-4762, 1979. Langston, C.A., Evidence for the subducting lithosphere under southern Vancouver Island and western Oregon from teleseismic P wave conversions, / . Geophys. Res., 86, 3857-3866, 1981. Langston, C.A., Scattering of teleseismic body waves under Pasadena, California, J. Geophys. Res., 94, 1935-1951, 1989. Lapp, D.B., R.S. Crosson, and T.J. Owens, P-waveform analysis for local subduction geometry south of Puget Sound, Washington, Pure Appl. Geophys., 133, 349-365, 1990. Lewis, T.J., W . H . Bentkowski, E.E. Davis, R.D. Hyndman, J.G. Souther, and J.A. Wright, Subduction of the Juan de Fuca Plate: thermal consequences, / . Geophys. Res., 93, 15,207-15,225, 1988. Ludwig, W.J., J.E. Nafe, and C.L. Drake, Seismic refraction, in The Sea, V.4, edited by A . E . Maxwell, New York, pp. 53-84, 1970. Lueschen, E., F. Wenzel, K.J. Sandmeier, D. Menges, T. Ruehl, M . Stiller, W. Janoth, F. Keller, W. Soeller, R. Thomas, A . Krohe, R. Stenger, K . Fuchs, H . Wilhelm, and G. Eisbacher, Near-vertical and wide-angle seismic surveys in the Black Forest, SW Germany, / . Geophys., 62, 1-30, 1987. Manghnani, M . H . , R. Ramananantoandro, and S.P. Clark, Jr., Compressional and shear wave velocities in granulite facies rocks and eclogites to 10 kbar, / . Geophys. Res., 79, 5427-5446, 1974. McMechan, G.A., and G.D. Spence, P-wave velocity structure of the Earth's crust beneath Vancouver Island, Can. J. Earth Sci., 20, 742-752, 1983. Monger, J.W.H., Georgia Basin: regional setting and adjacent Coast Mountains geology, British Columbia, in Current Research, Part A , Geological Survey of Canada, Paper 90-1F, 95-107, 1990. Monger, J.W.H., Georgia Basin Project: structural evolution of parts of south Insular and southwestern Coast Belt, British Columbia, in Current Research, Part A , Geological Survey of Canada, Paper 91-1A, 219-228, 1991. Monger, J., R. Price, and D. Templeman-Kluit, Tectonic accretion of two major metamorphic and plutonic welts in the Canadian Cordillera, Geology, 10, 70-75, 1982.  References  165  Monger, J., R. Clowes, D. Cowan, C. Potter, R. Price, and C. Yorath, Continent-ocean transitions in western North America between latitudes 46 and 56 degrees, in Phanerozoic Evolution of North American Continent-Ocean Transitions, Geol. Soc. A m . Centen Transect Vol. CTV-1, edited by R.C. Speed, in press, 1991. Moos, D., and M . D . Zoback, In situ studies of velocity in fractured crystalline rocks, J. Geophys. Res., 88, 2345-2358, 1983. Muller, J.E., Evolution of the Pacific margin, Vancouver Island, and adjacent regions, Can. J. Earth Sci., 14, 2062-2085, 1977. O'Connell, R J . , and B . Budiansky, Seismic velocities in dry and saturated cracked solids, / . Geophys. Res., 79, 5412-5426, 1974. Owens, T.J., Determination of crustal and upper mantle structure from analysis of broadband teleseismic P-waveforms, Ph. D. thesis, 145 pp., Univ. of Utah, Salt Lake City, Utah, 1984. Owens, T.J., and R.S. Crosson, Shallow structure effects on broadband teleseismic P waveforms, Bull. Seismol. Soc. Am., 78, 96-108, 1988. Owens, T.J., and G . Zandt, The response of the continental crust-mantle boundary observed on broadband teleseismic receiver functions, Geophys. Res. Lett., 12, 705-708, 1985. Owens, T.J., R.S. Crosson, and M . A . Hendrickson, Constraints on the subduction geometry beneath western Washington from broadband teleseismic waveform modeling, Bull. Seismol. Soc. Am., 78, 1319-1334, 1988. Owens, T.J., S.R. Taylor, and G . Zandt, Crustal structure at regional seismic test network stations determined from inversion of broadband teleseismic P waveforms, Bull. Seismol. Soc. Am., 77, 631-662, 1987. Owens, T.J., G . Zandt, and S.R. Taylor, Seismic evidence for an ancient rift beneath the Cumberland Plateau, Tennessee: a detailed analysis of broadband teleseismic P waveforms, J. Geophys. Res., 89, 7783-7795, 1984. Pennington, W.D., Role of shallow phase changes in the subduction of oceanic crust, Science, 220, 1045-1047, 1983. Phinney, R.A., Structure of the earth's crust from spectral behaviour of long-period body waves, / . Geophys. Res., 69, 2997-3017, 1964.  References  166  Riddihough, R.P., Gravity and structure of an active margin - British Columbia and Washington, Can. J. Earth Sci., 16, 350-363, 1979. Roddick, J.A., J.E. Muller, and A . V . Okulitch, Fraser River, British Columbia-Washington, 1:1,000,000, Geological Survey of Canada, Map 1386A, 1979. Rogers, G.C., Seismotectonics of British Columbia, Ph. D. thesis, 247 pp., Univ. of British Columbia, Vancouver, B.C., Canada, 1983. Rogers, G.C., A n assessment of the megathrust earthquake potential of the Cascadia subduction zone, Can. J. Earth Sci., 25, 844-852, 1988. Rogers, G.C., C. Spindler, and R.D. Hyndman, Seismicity along the Vancouver Island Litho-  probe Corridor, in Proc. of the Project Lithoprobe: Southern Canadian Cordillera Tra sect Workshop, Calgary, 3-4 March 1990, 166-169, 1990. Ruff, L., and H . Kanamori, Seismicity and the subduction process, Phys. Earth Planet. Inter., 23, 240-252, 1980. Schmeling, H . , Numerical models on the influence of partial melt on elastic, anelastic and electric properties of rocks. Part I, Elasticity and anelasticity, Phys. Earth Planet. Inter., 41, 34-57, 1985. Schmeling, H . , Numerical models on the influence of partial melt on elastic, anelastic and electric properties of rocks. Part n, Electrical conductivity, Phys. Earth Planet. Inter., 43, 123-136, 1986. Spence, G.D., Clowes, R . M . , and R . M . Ellis, Seismic structure across the active subduction zone of western Canada, / . Geophys. Res., 90, 6754-6772, 1985. Stanley, W.D., W.D. Mooney and G.S. Fuis, Deep crustal structure of the Cascade Range and surrounding regions from seismic refraction and magnetotelluric data, / . Geophys. Res., 95, 19,419-19,438, 1990. Sutherland Brown, A . , Tectonic history of the Insular Belt of British Columbia, Can. Inst. Mining Met., 8, 83-100, 1966. Taber, J.J., and S.W. Smith, Seismicity and focal mechanisms associated with the subduction of the Juan de Fuca plate beneath the Olympic Peninsula, Washington, Bull. Seismol. Soc. Am., 75, 237-249, 1985. Tseng, K . H . , A new model for the crust in the vicinity of Vancouver Island, M . Sc. thesis, 83 pp., Univ. of British Columbia, Vancouver, B.C., Canada, 1968.  References  167  Varsek, J.L., F. Cook, R. Clowes, E. Kanasewich, C. Spencer, J. Monger, M . Journeay, and R. Parrish, LITHOPROBE crustal reflection structure of the southern coast mountains, British Columbia, in Proc. of the Project Lithoprobe: Southern Canadian Cordille Transect Workshop, Calgary, 3 ^ March 1990, 139-140, 1990. Vera, E.E., J.C. Mutter, P. Buhl, J.A. Austin Jr., I.W.D. Dalziel, and P.L. Stoffa, Deep seismic structure of the southernmost Andes (abstract), EOS Trans. AGU, 70, 1314, 1989.  Waldron, D.A., R . M . Clowes, and D.J. White, Structure of a subducting oceanic plate off Western Canada, in Studies of laterally heterogeneous structures using seismic refr and reflection data, edited by A . G . Green, Geological Survey of Canada, Paper 89-13, 105-113, 1990. Weaver, C.S., and G.E. Baker, Geometry of the Juan de Fuca plate beneath Washington and Oregon from seismicity, Bull. Seismol. Soc. Am., 78, 264—275, 1988. White, W.R.H., and J.C. Savage, A seismic refraction and gravity study of the Earth's crust in British Columbia, Bull. Seismol. Soc. Am., 55, 463—486, 1965. Wickens, A.J., The upper mantle of southern British Columbia, Can. J. Earth Sci., 14, 11001115, 1977. Wong, I.G., and D.S. Chapman, Deep intraplate earthquakes in the western United States and their relationship to lithospheric temperatures, Bull. Seismol. Soc. Am., 80, 589-599, 1990. Yole, R.W., and E. Irving, Displacement of Vancouver Island: paleomagnetic evidence from the Karmutsen Formation, Can. J. Earth Sci., 17, 1210-1228, 1980. Yorath, C.J., R . M . Clowes, A . G . Green, A . Sutherland Brown, M.T. Brandon, N.W.D. Massey, C. Spencer, E.R. Kanasewich, and R.D. Hyndman, LITHOPROBE - Phase 1: southern Vancouver Island: preliminary analyses of reflection seismic profiles and surface geological studies, Geological Survey of Canada, Paper 85-1 A, pp. 543-554, 1985. Zandt, G., and T.J. Owens, Comparison of crustal velocity profiles determined by seismic refraction and teleseismic methods, Tectonophysics, 128, 155-161, 1986. Zelt, C.A., Seismic structure of the crust and upper mantle in the Peace River arch region, Ph. D. thesis, 223 pp., Univ. of British Columbia, Vancouver, B.C., Canada, 1989.  Appendix A Table A . l : Teleseisms recorded P = PGC-B, A = ALB-B, L = LAS, E = EGM No.  1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30  Date  Origin Time  D MY  HMS(UT)  Lat. °N  Long. °E  Z (km)  27 06 87 06 07 87 06 07 87 08 07 87 10 07 87 11 07 87 13 07 87 15 07 87 24 07 87 25 07 87 08 08 87 12 08 87 13 08 87 15 08 87 18 08 87 01 09 87 03 09 87 04 09 87 07 09 87 08 09 87 08 09 87 11 09 87 13 09 87 22 09 87 22 09 87 28 09 87 30 09 87 02 10 87 04 10 87 06 10 87  09 09 06.1 01 06 07.6 02 49 42.7 11 50 14.7 18 49 53.9 06 15 51.0 19 14 57.9 07 16 13.5 05 25 10.7 01 11 48.5 15 48 56.7 15 23 11.5 15 23 06.9 18 04 23.1 02 18 48.7 04 26 09.7 06 40 13.9 04 27 08.8 11 57 09.4 02 58 51.1 03 08 07.2 00 34 52.1 11 20 51.1 13 43 37.6 16 21 35.1 11 47 08.6 01 39 25.5 07 38 27.8 08 15 16.6 04 19 06.0  -14.1 -26.1 -14.1 -27.0 55.1 82.2 -15.3 17.5 56.3 60.1 -19.2 14.1 -17.9 -28.2 -5.5 -23.2 -58.9 49.3 -31.1 6.6 6.5 -22.3 14.2 1.0 1.1 -18.4 -18.2 27.3 10.7 -17.9  -76.1 -108.3 167.9 -108.2 165.5 -17.5 -70.1 -97.2 -153.6 -153.9 -70.1 -59.3 -70.9 -70.8 151.7 -66.6 158.5 -156.4 178.0 -82.4 -82.4 -68.4 -90.0 -78.1 -78.1 168.1 167.9 139.9 -86.0 -172.2  61 10 38 10 61 10 241 67 33 165 82 54 37 39 33 226 33 110 249 97 82 128 117 10 10 31 33 465 33 33  168  Region  5.9 Peru 6.2 Easter Island 5.9 Vanuatu Island 6.1 Easter Island 6.0 Kamchatka 5.4 Greenland 5.1 Peru 5.9 Mexico 5.5 Kodiak Island 5.0 South Alaska 6.4 Chile 5.7 Windward Island 6.1 Peru Chile 6.3 5.4 New Britain 5.8 Argentina 5.9 Macquarie Island 5.8 Kuril Island Kermadec Island 5.8 Panama 5.3 5.4 . Panama 5.4 Chile 5.0 Guatemala 6.1 Ecuador 5.9 Ecuador 5.7 Vanuatu Island 5.4 Vanuatu Island 5.6 Bonin Island 5.4 Costa Rica 6.7 Tonga  Stations  P P P P P P P P P P P P P P P P P,E P,E P,E P,E E E P.E E P P P E E E  169  Appendix A No.  31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68  Date  Origin Time  Lat.  DMY  HMS (UT)  °N  Long. °E  52.9 -19.6 -7.3 -6.3 52.6 -28.7 -17.0 12.2 -28.9 -24.8 -24.9 -17.7 50.7 -21.0 13.4 55.2 -17.2 10.4 -20.7 77.6 -17.2 -7.9 18.6 -26.6 11.4 -13.3 42.6 -15.4 -15.1 -10.8 13.5 27.0 25.0 -14.3 9.0 -17.5 -6.4 -17.6  160.0 -173.1 154.4 149.1 172.3 -62.9 -173.9 -87.2 . -177.5 -70.7 -70.5 -67.1 173.4 -69.9 124.6 167.4 -74.2 -60.5 -178.6 125.5 -72.5 157.6 145.8 -113.1 -86.0 -76.2 143.7 167.3 -173.5 165.2 -91.2 -110.9 -45.8 167.2 137.9 -71.8 148.9 -178.9  06 10 87 08 10 87 12 10 87 16 10 87 20 10 87 27 10 87 03 11 87 17 11 87 12 01 88 19 01 88 05 02 88 06 02 88 07 02 88 22 02 88 24 02 88 29 02 88 09 03 88 10 03 88 10 03 88 21 03 88 12 04 88 25 04 88 04 05 88 05 05 88 06 05 88 06 05 88 07 05 88 05 06 88 11 06 88 12 06 88 18 06 88 18 06 88 21 06 88 02 07 88 03 07 88 04 07 88 05 07 88 06 07 88  20 11 34.6 03 20 44.5 13 57 04.7 20 48 01.6 09 23 36.6 21 58 17.4 08 15 01.9 03 40 09.6 07 29 30.0 07 30 29.0 14 01 01.2 18 03 54.2 18 15 05.4 19 13 16.2 03 52 04.3 05 31 41.4 21 33 54.2 06 17 22.4 10 25 04.3 23 31 21.6 23 19 55.3 10 10 30.3 23 47 04.6 10 04 18.9 14 46 17.8 16 34 05.7 01 59 26.9 18 22 47.9 12 17 25.8 13 39 39.8 18 42 02.7 22 49 43.2 06 26 14.6 10 01 30.4 11 43 15.0 13 54 21.0 20 32 03.9 01 10 53.0  Z  Region  Stations  (km)  33 33 25 48 33 611 101 98 33 18 33 282 33 65 33 33 33 48 600 10 33 33 142 33 105 51 78 114 33 33 55 10 10 156 33 68 33 546  6.2 6.2 6.3 5.9 5.4 5.9 5.9 5.6 5.8 6.4 6.2 6.0 6.3 6.2 6.0 6.1 6.0 6.2 6.0 6.1 6.1 6.2 5.8 6.2 5.5 5.9 6.1 5.7 6.0 5.7 5.6 5.7 5.9 5.8 5.8 5.8 6.2 5.5  Kamchatka Tonga Islands Solomon Island New Britain Aleutian Islands Argentina Tonga Islands Nicaragua Kermadec Island Chile Chile Bolivia Aleutian Islands Chile Philippines Kamchatka Peru Trinidad Fiji Laptev Sea Peru Solomon Island Mariana Islands Easter Island Nicaragua Peru Japan Vanuatu Island Tonga Islands Santa Cruz Guatemala Gulf of Cai. Mid-Atlantic Pango Pango Caroline Island Peru New Britain Fiji Islands  E E E E P P P A,L,E L,E A,L,E A,L A,L A,L A,L,E A,L A,L E A,L E L A,L,E A,L,E A,L,E A,L,E A,L,E A,L,E A,L,E A,L,E A,L,E A.L.E A,L,E A.L.E A,L A,L,E A,L,E A,L,E A,L,E A,L,E  Appendix A No.  69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107  170  Date  Origin Time  DMY  H M S (UT)  06 07 88 19 07 88 20 07 88 23 07 88 23 07 88 25 07 88 27 07 88 28 07 88 30 07 88 06 08 88 06 08 88 10 08 88 10 08 88 10 08 88 14 08 88 14 08 88 17 08 88 20 08 88 27 08 88 27 08 88 05 09 88 07 09 88 13 09 88 14 09 88 15 09 88 08 10 88 10 10 88 03 11 88 07 11 88 05 12 88 07 12 88 08 12 88 04 02 89 10 02 89 14 02 89 25 02 89 28 02 89 11 03 89 05 04 89  15 54 01 00 23 15 14 25 15 17 06 46 21 55 17 12 21 07 00 36 09 03 04 38 11 46 13 11 10 56 17 53 11 34 23 09 01 25 16 30 06 13 11 53 00 58 22 14 18 48 04 46 18 20 14 47 03 50 16 05 07 41 12 58 19 24 11 15 06 20 11 26 13 01 05 04 23 47  22.0 20.5 34.8 39.0 11.3 07.6 12.3 32.7 21.4 26.9 21.7 26.0 46.5 09.1 57.5 09.4 52.4 10.3 17.7 17.6 18.7 25.4 45.9 07.6 03.2 24.4 30.0 10.6 02.8 31.4 24.2 59.8 10.9 22.8 26.9 37.7 57.9 59.7 48.3  Lat. °N  Long. °E  Z (km)  41.8 -19.4 23.9 -22.2 -6.5 -6.2 -13.1 -22.3 44.7 25.1 36.5 -10.3 -28.2 -14.9 54.6 -27.4 -26.9 26.7 11.3 -15.8 18.5 30.3 29.8 -23.4 -1.4 -18.7 -28.3 13.8 -22.2 -15.3 40.9 6.9 5.9 2.3 -10.4 -29.8 -23.1 -17.7 -21.1  144.2 -175.2 121.6 174.9 152.8. 133.8 167.0 -65.8 149.9 95.2 71.1 160.8 -112.8 167.5 152.7 -70.9 -70.9 86.6 141.5 -172.1 -70.4 137.4 138.4 -67.9 -77.9 -172.4 -177.7 -90.6 175.0 -173.4 44.3 -82.7 -82.6 126.7 161.3 -177.9 -61.6 -174.8 -69.2  54 142 33 33 39 38 196 ' 285 64 115 194 33 10 33 644 38 39 70 33 33 33 499 447 124 189 33 57 68 33 33 10 10 33 33 78 47 575 178 114  b  m  5.8 6.0 5.9 5.8 6.4 6.3 6.0 5.8 6.2 6.8 6.0 6.0 6.0 6.1 5.5 6.0 5.6 6.5 5.2 6.0 5.5 6.0 6.1 5.8 5.8 6.7 6.5 5.4 5.6 6.1 6.2 5.6 5.9 6.1 6.1 6.4 5.5 6.3 5.8  Region  Japan Tonga Islands Taiwan Loyalty Island New Britain Aroe Island Vanuatu Island Argentina Kuril Island Burma Afghanistan Solomon Island Easter Island Vanuatu Island Japan Chile Chile India Caroline Island Samoa Island Dominica Japan Japan Chile Ecuador Tonga Islands Kermadec Island Guatemala Loyalty Island Tonga Islands Armenia Panama Panama Molucca Solomon Island Kermadec Island Paraguay Tonga Islands Chile  Stations  A,L,E A,L,E A,L,E A,L,E A,L,E A,L,E A,L,E L A A,E A,E A,L,E A,L,E A,L,E A,L,E A,L,E E A,L,E A,E A,L,E A,L A,L,E A,E A,E A,L,E A,L,E A.L.E L,E A,L,E E A,L,E A,L,E A,L A,L,E A,L A,L,E A,L,E A,L L  Appendix A  No.  171  Date DMY  108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145  06 04 89 09 04 89 11 04 89 15 04 89 20 04 89 20 04 89 25 04 89 25 04 89 27 04 89 30 04 89 03 05 89 14 05 89 23 05 89 24 05 89 24 05 89 08 06 89 09 06 89 16 06 89 25 06 89 26 06 89 03 07 89 06 07 89 22 07 89 03 08 89 14 08 89 21 08 89 21 08 89 29 08 89 30 08 89 09 09 89 16 09 89 16 09 89 17 09 89 20 09 89 22 09 89 25 09 89 07 10 89 09 10 89  Origin Time H M S (UT) 08 05 40.0 05 07 50.3 03 56 38.8 20 34 11.5 08 08 51.5 22 59 54.9 02 13 24.5 14 29 01.1 02 20 05.8 08 22 54.1 05 53 03.0 00 59 49.7 10 54 46.9 13 31 15.7 15 43 34.0 09 51 56.7 12 19 35.5 23 42 36.5 20 37 33.0 03 27 03.9 17 09 55.8 17 25 30.0 05 02 12.1 11 31 24.2 17 51 18.8 18 25 41.8 23 12 43.4 04 16 24.7 11 38 13.1 01 40 36.9 02 05 06.2 23 20 54.0 00 53 37.5 13 19 32.1 02 25 53.5 14 17 47.3 15 48 30.0 18 01 10.0  Lat. °N  Long. °E  -19.6 51.5 49.4 30.0 -9.2 57.1 30.0 16.9 30.7 11.0 30.0 -30.4 -52.6 56.2 56.3 -19.5 71.4 31.8 1.1 19.4 51.6 -16.6 2.3 23.2 -18.9 -4.1 24.2 18.1 55.6 2.6 40.3 16.6 40.2 51.2 31.6 -20.3 51.1 52.1  169.5 -178.4 159.3 99.3 -79.0 122.0 99.5 -99.4 140.7 -68.3 99.6 -178.3 160.4 164.4 164.1 -173.9 -4.7 138.1 -79.6 -155.1 -175.2 -177.5 128.2 122.1 176.4 154.4 122.5 -105.6 161.5 -79.7 51.6 -93.7 51.8 178.8 102.5 169.3 -179.1 172.1  Z  b  m  Region  Stations  (km)  33 32 33 33 64 33 33 23 93 18 33 33 10 33 33 24 10 373 19 9 33 33 146 33 114 500 59 33 80 10 33 110 33 33 33 33 33 33  6.3 5.2 6.3 6.2 5.7 6.0 6.1 6.4 6.0 5.7 6.2 5.9 6.2 5.8 5.6 5.6 5.5 5.9 5.8 5.9 5.6 5.3 6.4 6.0 5.7 6.0 5.3 5.5 5.9 6.0 6.2 5.9 6.0 5.5 6.0 6.0 6.3 5.8  Vanuatu Island South Alaska Kuril Island China Peru Russia China Mexico Japan Venezuela China Kermadec Island Macquarie Island Kamchatka Kamchatka Tonga Islands Greenland Japan Ecuador Hawaii Aleutian Islands Fiji Islands Halmahera Taiwan Fiji Islands Solomon Island Taiwan Mexico Kamchatka Panama Caspian Sea Mexico. Caspian Sea Aleutian Islands China Vanuatu Island South Alaska Aleutian Islands  A,L A A,L A,L A,L A,L A,L A,L A,L A,L A,L A A A A E A A,E A,E A,E A A A,L,E A,E A,E A A A A A A,E A,E A,E A,E A,E A,E A A  Appendix B The following 3 tables provide the final earth models developed for A L B - B , L A S , and E G M respectively. Note that the values given for Z, dip angle and dip direction refer to the top of the interface.  Layer No. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15  Vp km/s 4.00 6.06 6.50 6.10 7.10 6.29 6.80 7.40 6.59 7.40 6.46 6.86 7.08 7.30 8.10  Vs km/s 2.31 3.50 3.76 3.53 4.10 3.64 3.93 4.28 3.28 4.28 3.73 3.97 4.09 4.22 4.68  P g/cm 2.43 2.69 2.80 2.70 2.98 2.75 2.88 3.08 2.72 3.08 2.79 2.90 2.97 3.04 3.33  3  Z km 0.0 0.5 2.5 6.7 10.8 20.1 23.1 26.6 36.7 41.2 47.0 49.0 49.9 52.1 53.0  Dip Angle deg. 00. 15. 00. 25. 25. 15. 15. 15. 7. 7. 15. 15. 15. 15. 15.  Table B . l : A L B - B final model  172  Dip Dir. deg. 000. 155. 000. 320. 320. 320. 320. 320. 280. 280. 300. 300. 300. 300. 300.  Appendix B  Layer No. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16  Vp km/s 4.00 6.00 6.70 6.95 6.00 6.35 6.70 6.96 8.00 6.80 7.20 7.60 8.10 7.20 7.40 8.10  Vs km/s 2.31 3.47 3.87 4.02 3.47 3.67 3.87 4.02 4.62 3.93 4.16 4.39 4.68 4.16 4.28 4.68  P g/cm 2.43 2.68 2.85 2.93 2.68 2.76 2.85 2.93 3.30 2.88 3.01 3.15 3.33 3.01 3.08 3.33  3  Z km 0.0 0.5 6.3 11.7 26.0 30.1 31.1 32.7 36.2 45.2 48.2 50.2 52.2 59.7 63.2 64.9  Dip Angle deg 00. 00. 15. 00. 00. 00. 00. 00. 00. 10. 10. 10. 10. 22. 22. 22.  Table B.2: L A S final model  Dip Dir. deg 000. 000. 320. 000. 000. 000. 000. 000. 000. 300. 300. 300. 300. 300. 300. 300.  174  Appendix B  Layer No. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17  Vp km/s 5.30 5.90 6.70 6.00 6.80 7.20 6.10 6.60 7.20 8.05 6.80 8.20 6.80 8.25 7.00 7.40 8.25  Vs km/s 3.06 3.41 3.87 3.47 3.93 4.16 3.53 3.82 4.16 4.65 3.93 4.74 3.93 4.77 4.05 4.28 4.77  P g/cm 2.56 2.66 2.85 2.68 2.88 3.01 2.70 2.83 '3.01 3.32 2.88 3.37 2.88 3.39 2.94 3.08 3.39  3  Z km 0.0 1.0 6.5 10.4 15.4 19.4 26.4 29.4 31.4 36.4 43.8 46.8 53.4 57.4 69.5 73.0 74.7  Dip Angle deg 00. 00. 15. 00. 00. 00. 00. 00. 00. 00. 15. 15. 15. 15. 30. 30. 30.  Table B.3: E G M final model  Dip Dir. deg 000. 000. 100. 000. 000. 000. 000. 000. 000. 000. 280. 280. 300. 300. 300. 300. 300.  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0052621/manifest

Comment

Related Items