UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The crustal structure of the northern Juan de Fuca plate from multichannel seismic reflection data Hasselgren, Elizabeth 1991

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

Item Metadata


831-UBC_1991_A6_7 H38.pdf [ 38.47MB ]
JSON: 831-1.0052831.json
JSON-LD: 831-1.0052831-ld.json
RDF/XML (Pretty): 831-1.0052831-rdf.xml
RDF/JSON: 831-1.0052831-rdf.json
Turtle: 831-1.0052831-turtle.txt
N-Triples: 831-1.0052831-rdf-ntriples.txt
Original Record: 831-1.0052831-source.json
Full Text

Full Text

T H E CRUSTAL STRUCTURE OF T H E NORTHERN JUAN DE FUCA PLATE FROM MULTICHANNEL SEISMIC REFLECTION DATA By Elizabeth Hasselgren B. Sc.(Hons.) Geology/Physics, Concordia University  A THESIS SUBMITTED IN PARTIAL F U L F I L L M E N T O F T H E REQUIREMENTS F O R T H E D E G R E E O F M A S T E R O F SCIENCE  in T H E FACULTY O F G R A D U A T E STUDIES GEOPHYSICS A N D A S T R O N O M Y  We accept this thesis as conforming to the required standard  T H E UNIVERSITY  O F BRITISH COLUMBIA  April 1991 © Elizabeth Hasselgren, 1991  In  presenting this  degree at the  thesis in  University of  partial  fulfilment  of  of  department  this thesis for or  by  his  or  requirements  British Columbia, I agree that the  freely available for reference and study. I further copying  the  representatives.  an advanced  Library shall make  it  agree that permission for extensive  scholarly purposes may be her  for  It  is  granted  by the  understood  that  head of copying  my or  publication of this thesis for financial gain shall not be allowed without my written permission.  Department The University of British Columbia Vancouver, Canada  DE-6 (2/88)  ABSTRACT  The crustal structure of a young (<10 My) ocean basin is imaged by two multichannel seismic reflection lines comprising 230 km recorded over the central part of the northern Juan de Fuca plate off western Canada. The more northerly line ties previously interpreted deep seismic reflection lines across the Juan de Fuca ridge and the Cascadia subduction zone; the southern line ties with another interpreted line across the subduction zone. Both lines trend obliquely to the spreading direction. A marine refraction profile crossing the eastern end of the lines provides velocity constraints. The processing sequence applied to the data includes a prestack inside-trace mute of CMP gathers to reduce noise levels on the deep data, CMP stack, post-stack dip filtering, f-k migration and bandpass. Coherency-filtered stacks are helpful in tracing weaker reflectors. The stacked sections reveal a horizontally layered sedimentary sequence overlying a rugged and prominent basement reflector dipping slightly landward. A strong, fairly continuous reflection from the base of the crust at about 2 s two-way-time below the basement surface generally mimics the basement topography and shows the characteristic doubling and tripling of reflections seen in other similar surveys. Although in general the crust appears acoustically transparent, weaker, discontinuous intracrustal reflectors are observed over 40 km at the eastern end of the northern line, and are interpreted to arise from the oceanic Layer 3A/3B and Layer 2/3 boundaries. The impersistence of these reflectors is an indication of the complexity of the processes producing intracrustal reflectivity, and an indication of the lateral variability of crustal formation. Pseudofault traces of propagating rifts are crossed at three different locations on the two lines, the first MCS crossings of such structures. Crust associated with the pseudofault traces is related to both subhorizontal and dipping subcrustal events which are interpreted as zones of crustal thickening or underplating. Although the crustal thickness elsewhere on the lines varies by only about 10%, crust associated with the pseudofaults is as much as about 25% thicker ii  than average, suggesting that magma supply at transform-type offsets may at times be large. A small seamount discovered on the southern line may result from the excessive magma production at the ridge postulated at propagating rift zones.  iii  Table of Contents  ABSTRACT  ii  List of Tables  vi  List of Figures  vii  Acknowledgement 1  2  ix  INTRODUCTION  1  1.1  Study overview  1  1.2  Outline of oceanic crustal structure  .  2  1.2.1  Velocity structure of oceanic crust  2  1.2.2  Creation and evolution of oceanic crust  6  1.3  Previous MCS surveys  14  1.4  Tectonic setting of study area  16  1.5  Summary  21  ACQUISITION AND PROCESSING PARAMETERS  22  2.1  Acquisition  22  2.2  Data quality  24  2.3  Processing techniques  32  2.3.1  Processing systems  32  2.3.2  Prestack processing  32  2.3.3  Poststack processing  39  2.3.4  Velocity analysis  39  iv  2.3.5 2.4 3  4  Final processing stream  47  Summary  • • • •  47  OCEANIC CRUSTAL STRUCTURE OBSERVED ON MCS DATA  49  3.1  Introduction  49  3.2  Major reflectors  58  3.2.1  Sedimentary sequence . . .  58  3.2.2  Top of oceanic basement  60  3.2.3  Intracrustal reflectors  64  3.2.4  Basal reflection  73  .  3.3  Seamount  83  3.4  Crustal structure of pseudofault traces of propagating rifts (PR)  84  3.5  Summary  90  DISCUSSION AND CONCLUSIONS  95  4.1  Introduction  95  4.2  General crustal characteristics  95  4.2.1  Basal reflection  97  4.2.2  Intracrustal reflectivity  97  4.3  ;  Propagating rifts (PR)  100  4.3.1  Characteristics  100  4.3.2  Magma sources  103  4.3.3  Location of crustal underplating or thickening  107  4.3.4  Crustal structure of transform fault zones  108  4.4  Conclusions  109  4.5  Future work  110  REFERENCES  112  v  List of Tables  1.1  Oceanic crustal velocity structure  4  1.2  Oceanic crustal reflections  17  2.1  Acquisition parameters  25  3.1  Reflection coefficient estimates  82  vi  List of Figures  1.1  Tectonic setting of study area  3  1.2  Average oceanic crustal velocity structure  5  1.3  Age dependence of oceanic crustal velocity .  5  1.4  Comparison of oceanic crustal seismic velocities with rock type . . . . . . .  7  1.5  Large-scale MOR structure. . .  9  1.6  Along-axis discontinuities in MOR structure  12  1.7  Model of magmatic segmentation of MOR  13  1.8  MCS data examples  15  1.9  Magnetic anomaly pattern of the Juan de Fuca plate  19  1.10 Comparison of stable MOR with propagating rift  20  2.1  Location map of MCS lines  23  2.2  Shotpoint locations of MCS lines  26  2.3  Shot gather and stack from 85-09  27  2.4  Shot gather and stack from 85-07  28  2.5  CMP gather from line 85-07  29  2.6  Frequency content of data  31  2.7  Prestack f-k  2.8  Prestack running mix  35  2.9  CMP mute pattern  37  filter  2.10 Prestack energy balance  34  -  2.11 Post stack f-k filter and bandpass  38 40  vii  2.12 Post stack running mix and f-k migration  41  2.13 Semblance velocity analysis  43  2.14 Constant velocity analysis of CMP gather  44  2.15 Constant velocity stacks  45  2.16 Final processing stream .  48  3.1  CMP stack of line 85-07  51  3.2  CMP stack of line 85-09  53  3.3  Coherencyfilteredstack and line drawing of line 85-07  55  3.4  Coherencyfilteredstack and line drawing of line 85-09  57  3.5  Migration of sedimentary section: eastern end of MCS lines  62  3.6  True amplitude prestack depth migration: 85-07, SP 297-381  63  3.7  Interval velocities derived from stacking velocities: eastern end of 85-07 . .  69  3.8  Migration and line drawing interpretation: eastern end of 85-07  70  3.9  Refraction velocity modelfromeastern ends of MCS lines . . . . . . . . . .  72  3.10 Dipping intracrustal reflectors, 85-09  .  74  3.11 Dipping reflector at base of crust, 85-07  77  3.12 Velocity-density relations  79  3.13 Data used for RC estimates, line 85-09  80  3.14 Seamount: fine 85-09  85  3.15 Seamount caldera  86  3.16 Closeup of PR 7, line 85-09  93  3.17 Closeup of PR4, line 85-09 ,  94  4.1  Crustal model of multiple intrusions  99  4.2  Relative crustal volume generated by Cobb Hotspot  104  4.3  Paleopositions of Cobb Hotspot and propagator traces  105  viii  Acknowledgement  First and foremost, I would like to acknowledge the support and guidance provided by Ron Clowes, who supervised this research. Andrew Calvert provided invaluable assistance regarding all aspects of MCS data processing and interpretation. His work on the seamount and associated subcrustal reflectivity led to the interpretation of substantial magma supply to propagating rift tips presented here; he also provided thefiguresof the seamount and detail of PR4. Kristin Rohr of the Pacific Geoscience Center, Geological Survey of Canada, provided the demultiplexed data; we had numerous conversations about processing problems and the tectonic significance of the data. Thanks to the staff at the Lithoprobe Seismic Processing Facility at the University of Calgary for their assistance. My daughter Katherine has been remarkably patient and understanding, and has contributed cut-and-paste skills in a number of crucial situations. Finally, I would like to thank everyone who made my stay at UBC an enjoyable one. During the course of this research, the author was the recipient of an NSERC Postgraduate Award. Additional funds were provided through an NSERC Individual Operating Grant to Dr. Clowes. The Geophysics Research Processing Facility in the department is partially supported by an NSERC Infrastructure Grant to Dr. Clowes and others.  ix  Chapter 1  INTRODUCTION  1.1 Study overview  "  This study involves the processing and interpretation of two subparallel multichannel seismic reflection (MCS) profiles on the northern Juan de Fuca plate, a small oceanic plate located in the NE Pacific (Fig. 1.1). Deep crustal MCS experiments in ocean basins have been comparatively rare, mainly because of their high cost, but have added an important new dimension to our understanding of oceanic crustal processes because of their relatively high resolution compared with other subsurface remote sensing techniques. The Juan de Fuca plate, located between the much larger Pacific and North American plates, is formed at the intermediate spreading rate Juan de Fuca ridge and subducted beneath the North American plate at the Cascadia subduction zone. The recent tectonic history of the Juan de Fuca plate has been eventful, with breakup of the plate both to the north and south, as well as numerous episodes of ridge reorganization. The primary objectives of this work were to discern within the seismic sections any evidence for age-dependence in the structure of young (i.e., <10 My old) oceanic crust formed at an intermediate rate spreading center, for changes in crustal structure along-axis as expressed by differences along isochrons between the two lines, and also for any modifications to crustal structure due to the recent tectonic activity in the area.  1  Chapter 1.2  The  1.  INTRODUCTION  2  Outline of oceanic crustal structure  evolution of our  understanding of oceanic crustal structure and  processes has progressed  at the pace of improvements in deep water surveying and sampling techniques. Initial concepts of an essentially one dimensional oceanic crust gleaned from the velocity models of widely spaced seismic refraction surveys have given way to an appreciation of the variability in crustal structure as more detailed studies have become feasible. More systematic geophysical and geological surveys of the oceans have resulted in the concept of 'spreading cells' - focii of magma upwelling and crustal creation separated by transform faults. The advent of sophisticated bottom sounding techniques and multichannel seismic reflection surveys in the deep ocean together with detailed petrological studies involving systematic sampling of crustal materials has led to the viewpoint that crustal accretion processes may be differentiated on a much smaller scale than previously thought. Along-strike crustal discontinuities on a scale of 20-100 km may play an important role, perhaps delineating boundaries of distinct geochemical cells (Langmuir et al., 1986; MacDonald et al., 1988).  1.2.1  The  Velocity structure of oceanic crust  main source of geophysical information regarding the nature of the oceanic crust has  been the seismic refraction method, a relatively fast and  inexpensive survey technique which  provides direct velocity information. Velocity models established with this method were initially characterised by a two-layered oceanic crust (excluding the sedimentary layer) as shown in Table 1.1 (Raitt, 1963; Ewing and  Houtz, 1979). This layered crustal structure is in  part an artifact of survey methods: in more recent surveys, smaller shot and  receiver spacing  has resolved a number of thinner sublayers within layers 2 and 3. Modelling of amplitudes as well as inversion of the traveltime data has resulted in a velocity structure characterised by a series of velocity gradients rather than a few thick homogeneous velocity layers (Fig. 1.2)  Figure 1.1: Tectonic setting of the Juan de Fuca plate. Arrows indicate relative directions of plate motion. Seismic lines interpreted in this study are numbered.  Chapter  1.  INTRODUCTION  4  Table 1.1: Oceanic crustal velocity structure  Interval Layer 1 (sedimentary) Layer 2 Layer 3 Layer 4 (mantle) 1  Velocity  (kms )  about 2 5.07 ± 0.63 6.69 ± 0.26 8.13 ± 0.24  l  1  Thickness  (km)  variable 1.71 ± 0 . 7 5 4.86 ± 1.42  Data from Raitt (1963); Layers 2 and 3 comprise the oceanic crust; in this thesis, the term crust is used to refer to Layers 2 and 3 only.  (Ewing and Houtz, 1979; White, 1984). Also, the hundreds of surveys conducted reveal some systematic differences in oceanic crustal structure; e.g., fracture zones may have thinner than normal crust (Purdy and Ewing, 1986), although this is not the only possible interpretation of the data (Karson and Elthon, 1989). Crustal velocities also increase with age, particularly those of layer 2A, from 2-4 km/s in young crust, to about 5 km/s in older crust (Fig. 1.3) (Ewing and Houtz, 1979; Purdy, 1987). Information regarding the actual composition of the velocity model layers must come from other sources, since the correlation between compressional velocity and rock type is not unique. Dredging and deep sea drilling have provided samples of the rock types which make up the oceanic crust; drilling and down-hole logging have been particularly useful in providing information about both the composition and physical properties of Layer 2 and  uppermost Layer 3 (e.g., Becker et al., 1989; Kirby et al., 1988). However, dredging  provides little information about the arrangement of rock types within the crust, and drilling provides a very spatially limited sample. Ophiolites, considered to be ancient slices of oceanic crust tectonically emplaced upon the continents, provide a larger spatial and depth sample, although the uppermost crustal sublayers tend to be poorly preserved or nonexistent (Ewing and Houtz, 1979). The layered ophiolite sequence of rock types, ranging downward  5  Chapter 1. INTRODUCTION V (km/s) p  6  8  10  Figure 1.2: Average velocity structure of Mesozoic oceanic crust awayfromfracturezones and spreading centers (after Purdy, 1983).  Figure 1.3: Stacked compilations (by crustal age) of seismic velocity versus depth profiles (from White et al., 1984).  Chapter 1.  6  INTRODUCTION  from extrusive basaltic pillows and flows, through sheeted diabase dikes, gabbros and layered ultramafic rocks, furnishes a correspondence between seismic velocity and rock type (Fig. 1.4). Drilling has confirmed this correspondence for the upper crust: DSDP Hole 504B in the Panama basin has drilled interlayered pillow and flow basalts overlying diabase dikes (Becker et al., 1989). ODP Site 835 on the SW Indian Ridge has drilled gabbroic rocks which have apparently been tectonically emplaced at the surface and have velocities similar to those measured in ophiolites, i.e., about 6.5 km/s (Kirby et al., 1988; Collins, 1989). Details of the postulated genesis of oceanic crust follow.  1.2.2  Creation and evolution of oceanic crust  Seismic refraction studies indicate that the oceanic crust away from fracture zones and spreading centers is broadly uniform. However, detailed geophysical and geological surveys of mid-ocean ridges (MOR)  indicate that crustal formation processes at spreading centers are  quite variable, both temporally and along-axis. In certain cases, this discrepancy may be explained by age-dependent processes which contribute to increasing uniformity of the crust as it moves away from the ridge, e.g., the increase in velocity in uppermost Layer 2, considered to reflect the infilling of porosity by metamorphic alteration minerals (Purdy, 1987). In most cases, however, this contrast is probably due to the scale of sampling inherent in the surveys. Along the spreading ridge, very detailed surveys have been carried out, in an effort to discover the exact mechanisms of crustal formation; in the deep ocean basins, survey techniques have been much less detailed, and perhaps the crustal variability which does exist on afinerscale has gone undetected. Although the gross morphology of MOR  is correlated with spreading rate, fine-scale  structure and tectonics are largely independent of this parameter (Sempere and MacDonald, 1987). For all spreading rates, a 'plate boundary zone' in which crustal accretion and  Chapter 1.  INTRODUCTION  SEISMIC REFRACTION LAYER  ORIGIN  2A/2B  LITHOLOGY PILLOW BASALTS  2C  3A  1  SHEETED DIKES  MAGMATIC SECTION  ISOTROPIC GABBROIC ROCKS IR LAYERED CUMULATE  3B  INTERLAYERED GABBROIC & ULTRAMAFIC ROCKS  Moho  ULTRAMAFIC CUMULATES MANTLE  RESIDUE  HARZBURGITE TECTONITES  Figure 1.4: Generalized crustal section showingfirst-orderstructural and penological divisions present in ophiolites compared with seismic refraction model. Also shown are horizons referred to later in the text: PM-'petrologic Moho'; SM-'seismic Moho'; MR-'Moho reflections; and IR-'intracrustal reflections'. Horizontal lines in columnar section represent horizons of interpreted intracrustal reflections (modified from Karson and Elthon, 1989).  Chapter  1.  INTRODUCTION  8  deformation take place occupies a narrow strip up to 30 km wide parallel to the spreading center. Beyond this the plate is essentially rigid: only relatively minor seismicity consistent with relaxation of thermally induced stresses occurs (Bratt et al., 1985). In general, the slower the spreading rate, the larger the scale of discontinuity at MOR:  at slow spreading  ridges, volcanic edifices are more pervasively dissected by larger throw faults; at fast rate spreading centers volcanism is more continuous, with smaller fault-generated discontinuities (Fig.  1.5). Studies of magnetic transition widths for intermediate to fast rate spreading centers indi-  cate that the width of the 'crustal accretion zone' (that segment of the crust in which at least 95% of the material is of Holocene age) is about 2 km in the plutonic layer and  about 1 km in  the extrusives (the 'neovolcanic zone') (MacDonald, 1982). A body of geophysical and geological evidence suggests that crustal rocks are formed by differentiation of mantle-derived parent magmas in a crustal magma chamber, rather than by eruption directly from the mantle. Seismic refraction surveys have identified low velocity zones, probably associated with partially molten rock, beneath the spreading center axis on fast and intermediate rate spreading ridges (e.g. Orcutt et al., 1975). Recent 3-D tomographic inversions of refraction data on the East Pacific Rise correlate an axial low velocity zone with a high amplitude phase-reversed reflector imaged by a previous multichannel seismic reflection survey (Toomey et al., 1989; Derrick et al., 1987). These data are consistent with the presence of a magma chamber of a maximum 3-4 km total width at 1-2 km below the sea bottom. Further analysis of the diffraction hyperbolae at the edges of the proposed magma chamber indicate that it may be only 800-1200 m wide and 10-50 m thick (Kent et al. 1990). No such evidence has yet been found on the slowly spreading mid-Atlantic ridge; however, petrological data seem to require at least a temporary magma chamber to account for the varying degrees of fractionation of basalts at MOR  (MacDonald, 1982). Morton et al. (1987) interpreted a shallow  reflection about 2.5 km deep and 1-2 km wide from the intermediate spreading rate southern  Figure 1.5: Comparison of the plate boundary zone at different spreading rates. The central volcano is highly discontinuous at slow rates, moderately continuous at intermediate rates, and almost perfecdy continuous at fast rates. Labels V and F indicate the limits of the neovolcanic and crustal Assuring zones (from MacDonald, 1982).  VO  Chapter  1.  INTRODUCTION  10  Juan de Fuca ridge as the top of a magma chamber. However, at the Endeavor segment of the same ridge, about 300 km north, White and Clowes (1990) found no evidence for a strong low velocity zone from a tomographic inversion of refraction data, even though an along-axis topographic bulge and hydrothermal vent fluids of 400° C have been observed at this location. The crust undergoes several stages of evolution perpendicular to the spreading axis (Fig. 1.5). The neovolcanic zone is comprised of fresh glassyflowswith sparse sediment cover. Central volcanoes, built by cycles of eruption of sheetflowsfollowed by extrusion of pillow basalts to the surface through lava tubes, are found in the axial zone. At intermediate rate spreading centers, the central volcanoes are often displaced parallel to the ridge axis by en echelon faults, suggesting periodic freezing of the magma chamber, with a frequency of eruptive cycles estimated at about one every 300-600 years. At the edges of the neovolcanic zone, the crust is pervasively fractured byfissuresdue to the tensional stresses associated with the acceleration of the crust from zero at the axis to the full spreading rate at the plate boundary zone. Thefissuresallow circulation of cold water into the crust and are probably the cause of the seismically defined low velocity layer 2A (2-4 km/s). As thefissuresare filled with sediments and hydrothermally derived metamorphic minerals, the P-wave velocity increases to about 5.5 km/s, resulting in the apparent thinning of layer 2A with increasing age of the plate (MacDonald, 1982; Purdy, 1987). An axial reflector over the Endeavor segment of the Juan de Fuca ridge has been interpreted as the possible lower boundary of the hydrothermally cooled crust (Rohr et al., 1988). Active normal faulting of the crustalfissuresin a zone of crustal extension begins 1-4 km from the axis and continues to about 10 km. The offset on these faults at intermediate rate spreading centers is about 50 m; fault dips are generally vertical at the top, indicating tensional failure, but change on the larger throw faults to 50-60° about 20-100 m in the crust, indicating a transition to shear failure (MacDonald, 1982). Dipping crustally-penetrating reflectors in  Chapter 1. INTRODUCTION  11  the Atlantic have been interpreted as possible seismic expressions of this phenomenon (White et al., 1990; NAT Study Group, 1985). Crustal extension due to faulting is about 5-10%. Beyond this zone the plate is essentially rigid (MacDonald, 1982). Along-axis discontinuities in MOR appear to be related to fundamental partitions in the crustal formation process, correlated with topographic and petrologic variations along strike (Fig. 1.6).  Transform faults are large offset boundaries (20-50 km or more, i.e.,  large enough that the lithosphere along the plate boundary behaves rigidly) which persist for millions of years or more, and divide MOR into distinct tectonic and petrologic units ('spreading cells') thought to derive from different mantle sources. Propagatingriftsoccur when one of theriftsoverruns the transform boundary and begins spreading in the older crust. Shorter wavelength segments (10-300 km), defined by smaller offsets of the spreading center from 0.5-20 km have been hypothesized as the boundaries of distinct geochemical cells (Langmuir et al., 1986). MCS data over the East Pacific Rise show along-axis disruption of the inferred axial magma chamber at larger-scale discontinuities (Derrick et al., 1987), but geochemical evidence appears to require magma chamber interruptions at even smaller scales. MacDonald et al. (1988) have proposed a model invoking along-axis variations in magmatic budget that may solve this dilemma. They envisage a spreading center fed discontinuously, with the smallest order of segmentation resulting from along-axis waning of magma delivery, explaining the observed highly fractionated chemistry of the crust at these locations (Fig. 1.7). Mutter et al. (1989) interpret dipping structures which occur most commonly on isochron-parallel MCS lines as the bounds of crustal accretion cells. Oceanic crustal formation is thus increasingly viewed as a mree-dimensional process, compared with previous views of along-axis homogeneity in the accretion process.  I20°W  110°.  100°  Figure 1.6: Along-axis discontinuities in mid-ocean ridge structure: a) Map view of section of the East Pacific Rise (EPR) showing segmentation of the ridge by transform faults ('Rivera', 'Orozco', etc.) and smaller offset discontinuities called overlapping spreading centers (OSC) (circled). Migration rates of larger OSC are shown in speeds (mm/yr); these OSC are interpreted by MacDonald et al. (1988) as fundamental boundaries of magmatic segmentation. Arrowheads indicate two recent propagating rifts. At left, an enlargement of the section from 9° to 13° N shows the structure of the ridge and occurrence of an axial magma chamber detected inan MCS experiment. Where there is an axial magma chamber, the cross-sectional shape of the rise tends to be:broad, and an axial summit graben is usually present; this occurs along shallow parts of ridge segments, away from discontinuities. Where the magma chamber is absent, the ridge is narrow (triangular symbol) and deep, and the summit graben is absent, b) Axial depth profile of the EPR from 8° to 18° N. Discontinuities of order 1-3 (transforms and overlapping spreading centers) are identified from Sea Beam charts; 4th order discontinuities are identified geochemically. Long wavelength undulations in the axial depth profile are bounded by first- and sometimes second-order discontinuities, whereas short-wavelength undulations are marked by third-order discontinuities. Fourth-order discontinuities have little or no bathymetric signature (from MacDonald et al., 1988).  Chapter  1.  INTRODUCTION  a  13  Axial depth profile Long-wavelength undulation of the axis Short-wavelength undulations.of the axis  4  '  4  Figure 1.7: Model of magmatic segmentation of MOR. a) Schematic along-strike cross-section of ridge. First- to third-order discontinuities occur at local maxima in the depth profile and are indicated by large numbers; dashed circle shows area that is enlarged in b). b) Possible configuration for fourth-order segmentation (defined geochemically). Although the pattern of asthenospheric upwelling is unknown, the model of MacDonald et al. (1988) suggests that the source of upwelling is enhanced beneath the shallow portions of the ridge, but has some flux along most of the length of the ridge. Melt segregation events occur 30-60 km beneath the axis and replenish and inflate the axial magma chambers beneath separate ridge segments as shown (from MacDonald et al., 1988).  Chapter 1. INTRODUCTION  1.3  14  Previous MCS surveys  MCS surveys in the deep ocean have become more common over the past 10 years or so with the availability in academic and government institutions of computing facilities previously exclusive to industry. This type of survey provides an intermediate level of resolution between large scale refraction surveys and the small scale geological mapping of ophiolites or drill cores and should provide a much clearer structural picture of the oceanic crust. MCS surveys have long been the primary tool in petroleum exploration, providing the most detailed remotely-sensed picture of the subsurface available. For this same reason they are a valuable but expensive procedure for determining oceanic crustal structure. MCS data reveal a high degree of variability in oceanic crustal reflectivity, ranging from almost acoustically transparent to highly reflective both temporally as expressed by distance normal to the ridge, and along the spreading axis (Fig. 1.8). In the 'average' refraction model, most of the layer boundaries are characterised by a change in velocity gradient rather than a discrete jump in velocity. The main exception is the Moho; it is often characterised by a velocity jump from about 7.5 km/s to above 8.0 km/s (White, 1984). On MCS data, a complex zone of strong reflectivity at about 2-2.5 s two-way time (TWT) below the top of the basement is often seen and has been interpreted as Moho (e.g. NAT Study Group, 1985) (Fig. 1.8). These reflections have been seen in many MCS surveys, and in crust of almost all ages, excluding zero-age crust at the ridge crest. However, normal incidence seismograms computed for the Bay of Islands ophiolite in Newfoundland show no correlation between reflectivity and the interface between the depleted upper mantle and the base of the magmatic material comprising the crust (i.e., the 'petrologic Moho', see Fig. 1.4). Based on the synthetics, the first high amplitude reflections may arise from a number of different lenses of mafic and ultramafic material in the crust, but are most often associated with the layered mafic and ultramafic zone near the base of the crust - the 'Moho transition zone'  Chapter  1.  INTRODUCTION  15  Figure 1.8: MCS data with interpreted events from a) Mesozoic North Adantic crust (NAT Study Group, 1985); b) Endeavor segment of the Juan de Fuca ridge (Rohr et al., 1988).  Chapter  1.  INTRODUCTION  16  (Collins et al., 1986)(see Fig 1.4). Given the gradient velocity model interpreted from refraction surveys, it may be surprising  that we see any reflectivity within the oceanic crust. Two factors make imaging reflectors  particularly difficult. First, the acoustically hard (e.g. reflection coefficient - 50%) and rough upper crust tends to scatter incident seismic waves, reducing energy propagation below this interface. Second, velocity contrasts between rock types which comprise the oceanic crust are small since its composition and physical properties such as porosity tend to change in a gradational manner. In areas where refraction models show velocity jumps between layers which would produce large reflection coefficients when modelled with synthetic reflection seismograms, the boundaries may not be discrete, first order interfaces. On the other hand, constructive interference (tuning due to thin layers) may produce reflectivity in areas where large velocity jumps are not seen in refraction models (Jones and Nur,  1984). In any case,  intracrustal reflectors have been imaged in the oceanic crust (although they are much less widely observed than the reflective package at the base of the crust) indicating that sharp transitions, relative to refraction wavelengths, exist within it. The best examples of intracrustal reflectors are seen on seismic data acquired with two ships and an expanded receiver array, which allows for a better signal-to-noise (S/N) ratio and velocity discrimination (e.g. the North Atlantic Transit (NAT)  experiment, Fig.  1.8). Table 1.2 summarizes reflectors imaged  in the oceanic crust.  1.4  Tectonic setting of study area  The  Juan de Fuca plate is a small oceanic plate located off the coast of British Columbia  and  reaching south as far as northern California (Fig. 1.1). It is bounded on the west by  the Juan de Fuca ridge, which extends 500 km between the Sovanco and Blanco fracture zones, and on the east by the Cascadia subduction zone. It is the northern remnant of the  17  Chapter 1. INTRODUCTION  Table 1.2: Oceanic Crustal Reflections Reference  Interval Velocity Proposed Origin (kmls)  2-Way Time Below Top of Basement (s) 0.3-0.5  Rohr et al., 1988  2.7-4.7  2A/2B boundary or metamorphic front  0.25-0.4 0.6-1.0  Musgrove and Austin, 1983 McCarthy et al., 1988  5 5.5  1.6-1.9 ?? 1.8-2.0  Musgrove and Austin, 1983 Sellevoll and Mokhtari, 1988 NAT Study Group, 1985; McCarthy et al.,1988 NAT Study Group, 1985, etc.  2B/2C boundary sheeted dikes/ gabbro contact 3A/3B boundary  6.0  fault zone or mafic/ ultramafic boundary Moho  2.0-2.5  6.3  Farallon plate, which was fragmented when the East Pacific Rise collided with the North American plate at about 27 My, creating the proto-San Andreas fault. Since that time it has continued to diminish in size due to northward migration of the southern triple junction to its present location at Cape Mendocino and continuing subduction beneath the North America plate (Riddihough, 1984). During the last 10 My, the Juan de Fuca plate has broken up into a number of small subplates near both the northern and southern triple junctions, and the ridge has undergone a clockwise rotation from north to N20°E (Riddihough, 1984; Wilson et al., 1984). Separation of the Explorer ridge to the north from the Juan de Fuca ridge began at about 7.5 My with the inception of the Sovanco fracture zone (Botros and Johnson, 1988). At about 4 My the northern Explorer segment of the plate was detached from the Juan de Fuca plate along the Nootka fault, probably because of greater resistance to subduction of its younger, more buoyant lithosphere (Riddihough, 1984).  Chapter  1.  INTRODUCTION  18  The spreading rate at the Juan de Fuca ridge has decreasedfromabout 80-90 mm/yr (half rate) at 40 My to its present rate of about 35 mm/yr. In a minimum energy configuration, spreading along the Juan de Fuca ridge would be parallel to the bounding Blanco and Sovanco transform faults, which trend about N120°E. However, the ridge axis trends N20°E, resulting in asymmetric spreading, ridge jumping and rift propagation as the ridge accommodates this disparity in plate orientation. The effects of ridge reorganization are not confined to the plate boundary zone, but create a swath of disrupted magnetic anomalies and disturbed seafloor extending out onto the plate (Kleinrock and Hey, 1989; MacDonald et al., 1988). Consumption of the plate at the Cascadia subduction zone is currently taking place in a N35°E direction and has decreased over the past 7 Myfromabout 60 mm/yr to about 40 mm/yr (Riddihough, 1984). The pattern of. magnetic lineations on the Juan de Fuca plate does not show the symmetrical, parallel stripes, offset at transform faults, which one might expect about a stable mid-ocean ridge; rather, they are in some cases offset and oblique to the current ridge or discontinuous (Fig. 1.9). Wilson et al. (1984) have successfully modelled the magnetic anomaly patterns in this area with a series of seven episodes of rift propagation, a process by which the ridge may have reoriented itself to the current direction of plate motion. Although the causes of rift propagation are not well understood, the interpretation of magnetic anomalies as well as the observed structures at well-studied propagating rifts (e.g., the propagator system near 95.5°W on the Galapagos (Cocos-Nazca) spreading axis) indicate that in these locations an offset at the ridge is followed by along-axis extension of one ridge segment and simultaneous cessation of spreading at the other (Wilson et al., 1984; Kleinrock and Hey, 1989). Fig. 1.10 compares a stable spreading system to one modified by a propagating rift. In the stable spreading regime, crustal accretion at independent spreading segments separated by transform faults produces magnetic anomaly patterns which are symmetric about each segment but offset at fracture zones. In a propagating rift environment,  135° W  130° W  125° W  Figure 1.9: Magnetic anomaly pattern in the Juan de Fuca plate area showing numbered pseudofault traces (solid lines) of propagating rifts as modelled by Wilson et al. (1984). Magnetic polarity time scale (right) shows both detailed scale from Mankinen and Dalrymple (1979) (left) and simplified version used by Wilson et al. (1984) for forward modelling.  P R O P A G A T I N G RIFT Doomed Spreading Axis  b)  a)  3 §  STABLE SPREADING S Y S T E M Spreading Axis  . . Transform Zone Propagator Tip Failing i \ Spreading t ' ' * Axis -  \ \  * " •$  Iii  Spreading Axis  Failed Spreading Axis Inner Pseudofault Zone  Propagating Spreading Axis Outer Pseudofault Zone  Figure 1.10: a) Stable MOR spreading system: magnetic stripes are parallel to the spreading axis and offset only at transform faults; b) Propagating rift: magnetic stripes are disrupted at the V-shaped 'pseudofaults' which mark the past position of the propagator tip and separate crust formed at the PR from the older crust formed at the failing spreading axis through which the ridge propagates.  1  Chapter  1.  INTRODUCTION  21  ridge reorganization is accommodated by changes in the length and orientation of the spreading  segments. The propagating segment lengthens into old crust previously formed at the  failing segment. Discontinuities in the magnetic anomaly pattern are interpreted as 'pseudofaults' - the contact between older crust created at the failing rift and new crust formed at the tip of the propagating rift. The  area of interest in this study is crossed by PRs 4 and 7 in the Wilson et al. (1984)  model, which both propagated northward at about N20°E (Fig.  1.9). PR7  propagated between  about 5-1.7 My, stopping when it intersected the Sovanco fracture zone; PR4 began propagation at about 4.5 My,  slowed down at about 1.65 My,  and now appears to be stalled in an  overlapping spreading center configuration at the Cobb offset (Karsten and Delaney, 1989). The  data interpreted in this study comprise the first MCS  data collected over pseudofault  traces, and these features will be considered in detail in a later section.  1.5 Summary The  MCS  data interpreted in this study provide new information about the structure of young  oceanic crust, both as it ages and along isochrons. Current information on the structure and evolution of oceanic crust has been summarized in this chapter. The following chapters will discuss the characteristics and implications of the data in detail.  Chapter 2  ACQUISITION AND PROCESSING PARAMETERS  2.1 Acquisition The 230 km of multichannel seismic reflection data interpreted in this study were acquired by the Geophysical Services Inc. vessel Edward  O. Wetter in 1985 for the Geological Survey of  Canada as part of the Frontier Geoscience Program. The lines (85-07 and 85-09) were shot in transit between primary survey targets (85-01 to 85-05) across the Cascadia subduction zone and the Juan de Fuca ridge; line 85-06 was another transit line (Fig. 2.1). The data were recorded to 16 s at a 4 ms sampling rate using a Texas Instruments DFS V system. A 3000 m neutral buoyancy streamer of 120 25 m groups, each consisting of 15 acceleration cancelling hydrophones, was deployed 15-22 m below the surface (Yorath et al., 1987). Since maintainance of the airgun array for quality control of the primary survey targets was carried out on the transit lines, the source varied over both lines. About one third of the data at the eastern end of line 85-07 and the western end of line 85-09 were recorded using the full array of airguns, which consisted of a four string, 25 element tuned array with a total capacity of 50 1. The other segments of the lines were recorded with either the port or starboard guns, depending on which were being serviced at the time. Both the full and partial arrays were operated at 2000 psi and deployed at 12 m depth below the sea surface. A source-streamer offset of 287 m was used. The shotpoint interval was virtually constant at 75 m on line 85-09, with only a few missed shots; line 85-07 had many missed shots and a highly variable shotpoint interval over some segments of the line where the ship slowed  22  Chapter  2. ACQUISITION  AND  130°  130°  PROCESSING  W  W  PARAMETERS  23  125°  W  125" W  Figure 2.1: MCS line locations in the northern Juan de Fuca plate area (line numbers 85-01 to 85-09); squares indicate OBS locations from refraction experiment (dotted line) used for velocity control at eastern ends of MCS lines interpreted in this study; numbered pseudofault traces (solid lines) of propagating rifts as modelled by Wilson et al. (1984) are also shown. C-E : Cobb-Eickelberg seamount chain.  Chapter  2. ACQUISITION  AND  PROCESSING  PARAMETERS  down to pull in the guns. The nominal common midpoint (CMP)  24  fold for this line geometry  is 20; ship speed variations which would increase this value were compensated for by firing only every other shot (particularly on line 85-07), so that only minor variations in the  CMP  fold occur. Ship navigation was by Loran-C as well as Transit satellite and doppler sonar systems; in general, the ship maintained course with minimum feathering of the streamer along the survey line. The data were acquired on tape with few field monitors, and had not been seen except in the form of single trace plots, until the demultiplexed records were stacked in the initial stages of this study. Table 2.1 summarizes the acquisition parameters. Fig.  2.2 shows shotpoint locations for lines 85-07 and 85-09.  2.2  Data quality  From the point of view of data processing, the most significant feature of oceanic crustal reflection data is the acoustic character of the top of the igneous basement. The  high reflection  coefficient (estimated by techniques detailed in section 3.2.4 to range from 20-63%) at this interface due to the large contrast in velocity between the sedimentary cover and  the oceanic  basement reduces the amount of energy travelling through the crust, so that careful processing is required to enhance any weak crustal reflectors. The  roughness of the extrusive upper layer  has a significant effect in producing scattered noise in the deeper section, both from in-line and  out-of-plane scatterers; diffraction tails from this interface dip down into the crustal  section, obscuring any weak, subhorizontal reflections within the oceanic crust. Both of these features can be observed on shot gathers; typical gathers from both the west end of line 85-09 and the east end of line 85-07 are shown with stacked sections for comparison (Figs. 2.3 - 2.4). The clean, high amplitude, hyperbolic sedimentary arrivals contrast with the complex, discontinuous, doubled or tripled arrivals from the top of the basement. Directly below this interface, the curved limbs of diffractions dominate the shot  Chapter  2. ACQUISITION  AND  PROCESSING  PARAMETERS  Table 2.1: Acquisition parameters M/V Edward 0. Vetter GSI Primary system - Loran C Secondary system - Argo Date shot May 1985 Seismic reflection Data type Recording system DFS V Recording format SegB Recording density 1600 B.P.I. Sample period 4 ms Record length 16 s Recording filters Low cut 3.5 Hz at 18 db/octave High cut 64 Hz at 72 db/octave Tape polarity Compression negative Seismic source Airgun array 4 string 3000 cu. in./2000 p.s.i 11 m Average source depth Cable length/avg depth 3017 m/15 m Number of groups 120 Normal group interval 25 m Normal shotpoint interval 75 m Shots per shotpoint 1 Nominalfieldmultiplicity 20 Average near group offset 287 m Fathometer 1477 m/s draft corrected Vessel Shot by Navigation system  25  Chapter 2. ACQUISITION AND PROCESSING PARAMETERS  r\  A  \V* ioi^—^ y  & 400 1/0OO / 800 / ^^1000 / ^ b  /  / I /  /  /  2301^ f  1357  2200  00 \ A1800 /l600  0 • s ^ / 1200  1  S \ 000 ^800  V  y^eoo VM00 y200 101  128  127  Figure 2.2: Shotpoint locations for MCS lines.  J  Figure 2.3: Typical shot gather from line 85-09 (SP 264) shown with stacked section for comparison. Labels indicate water bottom (W), sedimentary arrivals (S), top of oceanic crust (B), scattered arrivals (D) and base of oceanic crust (M). The shot gather has been corrected for geometrical spreading and plotted with a scalar gain. The stacked section is plotted with an automatic gain control (AGC).  K>  200  SP 160 3=5  175  150  n 3=5 t-o  4=0  4=0  to  I 4„5  4 =5  5D0  5=0  I e o  ES  7.0  5=5  5 =5  6=0  6=0  6=5  6=5  7o0  7 =0  Figure 2.4: Typical shot gather from line 85-07 (SP 160) shown with stacked section for comparison. Labels as in Fig. 2.3; also shown is an intracrustal reflector (IC). Display parameters as in Fig. 2.3. to  Chapter  2. ACQUISITION  AND  PROCESSING  3.5  - -  4c0  4-  29  PARAMETERS  4„5  5o5  -jr  6.0  6.5  7.0  i  Figure 2.5: CMP gather at SP 138; compare with Fig. 2.4. Note the hyperbolic arrivals which dominate the gather between 5-6 s. Most of these events stack in at the sedimentary or top-of-basement velocities, and therefore are interbed multiples or scattered events. The gather has been corrected for geometrical spreading and plotted with a constant scalar gain. gathers; the curvilinear nature of these arrivals indicates that they originate from scatterers on or near the seismic line. Between about 4.5 and 6 s two-way time (TWT), the shot gathers are  dominated by more or less linear energy, dipping at apparent slownesses of about  ms/trace (1/2500 m/s). When sorted to CMP observed on this segment of the data (Fig.  ±10  gathers, predominantiy hyperbolic events are  2.5). Lamer et al.  (1983) showed that this pattern  of noise (linear in shot gathers, hyperbolic in CMPs) could be explained by scattering from irregularities in the water bottom and sub-bottom. The apparent velocity of the scattered energy depends on the azimuth of the scatterer with respect to the streamer, with the highest velocities for scattered energy from direcdy ahead of or behind the boat and streamer. In  Chapter  2. ACQUISITION  AND  PROCESSING  PARAMETERS  30  general, the diffracted arrivals have a frequency content very similar to that of the data (Fig. 2.6), and so are difficult to remove by simple processing methods. At about 2 s TWT  below basement, a fairly high amplitude, low frequency arrival (M)  is seen on the mid- to far-offset traces in some of the shot gathers (Fig. 2.3). The normal moveout of this event is small, indicating a high stacking velocity. All of these features support an interpretation of this event as a reflection originating at the base of the crust or Moho transition zone (see Section 3.2.4 for further discussion). The  quality of the data is in general very good. Some shot gathers and thefirstbrute  stacks revealed the strong, reasonably continuous reflector at the base of the crust. Initial stacks also showed the very strongest segment of the intracrustal reflectors at the east end of line 85-07. Although the variation in the source array is noticeable on true-amplitude sections as an abrupt amplitude drop off, there does not appear to be any marked change in continuity of reflectivity within the crust: on line 85-07 the starboard guns were turned off  at SP 407, but the continuity of intracrustal reflectors and the basal reflection are not  affected. The source signature does not change noticeably at this point, nor is the amplitude on individual traces noticeably affected. The  general lack of intracrustal reflectors may  reflect  the relatively short streamer length and low fold of the data compared to experiments such as the NAT  experiment (NAT  Study Group, 1985), in which many more farther offset traces  could be stacked: stacking improves the signal-to-noise ratio by n , where n is the fold. 1/2  Therefore, the S/N for an experiment such as NAT,  which produced 60-fold data, would be  about twice that of these data. Such an improvement would increase the chances of imaging weak reflectors in the crust. On the other hand, the fact that intracrustal reflectors are imaged at one portion of 85-07 may indicate that the lack of reflectivity elsewhere is not due to some deficiency in the data but is rather a true representation of the character of the oceanic crustal reflectivity in this area (see Section 3.2.3 and 4.2.2 for discussion).  4.66-4.86  S (B)  0.4 0.2 0.0  HI Q D I— 40  Q LU  5.5-5.7  N _J  0.08  <  0.04  DC  0.00  o  80 S (NOISE)  40 6.14-6.34  80  120 40  S (M)  6.60-6.80  0.08  80  20  S (M)  0.04  0.04  0.02  0.00 40  80  120  0.00 40  80  120  FREQUENCY (HZ) Figure 2.6: Frequency content of selected time segments of the data normalized to highest amplitude of trace at water bottom (note changes of scale), a) Trace 40 of SP 264, line 85-09. The frequency content of the noise is similar to that of the data, b) Trace 71 of SP 137, line 85-07. Labels as in Fig. 2.3.  Chapter 2.3  2. ACQUISITION  AND PROCESSING  PARAMETERS  32  Processing techniques  2.3.1 Processing systems Two different processing systems were used in the analysis of this dataset. Thefirst,in use at UBC, consists of a MicroVAX  II minicomputer and MiniMAP array processor with  Inverse Theory and Applications (ITA) LNSIGHT/1 software, designed for interactive use. With this system, the user may pick such processing parameters as velocities and mute patterns and design deconvolution operators, etc., interactively at the terminal using a highresolution graphics processor, in this case a METHEUS Q500, and Sony monitor. The second processing system comprises a CYBER 835 computer and MAPV array processor with Cogniseis DISCO software. This system is in use at the Lithoprobe Seismic Processing Facility located at the University of Calgary and is available for remote use via DATAPAC. Display of processing results is submitted to the plotter in Calgary; plotted output is then sent to the user by courier.  2.3.2  Prestack processing  The major processing problem for this data set is the removal of dipping noise within the crustal section due primarily to scattering from the rough, acoustically hard surface of the igneous basement. Since the frequency content of the noise is similar to that of the signal (Fig. 2.6), a simple bandpassfilteris not suitable. The most successful technique for removal of such noise in similar datasets has been the application of a prestack f-k or dipfilter(Larner et al., 1983; Rohr et al., 1988; etc). Such afilteris ideal for removing noise which dips with a different apparent velocity than the signal. Reflectors within the igneous crust are expected to be, in general, subhorizontal, and so a judicious selection offilterparameters should allow removal of the dipping noise but retention of the signal. The resulting improvement in data quality should aid in picking stacking velocities for weak reflectors, since scattered noise  Chapter  2. ACQUISITION  AND PROCESSING  PARAMETERS  33  (which may have high apparent velocities, depending on the azimuth of the scatterer) will be reduced. A two-step prestack implementation will ideally result in the tightest possible f-k filter: first, the noise dipping opposite to the direction of normal moveout in the shot gathers is filtered, then the traces are sorted to receiver gathers and the noise dipping in the opposite direction is filtered. However, the geometry of this experiment precluded this two-step process, because the 75 m trace spacing in the receiver gathers led to spatial aliasing in the filtered data. Various dipfilteringtechniques were tried, including the ITA f-kfilter,and the DISCO f-k filter and time-domain dipfilter.None of these implementations gave satisfactory results. Although the results looked promising infilteredshot gathers, in the stacked data the spurious noise introduced by thefiltersappeared worse than the original noise problem (Fig. 2.7). The main source of this spurious noise is probably the taper applied to the edges of the data. In the ITA prestack f-k application, the length and shape of the taper are not controlled by the user. With DISCO, one may design a taper function fairly precisely, but tests proved unsatisfactory. All the filters tended to produce artifacts dipping at the same apparent velocity as the edges of thefilter;this made it impossible to apply a tightfilterto the crustal section, because the artifacts began to look like 'real' subhorizontal events. Since the results of the dipfilteringattempts were unsatisfactory, a less sophisticated approach was tried: a running mix of shot gathers. The philosophy behind this approach is simply one of applying a spatialfilterto the data, much the same as an f-kfilter.Although this process will degrade the frequency content of the data, the frequency content of the signal below the top of the igneous crust is limited (Fig. 2.6), and higher frequencies were deemed expendable in the search for the elusive intracrustal reflectors. Again, the stacked results proved unacceptable. Using a small number of traces (e.g., 3) did not improve the S/N ratio of the stack substantially, while using larger numbers of traces resulted in excessively wormy, low frequency sections (Fig. 2.8).  Figure 2.7: Prestack implementation of the f-k filter, a) SP 160, line 85-07, filtered to exclude apparent velocities outside the range 12000 m/s and -3000 m/s. Positive values indicate dipsfromright to left; negative values, dips from left to right. Compare with Fig. 2.4. Note the linear noise dipping from right to left, at the same apparent velocity as the edge of the filter. This noise is more obvious in b) stack of prestack f-kfiltereddata. Although in this display lateral continuity of some of the reflectors is improved, this was not the case in general. Display parameters are the same as in Fig. 2.4.  the range 12000 m/s and -3000 m/s. Positive values indicate dips fromrightto left; negative values, dips from left to right. Compare with Fig. 2.4. Note the linear noise dipping fromrightto left, at the same apparent velocity as the edge of the filter. This noise is more obvious in b) stack of prestack f-k filtered data. Although in this display lateral continuity of some of the reflectors is improved, this was not the case in general. Display parameters are the same as in Fig. 2.4.  SP  200  175  150  Figure 2.8: Prestack 5 trace running mix of shot gathers. The shot gathers are first corrected for normal moveout at stacking velocities, and then laterally summed across each gather, 5 traces at a time. The data are displayed with an 800 ms window AGC.  2. ACQUISITION  Chapter  AND  PROCESSING  PARAMETERS  36  Ultimately, the best solution to the removal of scattered noise from the top of the crust was discovered in the attempt to find possible intracrustal reflectors. Over the course of the velocity analysis, far-offset constant velocity stacked sections were produced. The rationale behind this analysis is that on a section which has the normal moveout (NMO) and  been stacked at a constant 'crustal' velocity,  of noise such as interbed multiples from the sedimentary column  scattered energy from irregularities nearly 90°  from the survey line will be overcorrected,  with a greater overcorrection for farther offsets. Therefore, in a far offset stacked section, this noise should be attenuated since it is improperly stacked. Real events, on the other hand, should be enhanced when stacked at 'crustal' velocities. This was indeed the case for this data set: the far-offset sections were less noisy in the crustal section, and intracrustal events were discerned in some areas. In fact, study of the shot gathers also shows this pattern: the near-offset traces are noisier than the farther offset traces (e.g., Fig. mute pattern for the CMP  2.3). When designing a  gathers, this information was taken into consideration: in addition  to the standard outer trace mute in the upper sections of the gathers to remove those segments distorted by NMO Fig.  'stretch', the inside traces were muted in the lower sections of the gathers.  2.9 shows an example of this mute pattern. Other processing steps applied to the data included a spherical divergence correction and  a prestack trace amplitude equalization. The spherical divergence correction was applied to some of the data before stacking in order to correct trace amplitudes for the geometrical divergence of the seismic waves as they travel from source to receiver, and t 'VRMS  2  involves a simple  multiplication. Prestack equalization of trace amplitudes across individual  CMP  gathers was performed to remove amplitude variations due to variable receiver response or shot amplitude. Thefilterwas designed on a time window within the crust which was free of reflectors, in order to equalize noise levels on all the traces (Fig. 2.10).  Figure 2.9: Mute pattern for CMP gathers (offset decreases to the right): a) NMO applied to gather using a time-varying velocity-function, b) Mute applied to the moved-out gather. The CMPs appear different because of the gain applied within the plotting routine, which depends on the maximum amplitude of the gather.  Figure 2.10: Prestack energy equalization across SP 264, line 85-09 (nonequalized shot at right). Note the more uniform amplitude levels across the gather, particularly the noisy trace. Data have been corrected for geometrical spreading and displayed using a constant scalar gain.  LO OO  Chapter 2.3.3  2. ACQUISITION  AND  PROCESSING  PARAMETERS  39  Poststack processing  Poststack processing routines were tested for improvement of the final stacked section. The most important post-stack process was f-k migration at the water velocity. The data were also f-k filtered and/or bandpass filtered, particularly for enhancement of the deep structure. Poststack coherencyfiltering(A. J. Calvert, unpub. algorithm, 1989) was performed on  stacked sections, and proved useful in the final presentation and interpretations. The  coherencyfilteringalgorithm was also used to perform a 3- or 5-trace running mix of the stacked data along the dips of greatest coherency excluding large dips. This enhanced the lateral coherency of the very weak reflectors. Figs. 2.11-2.12 provide examples of these results. Final sections were displayed using either a scalar gain or automatic gain control (AGC), using an LI norm and 0.800 s window length.  2.3.4 The  Velocity analysis velocity model used to obtain initial stacks was derived from the results of an earlier  refraction experiment in which ocean bottom seismographs (OBS) were laid out near the eastern ends of both reflection lines (Fig.  2.1) (White and  Clowes, 1988). Various techniques  were used to perform further velocity analysis using the reflection data. However, because of the decrease in resolution with depth of the velocity function derived from reflection analysis (Fig.  2.13), the velocities of the deep reflectors are not well-constrained by these methods.  This decrease in velocity resolution arises from the short streamer length which is standard for  imaging the relatively shallow targets common in industrial acquisition. Distinguishing  velocities using reflection data relies on differential moveout between the near and  far offset  receivers, i.e., the difference in distance travelled by the seismic wave from the source to the near and far ends of the streamer. For  the higher stacking velocities, the time difference  between travel paths to the near and far offset receivers for different velocities is small and  a)  b)  Figure 2.11: a) Post stack f-k filter, passing apparent velocities between +2000 m/s. Note the reduction in dipping noise within the crust, b) Post stack bandpass 5-20 Hz. Although the bandpass improves the continuity of the weaker signal, for example at 6.0 s, it also enhances some of the dipping noise. Data are displayed using AGC, window length 800 ms.  ©  Figure 2.12: a) Post stack 5 trace running mix along dip of greatest coherency, excluding dips with apparent velocities greater than ±2000 m/s. This filtering technique does a somewhat better job than the poststack f-k filter, but results in a somewhat wormy-looking section, b) F-k migration at water velocity, using 300 trace window with 150 trace overlap. Although the basement surface is overmigrated at this velocity, the noise within the crustal section is quite nicely removed, except for some migration 'smiles'. The use of this low velocity also prevents aliasing problems associated with f-k migration. Display parameters as in Fig. 2.11.  Chapter  2. ACQUISITION  AND PROCESSING  PARAMETERS  42  therefore the resolution between travel paths is poor, i.e., a large range of velocities will have a similar moveout. Velocity analysis was carried out in a number of different ways. Analysis of CMP gathers was performed by merging adjacent CMPs to give 40-fold gathers, and then correcting them for normal moveout at a series of constant velocities. Events which appear horizontal on such displays are properly corrected for NMO;  one may then compute interval velocities from  these stacking velocities via the Dix formula. In the search for intracrustal reflectors, several adjacent 40-fold CMPs were stacked to improve the signal-to-noise ratio; for very weak reflectors, a 3- or 5-trace running mix was applied across stacked gathers which had been moved out at constant velocities (Fig. 2.14). In such analyses, typical velocity increments used would be 50 m/s between YRMS of 1450-1600 m/s, 100-200 m/s between 1700-2500 m/s and 500-1000 m/s between 2500-4500 m/s. Greater resolution of the velocity function at shallower depths (i.e., within the sedimentary section) was possible using semblance velocity analysis, a process in which the coherency of the data is computed across a gather which has been NMO-corrected over user-defined increments of velocity and time, assuming hyperbolic moveout. A sample of such an analysis is shown in Fig. 2.13. At certain locations, panels of data stacked at different constant velocities were produced and displayed using a constant scalar gain. The most useful of these sections included only the ten far offset traces. Panels which showed potential crustal reflectors were also migrated at the same constant velocity, to ascertain whether the reflectors were properly focussed at approximately this velocity. This analysis was probably the single most important step in confirming the presence of intracrustal reflectors in these data (Fig. 2.15).  Figure 2.13: Semblance analysis for CMP 250 (SP 101), line 85-07. Note the smearing of the semblance peaks at greater times, indicating a decrease in resolution of the velocity function with depth. Improved resolution of the peaks at crustal depths requires different acquisition parameters, i.e., a longer streamer (see text).  £  Figure 2.14: 5 trace running mix of two-fold stack of adjacent CMP gathers from SP 177, line 85-07, merged to 40-fold and corrected for normal moveout at constant velocities. Note the apparent lack of reflectors within the crust, except perhaps at about 6 s. Data are displayed with AGC length of 800 ms.  Figure 2.15: Constant velocity stack of east end of line 85-07: a) 1550 m/s. This is approximately the velocity to the top of the basement; any intrabed multiples between the water and the basement, or scattered energy from in-plane or nearby sources should stack in at approximately this velocity. Note that linear energy stacking in at this velocity is confined to within 800 ms of the top of the basement; b) 2500 m/s. A number of linear features appear in the data stacked at higher velocities (see also (c), following page) which are not evident on CMPs moved out at constant velocity (Fig. 2.14). These events are not seen on the constant velocity stack at upper crustal velocity (a); therefore they cannot be multiply reflected energy from this interface, and may be intracrustal reflectors. Note the high apparent velocity dipping noise in both (b) and (c); these events would be due to scatterers nearly directly ahead of or behind the streamer; some of this noise may be reduced by the mute pattern used before stacking the data (Fig. 2.9), which cuts off the outside traces near the top of the basement; c) 3500 m/s (following page).  Chapter 2.3.5 The  2. ACQUISITION  AND  PROCESSING  PARAMETERS  Al  Final processing stream processing stream used to produce final stacked sections is shown in Fig. 2.16. Noisy  traces were edited out in areas of particular interest, although in general this was unnecessary. The  shot gathers were also resampled from 4 ms to 8 ms to reduce the amount of data. The  Nyquist frequency after resampling is 62.5 Hz, (Fig.  well above the frequency content of the data  2.6). There is some variation in the poststack processing steps performed depending  on the desired emphasis, e.g., resolution of the sedimentary structure vs. enhancement of deep reflectors. The geometrical spreading correction was in certain cases applied, prestack and  in others applied post stack. An automatic gain control with 800 ms window length was  generally used for display purposes.  2.4 Summary The  basic processing stream was applied to the data to arrive at stacked sections for the  entire 230 km of lines 85-07 and 85-09, enabling an overview interpretation of both of the lines. Particular segments of the sections were selected for further processing (primarily poststack techniques) to enhance features of particular geological interest. Discussion of both the regional and more local features of this dataset follows.  Chapter  2. ACQUISITION  AND PROCESSING  PARAMETERS  48  TRACE EDIT RESAMPLE TO 8MS GEOMETRICAL SPREADING CORRECTION CMP  SORT MERGE ADJACENT CMPS TO 25M SPACING VELOCITY ANALYSIS  NORMAL MOVEOUT CORRECTION MUTE STACK (20-FOLD)  1 CRUST  1  F-K  MIGRATION  BANDPASS (5-40 HZ)  POSTSTACK P R O C E S S E S : - RUNNING MIX - F-K FILTER - F-K MIGRATION - COHERENCY FILTER - BANDPASS (5-20) HZ  Figure 2.18: Final processing stream  Chapter 3  OCEANIC CRUSTAL STRUCTURE OBSERVED ON MCS DATA  3.1  Introduction  The high quality of the data set acquired in this survey is shown by stacked sections of both lines (Figs. 3.1-3.2), which sample young oceanic crust from about 1.4-3.7 My (85-07) and 1.8-5.3 My (85-09) in age. The stacks show a reflective, subhorizontal sedimentary sequence dipping slightly landward and overlying a rough, prominent basement reflector. Reflections interpreted as originating at the base of the crust (Section 3.2.4) are also evident at about 2 s two-way-time (TWT) sub-basement in the sections. The water bottom multiple occurs at about 7 s TWT, generally below features of interest, except at the eastern end of 85-09 where it obscures the basal reflection. Intracrustal reflectors are absent or much less prominent, except on the eastern third of line 85-07. Dipping coherent noise, attributed to scattering from the rough basement surface, appears on the sections below the top of the basement. As a visual aid to interpretation of the data, several different displays are presented: Figs. 3.3 and 3.4 show coherency-filtered panels accompanied by line drawings which were derivedfromcomparison of stacked sections using different post stack processing parameters as discussed in Chapter 2 (e.g., bandpass, f-kfilteror f-k migration) and coherency sections filtered at different threshold levels. The following discussion of the important features of the data relies on these figures plus a number of detailed closeups.  49  Chapter3.  OCEANIC  CRUSTAL  STRUCTURE  OBSERVED  ON MCS  DATA  50  Figure 3.1: CMP stack of line 85-07 (following page). Note the irregular shotpoint interval due to acquisition parameters (see Fig. 2.2). S = sedimentary section; B = oceanic basement; IC = intracrustal reflectors; M = crust-mantle transition; PT = outer pseudofault trace of propagating rift; M y = approximate crustal age in My before present.  Chapter  3. OCEANIC  CRUSTAL  STRUCTURE  OBSERVED  ON MCS  DATA  Figure 3.2: CMP stack of line 85-09 (following page). Symbols as in Fig. bottom multiple.  52  3.1; W = water  TIME ( S ) cn  a  cn  O  cn  CD •  a  cn  a  O  o  cn  o  CJ a  •  cn  o  ++  CO  5®S  SI  fillip® ro co  BB  1- v-^. ^:  JHBfl  1§P  l i i i  IH S  '  #11  lie §;§W i'.t ' S "  ro  Ax  Ill  1 m  I I I I I I I I I cn a  O  cn  CD  [I I I I I I I cn a CJ1  11 I II I I  i M M1111111111  cn o o  (JI  O  CO a  CO a  cn  O  TJ  LA)  Chapter  3. OCEANIC  CRUSTAL  STRUCTURE  OBSERVED  ON MCS  DATA  54  Figure 3.3: (Following page): a) Coherencyfilteredstack of line 85-07. Semblance maxima of most coherent dip between dips of ±0.002 s/tr are summed over 5 traces; the plot is a 2-trace sum of the result, b) Line drawing of interpreted primary reflections. S = sedimentary section; B = top of igneous oceanic crust; IC = intracrustal reflectors; D = disrupted sediments; CMT = crust-mantle transition; U = interpreted underplated zone; W = water bottom multiple; PT = pseudofault zone of propagating rift.  NE  Chapter  3. OCEANIC  CRUSTAL  STRUCTURE  OBSERVED  ON MCS  DATA  56  Figure 3.4: (Following page): a) Coherency filtered stack and b) line drawing of line 85-09. Parameters and abbreviations as in Fig. 3.3.  SW a)  SP  NE 200  i  1—  V.E. -2:1  400 1  600 1  800 1  :  1000 i  1200  _L  1400 i  1600 i  1800 l  2000  1  1  1  Chapter 3.2  3. OCEANIC  CRUSTAL  STRUCTURE  OBSERVED  ON MCS  DATA  58  Major reflectors  3.2.1 Sedimentary sequence The  sedimentary sequence on both lines is wedge-like in shape, increasing in thickness and  dipping slighdy toward the east. The thickness of the sedimentary unit varies from about 200 m at the western ends of the lines to about 1500 m at the eastern ends. The sedimentary cover near the western ends of the lines is fairly uniform with only sparse reflectivity. This is typical of deep water sediments, which are comprised of fine-grained pelagic ooze and clay particles deposited slowly and steadily from suspension. High sedimentation rates on the west coast of B.C.  and  Washington provide a large amount of sediment to the shelf and  continental slope, resulting in episodic dumping of sediments further offshore as turbiditydriven flows deliver material to the nearby basin. On line 85-09, the sediments increase in thickness towards the east to SP 300, areflatlyingbetween SP 300-900 and  dip towards the  continent east of SP 900. At the latter position, where a slight undulation of the seafloor and  associated diffracted arrivals may indicate a submarine channel, onlap of sediments  from the east and an increase in the frequency of layering and apparent coarseness of the sediments towards the east indicate an increase in the quantity of land-derived material (Fig. 3.2). The thickness of individual layers decreases from about 100-200 ms west of SP 900 to less than 50-100 ms east of about SP 1100 in the upper layers. West of about SP 550, line 85-07 has thinner sediments than 85-09, probably because of its somewhat younger age and  consequently smaller accumulation of pelagic sediments; an abrupt increase in thickness  and  frequency of layering in the sediments east of that point is again apparently due to an  increase in land-derived sediments. The interval velocities derived from semblance velocity analysis of the reflection data range from about 1450 m/s (water velocity), appropriate for water-filled, unconsolidated material in the uppermost layer, to as high as 2850 m/s directly above the basement at the base of the thickest sedimentary sequences. The average velocity  Chapter  3. OCEANIC  CRUSTAL  STRUCTURE  OBSERVED  ON MCS  DATA  59  directly above the top of the basement is about 2000 m/s. These values are within the range of velocities appropriate for shaley deep water facies. The  sediments are generally flat lying and undisturbed, except at the eastern end of both  lines. Slumping and discontinuous reflectors from SP 2000-2200 on line 85-09 appear to indicate rapid deposition. The disruption of the layering appears to follow the basement contours in a general sense, with the sediments slumping into relative lows in the basement topography (Fig. 3.5). However, at about SP 2135-2200, for example, the disruption of the upper layers is more cryptic, with haphazard settling of the sediments and some higher amplitude reflectors. The sediments at the eastern end of line 85-07 between SP 225 and 430 also show interesting features, where many diffraction hyperbolae indicate faulting and disruption of the layering (Fig.  3.1). Migration of the section at the water velocity collapses  these diffractions quite successfully (Fig. 3.5), and fault disruptions in the layering can be discerned. At about SP 430, the faulting appears to be due to sediment draping and differential compaction over a topographic high, with the faults 'healing' upwards. Between about SP 300 and 375, the faults have larger offsets for a smaller degree of differential topography, and so such an explanation does not have the same appeal. Bright spots of higher amplitude within the sedimentary section occur at the eastern end of both lines (Figs. 3.5 - 3.6). The  polarity of the brighter spots on 85-09 is difficult to discern  because of the tightness of reflectors; however, the wavelet arising from the water bottom reflector is a doublet comprised of 2 positive and one negative peaks, whereas the bright spot at e.g., ~SP 2195 at 3.7 s has a single positive peak, in contrast with the single negative of the normal polarity water bottom, suggesting reverse polarity. Polarity on the twin bright spots overlying the propagator trace zone on 85-07 is reverse (Fig. 3.6). Preliminary results of an AVO  study on these data indicate that the polarity reversal is probably due to gas  trapped in the sediments (D. Lumley, pers. comm., 1990). The  origin of the disruption of the sediments and higher amplitude reflections is not  Chapter 3. OCEANIC CRUSTAL STRUCTURE OBSERVED ON MCS DATA  60  clear. On 85-07, the bright spots appear to be associated with the topographic high at the propagator location. Davis et al. (1989) show a correlation between basement highs and upwelling of hydrothermal convection cells; they conclude that the controlling factor in location of the cells is the thickness of sediment, and not the basement topography per se. The apparent localisation of gas at the 85-07 PR crossing could be a result of such topographically controlled circulation causing upward flux of gaseous fluids, which then reach an impermeable layer in the sediments. On 85-09, however, the brighter reflectivity is not exclusively associated with topographic highs. At ~SP 2050 there appears to be an association between a topographic high and bright spots; at SP 2167 the association is between possible crustally pervasive faults (see Section 3.2.3) and bright reflectors, suggesting the possible localisation of fluid flow in areas of such faulting.  3.2.2  Top of oceanic basement  The top of the oceanic basement is a prominent feature on the seismic sections, because of the large impedance contrast across this boundary: its reflection coefficient varies between 20%-63% (see Table 3.1, p. 82; Section 3.2.4 gives details of the method of estimation), because of the variability in roughness and composition of the surface. Blocky pillow basalts filled with voids would have a comparatively low reflection coefficient because of their rough diffracting and potentially altered surfaces, while harder massive flows would have a much higher reflection coefficient. The basement dips slightly landward and relief is generally on the order of 100-200 ms TWT (i.e., 100-200 m). It is rough at the eastern end of line 85-07, for example at -SP 550 the relief is 600-700 ms, about 700-800 m (Fig. 3.1). The relief of the basement surface of 85-09 also exhibits several fairly abrupt changes in character, the obvious example being the seamount at SP 700 which rises -1 km above the surrounding basement. Other changes in relief occur at the two line 85-09 crossings of pseudofault  Chapter  3. OCEANIC  CRUSTAL  STRUCTURE  OBSERVED  ON MCS  DATA  61  Figure 3.5: (Following page): a) Sedimentary section from the eastern end of 85-07, corrected for spherical divergence, f-k migrated at the water velocity and plotted with a scalar gain. Note association of high amplitude reflectors with PR trace at about SP 330. Disruptions of the sedimentary layering are also associated with the PR zone, b) Sedimentary section from the eastern end of 85-09. Processing parameters as in a). Note the brighter amplitude reflectors in the sedimentary section between SP 2020-2050 and SP 2100-2200.  f  O  SP  SW  2500-0  375  350  325  NE  300 T-  2500a0  2750D02750„0 £  a.  9 | 6*3  3000nO$ 3000a0  LU Q  3250aO4;  3250a0  3500a0 3500„O  § Co  Q  Figure 3.6: True amplitude prestack depth migration of 85-07, SP 297-381 [provided courtesy D. Lumley; see Lumley (1989) for method]. The bright reflectors at about 2675 m are phase-reversed. ON  Chapter  3. OCEANIC  CRUSTAL  STRUCTURE  OBSERVED  ON MCS  DATA  64  traces. PR4 at about SP 300 (-300 m relief) separates 'normal', highly diffractive basement topography from an anomalously deep, flat and relatively high amplitude region about 20 km  long between SP 360-820 where the relief is at most 30 ms (30 m). PR7 at ~SP 1400 is  associated with relief of about 500 m. The  basement interface appears attimesto be a 'simple' boundary (i.e., the shape of the  wavelet is the same as that of the direct water arrival) which one would expect from a discrete interface. In other cases the wavelet is more complex, probably due to the interference effects of roughness at the top of the basement. In some areas (e.g., Fig.  3.7, SP 200-225), velocity  analyses appear to indicate the presence of a low velocity upper crustal layer about 50 ms thick (about 75-90 m) with velocities of about 3000-3500 m/s overlying average velocities of about 5500-6500 m/s within the crust to the basal reflector. These low upper crustal velocities are consistent with those of highly porous or fractured basalts with the thin cracks and  voids filled in with metamorphic or sedimentary materials, the expected composition of  the top of the igneous crust as it moves away from the ridge (e.g., Purdy, 1987). In any case, these low velocity events occur so close to the basement surface that they tend to be obscured by diffractions arising from that interface; it is therefore difficult to tell whether they are real reflectors, because as noted in Section 2.2, out-of-plane diffractions may have higher than normal apparent velocities to the top of the basement On the other hand, they may  be real, laterally persistent events which appear intermittent because they are obscured  by the diffractions.  3.2.3  Intracrustal reflectors  Much of the effort involved in this research went into a search for intracrustal reflectors. Previous work reviewed in Section 1.3 indicated that reflective interfaces might exist within the oceanic crust. It seemed obvious that any crustal variability or evolutionary trends would  Chapter  3. OCEANIC  CRUSTAL  STRUCTURE  OBSERVED  ON MCS  DATA  65  be more readily apparent if intracrustal reflectors could be imaged. Certain features in the data visible on far offset constant velocity stacks (Fig. 2.15:(b) and (c)), led to optimism that discontinuous crustal reflectors might exist. Further impetus for this effort derived from the fact that a shallow intracrustal reflector had already been resolved on MCS  line 85-03 across  the Juan de Fuca ridge which ties with 85-07 (see Figs. 1.8, 2.1 and Table 1.2) Q3.ohr et al., 1989). Studies of the velocity structure and composition of young ocean crust had led to the prediction that the shallow reflector near the ridge, interpreted as possibly the layer 2A/2B boundary, might shoal and disappear within several My of crustal formation due to infilling of the porous basaltic upper layer with metamorphic material and sediments, and therefore decreasing velocity contrast, as the crust moves awayfromthe ridge (Purdy, 1987). Support for  this hypothesis from MCS  data would have been a valuable result.  After much investigation of the western ends of both 85-07 and 85-09, no shallow reflectors were resolved. Although preliminary results of a bottom-source refraction survey in this area show no indication of the presence of such a reflector (Purdy et al., 1989), other MCS  data recently acquired on a streamline out into the basinfromline 85-03 show a similar  event to 3.5 My (Rohr et al., 1990). In fact, most of the crust on lines 85-07 and 85-09 appears quite transparent acoustically, indicating that transitions in the crust are gradual at reflection wavelengths. However, a number of reflectors have been resolved on the eastern third of line 85-07 (Fig. 3.8), a highly reflective zone in comparison to the rest of the two lines. Dipping events within the crust are also imaged near the eastern ends of both lines and  will be considered later. The  enhanced reflectivity within the crust at the eastern end of 85-07 (SP 101-600) is  apparent on the coherency-filtered stack (Fig. 3.3), which is dominated by coherent energy at this location at a threshold level which shows only sparse coherency within the crust at other locations. Fig. 3.8 should be referred to for the following discussion of intracrustal reflectors. Fig. 3.7 presents a summary of the results of velocity analyses from the reflection  Chapter 3. OCEANIC CRUSTAL STRUCTURE OBSERVED ON MCS DATA  66  data, and is helpful in discerning the continuity of the reflectors. A package of reflectivity between 1.3-1.5 s TWT sub-basement (about 6 s TWT; reflector 'a', Fig. 3.8) occurs at the easternmost end of line 85-07 and is continuous for about 10 km to SP 260, where it becomes difficult to extend because of an apparent bifurcation in the reflectivity. At this point, one event shallows to -1.2 s sub-basement at SP 300, mimicking the basement topography (reflector 'b', Fig. 3.8). The continuity of the events here is poor, but if we assume these events are correlated, the velocity resolution is too poor to definitively ascertain whether the 100 ms difference in time thickness sub-basement between the reflector from SP 101-260 and SP 275-350 is due to a change in upper crustal velocity structure or thickness. The other event dips downward to the west, to -2.1 s sub-basement ('Moho' depth) at about SP 300 (reflector 'c', Fig. 3.8), but becomes difficult to trace in the disrupted zone associated with the propagator between SP 300-450. Within the propagating rift zone, at about SP 350, velocity analyses provide an estimate of about 4800 m/s (±20%) to an event at about 1.2 s sub-basement (reflector'd', Fig. 3.8). At SP 500, a reflective package at about 1.2 s sub-basement appears, and continues to about SP 610 (reflector 'e', Fig. 3.8). The only velocity estimate for this event is 5000 m/s (±20%), about 10% slower than estimates taken further east (Fig. 3.7). On the basis of position (relative to the top of the basement), these events may be correlated with the shallow extension of the 6 s reflector. At about SP 515, a short reflector dips east from about 1.3 s sub-basement to about 1.5 s sub-basement at SP 500 (reflector T , Fig. 3.8), and may arise from a similar interface as the 6 s reflector to the east. Between SP 400-500, the basement surface is less diffractive than at the pseudofault trace and short, shallow reflectors (1 s or less below the top of the basement) are discernable between SP 390-650 (0.75 s sub-basement at SP 390-400 (reflector 'g', Fig. 3.8), 0.8 s sub-basement at SP 425-430 (reflector 'h', Fig. 3.8), 1.0 s sub-basement at SP 450-455  Chapter  3. OCEANIC  CRUSTAL  STRUCTURE  OBSERVED  ON MCS  DATA  (reflector T , Fig. 3.8), 0.8 s sub-basement at SP 470-500 (reflector 'j', Fig. basement at SP 550-560 (reflector 'k', Fig. T,  67  3.8), 0.7 s sub-  3.8) and 0.9 s sub-basement at SP 650 (reflector  Fig. 3.8)). The velocity estimates to two of these events at SP 425 and 450 are among  the best-constrained of the crustal reflectors (within ± 5-10%), and are 5000 m/s and 4700 m/s respectively (Fig. 3.7). A westward dipping event is imaged from 5.8 s TWT  at SP 390 (1.5 s sub-basement) to  6.0 s at SP 475 (1.65 s sub-basement) (reflector'm', Fig.  3.8). Interval velocity estimates to  this event range from 5900 m/s (± 10-20%) from the basement surface (at SP 400) to 6500 m/s (± 10-20%) from an intermediate reflector (at SP 450). Stacking velocities range from 3150-3300 m/s. A number of lines of evidence suggest that the above events are due to reflectivity within the crust rather than noise. They occur at different times in the section than those at which we would expect simple pegleg multiples from the top of the basement. Velocity analyses indicate stacking velocities for these events ranging between about 2000-3500 m/s, which is too high for multiples generated within the sedimentary section, as these would have stacking velocities less than 1700 m/s. It is possible that these events are due to scattered arrivals from lineations at the basement interface out of the plane of the section. However, we would expect linear scatterers in this survey to be oblique to the survey line; such scatterers would produce dipping events rather than the subhorizontal events imaged here. Excluding the reflectors within 1 s of the basement surface (reflector 'g' to '1', Fig. 3.8) and the short dipping reflector at 1.3-1.5 s TWT  sub-basement at about SP 500 (reflector T , Fig. 3.8),  the events are laterally persistent over ranges of 5-10 km, with fairly consistent interval velocities. The frequency content of the crustal reflectors ranges from about 10-30 Hz (Fig. 2.6), with a dominant frequency of 20 Hz,  consistent with the frequency range expected for  signal at these depths. It is difficult to interpret the origin of the crustal reflectivity on line 85-07, because of  Chapter  3. OCEANIC  CRUSTAL  STRUCTURE  OBSERVED  ON MCS  DATA  68  the low S/N, the apparent complexity and lack of continuity of the reflectors, as well as the lack of coincident refraction results. The appearance of the crustal section between about SP  260-600 is that of a number of lenticular bodies of different sizes, particularly in the  lower crust below 1 s sub-basement. The  eastern section of the line shows a highly reflective  lower crustal section, with considerable coherent energy from 6 s downward, as well as some intermittent reflectivity above that. Based on velocity analyses, which correspond roughly to the refraction results further east (Fig.  3.9), the shallow crustal reflectors (<1 s sub-basement) seen on 85-07 with overlying  interval velocities of <5000 m/s may arise from the Layer 2-3 boundary. However, these reflectors are in several cases significantly deeper than 0.7 s TWT refraction results at OBS  sub-basement which the  2 indicate, implying a -0.5 km variation in the Layer 2 thickness if  they arise from the same boundary. Such variability in Layer 2 thickness is observed in both refraction surveys (e.g., Whitmarsh, 1977) and other MCS  results (McCarthy et al., 1988)  (see Table 1.2). The  refraction velocity model for OBS  2 has a higher gradient zone (0.6 s ) at 7500 m _1  depth (about 3750 m sub-basement), corresponding to about 5.9-6 s TWT,  between layers  with gradients of about 0.1 s . This higher gradient zone separates velocity layers of about -1  6700 m/s and 7200 m/s, and is interpreted as the layer 3A/3B boundary. Velocities derived from the refraction model compare well to those derived from analyses of the reflection data (Fig. 3.7), and the 6 s reflector occurs at the same traveltime as the refraction 3A/3B boundary. The limit of resolution of the interface at 6 s TWT  is about 1 km, assuming a  stacking velocity of 3000 m/s. It has a more complex signature at the east end of 85-07. By analogy with the basal reflective sequence (Section 3.2.4), and  considering the geological  model of a Layer 3B comprised of layered cumulates, this complex reflectivity may arise from interlayered materials of contrasting mafic content within the crust. The  westward dipping reflector at 1.5-1.65 s sub-basement between SP 390-475 (reflector  Figure 3.7: Interval velocities derived from stacking velocities of the eastern end of 85-07 superimposed on section f-k migrated at 1450 m/s and bandpassed 5-30 Hz. Reliability of the velocity estimates is coded with type style and size: larger boldface type indicates that the estimates are the most reliable, within ±10%; italics, within ±20%; smaller plain type, worse than ±20%. These error bars are based on the width of semblance analysis peaks, with no independent information, and as ON  such are not exact, but are included to provide an indication of the interpreter's confidence in the results.  ^  Chapter  3. OCEANIC  CRUSTAL  STRUCTURE  OBSERVED  ON MCS  DATA  70  Figure 3.8: a) F-k migration at 1450 m/s of eastern end of 85-07, bandpassed 5-30 Hz. b) Line drawing interpretation of a). Letters refer to reflectors discussed in text.  Chapter 3. OCEANIC  CRUSTAL  STRUCTURE  OBSERVED  ON MCS  DATA  71  'm', Fig. 3.8) probably arises from layering within Layer 3B. It is deeper than the 3A-3B boundary from the refraction model, yet too shallow for the basal reflector. Although the velocities derived from the reflection data are more typical of layer 3A (i.e. less than 7 km/s) they are accurate only to 10% at best, and therefore two-way time is a more reliable indicator of the origin of the reflectivity.  Dipping events within the crust Dipping events at the eastern end of both lines may correspond to other such events seen in the Atlantic (NAT Study Group, 1985; McCarthy et al., 1989; Mutter et al., 1989; White et al., 1990). Such features have been interpreted as either delineating crustally pervasive normal faults induced by crustal failure due to thermal contraction as the newly-formed plate moves away from the spreading ridge, or as the boundaries of axial magma chambers. In the Atlantic, these dipping features are also associated with normal faulting at the basement surface. On line 85-09, between about SP 2100-2250 and SP 1800-1900, there are two occurrences of dipping events which appear to be associated with offsets of the basement (Fig. 3.10). The easternmost-of these reflectors appears to steepen in dip near the basement surface, which is typical of listric normal faults. Two more dipping events occur at the eastern end of line 85-07 between SP 101-250 from 5.3-6.0 s (reflectors 'n' and 'o', Fig. 3.8), but are difficult to trace near the basement surface. Efforts to obtain interval velocities for these eventsfromthe reflection data proved fruitless. Modelling carried out by Andrew Calvert (pers. comm., 1989) suggests that the dipping events seen on 85-09 and 85-07 may not be real. Reflected refractions and refracted reflections arisingfromlinear scatterers at the top of the basement at azimuths different from that of the seismic survey line could produce dipping energy within the section at high apparent velocities. Since the lines of this survey were shot oblique to the presumed structural trends  LAYER  WATER  1  2  3A  3B  MANTLE  Figure 3.9: Refraction velocity model from eastern ends of MCS lines (from Clowes and White, in preparation). Velocities at the top of layers and gradients (brackets) are displayed. MCS two-way times are shown in italics at line boundaries. See Fig. 2.1 for line location. Note increased thickness of lower crustal layer 3B, the layered cumulate zone at the base of the crust. The refraction model intersects a propagating rift pseudofault trace at about 80 km, corresponding with the interpreted thickening of the lower crust.  Chapter 3. OCEANIC CRUSTAL STRUCTURE OBSERVED ON MCS DATA  73  of the basement, one cannot be certain that these events are real. However, since the dipping reflectors on 85-09 are associated with offsets in the basement, they may arise from similar features as the events in the Atlantic, although it is impossible to determine the nature of these offsets. The dipping events on line 85-07 are not visible near the top of the basement. The contact between the dipping events and the base of the crust might also be useful in evaluating their origin: since the base of the crust is assumed to be a fundamental boundary between materials resulting from different processes of formation in the lithosphere, it is unlikely that reflectors would cross-cut it. However, the relationship of the dipping events to the base of the crust is either obscured by the water bottom multiple or unclear, providing no further information.  3.2.4  Basal reflection  Perhaps the most outstanding feature of this dataset is the exceptional clarity and continuity of the reflective package at about 2 s TWT below the top of the basement (Figs. 3.1-3.4). For reasons summarized in this section, this package of reflectors is interpreted as arising from the base of the crust at the 'Moho transition zone' (Fig. 1.4). In general, the basal reflector mimics the basement topography, implying a fairly constant crustal thickness. However, a number of exceptions to this uniformity are associated with propagator traces, indicating that such zones may be characterised by thicker than normal crust. The basal reflector can be traced for almost the entire length of 85-09, a distance of 170 km. It is disrupted for a distance of about 4 km at the pseudofault trace of PR 4 (SP 275-325) and on the eastern side of the seamount between SP 740-820. At about SP 2100 it is obscured by the water bottom multiple because of increased sediment thickness. Results of the refraction survey at the eastern end of the line indicate that the Moho should occur at about 6.95 s TWT (Fig. 3.9). A weak event corresponding to the basal reflection stacks  Figure 3.10: F-k migration using depth-varying velocity function showing dipping events (arrows) at eastern end of 85-09. Data is resampled to 2 ms. Events may be due to reflectivity within the crust or side-scattered noise; however, their association with offsets of the basement surface is consistent with events on other surveys interpreted as crustally pervasive faults.  Chapter  3. OCEANIC  CRUSTAL  STRUCTURE  OBSERVED  ON MCS  DATA  75  in above the water bottom multiple at about 6.8 s between SP 2100-2150 (Fig. 3.2). As it dips slightly eastward to the location of the refraction results, there is close correspondance between the basal reflector and the refraction Moho.  can  At the east end  of 85-07, the water bottom multiple is deep enough that the basal reflection  be traced (Fig.  3.1). In the vicinity of the pseudofault trace of PR 7, between SP 250-600,  its position is difficult to determine because of poor S/N and lack of reflector continuity. A number of weak events may correspond to the basal reflector (events 'p' and 's'-V on Fig. 3.8). Between SP 610-1200, it is strong and  can be traced without difficulty (Fig. 3.1). West  of SP 1200 it becomes extremely weak and difficult to locate. Coherencyfilteredsections are helpful in tracing it (Fig. 3.3). The  basal reflector has a variable seismic signature. The wavelet is often doubled or  tripled, indicating a geologically complex boundary. The thickness of the reflective package varies from about 150-300 ms in 'normal' crust away from PR traces and its frequency content is about 10-30 Hz,  with a dominant frequency of about 20 Hz (Fig. 2.6). Assuming  an average velocity to this interface of about 3500 m/s, the limit of spatial resolution on stacked data is on the order of 1.25 km.  Estimates of interval velocities through the igneous  crust to this reflective package vary from about 5800 m/s to 6500 m/s. The  refraction models  indicate velocities of about 6500 m/s through the crust. Collins et al. (1986) have shown for the Bay  of Islands ophiolite that the 'Moho' reflection  may arise from different layers within the lower crust, but is most likely generated at the Moho transition zone, i.e., the layered mafic and ultramafic material at the base of the crust (Fig.  1.4). The coincidence of the base of the reflective package with the refraction Moho  at the eastern end of 85-07 supports the idea that the top of the reflective sequence coincides with the layered ultramafic sequence at the base of the crust. If we assume that the bottom, rather than the top, of the basal reflective sequence corresponds to the seismic Moho, the total change in thickness of crust over line 85-07, excluding the propagating rift (PR) pseudofault  Chapter  3. OCEANIC  CRUSTAL  STRUCTURE  OBSERVED  ON MCS  DATA  76  trace area, and assuming an average velocity through the crust of 6000 m/s, is about 600 m. If we interpret the basal reflector's position at the PR trace to be at 6.35 s (reflector 'p', Fig. 3.8), there is 600 m of increased thickness, assuming constant velocities within the crust. If it is at 6.75 s (reflector 't'-'v', Fig. 3.8), the increase in thickness is 1800 m. A zone of reflectivity between 6.5-7.1 s TWT,  (i.e., beneath the 'normal' base of the crust)  from SP 300-900 on line 85-09 is interpreted as an indication of extensive underplating of the pre-existing oceanic crust (see Section 3.4 and  Figs. 3.4 and 3.17). The change in thickness  over 85-09 is about 1.75 km if the underplated region has a velocity of 7.3 km/s. Excluding the underplated region and  the PR trace area, the crustal thickness on line 85-09 varies about  600 m. This range excludes the seamount, which is interpreted as having a nearly normal crustal thickness beneath the volcanic edifice (Section 3.3). In the areas near PR traces, the base of the crust appears to have a different character than at 'normal' crust elsewhere. In all cases, the PR crossings are associated with dipping events at the base of the crust. At the 85-09 PR 7 crossing (about SP 1370-1410), the brightest segment of the basal reflection on either line occurs directly below the surface location of the PR trace. To the west of this bright spot, between SP 1350-1400, a westward dipping event is observed beneath the basal reflector which appears to be continuous with it. At about SP  1130-1240, a lenticular reflective body is observed beneath the basal reflector. At the  85-07 PR 7 Crossing, the basal reflection is obscured by noise from the top of the basement, however, strong dipping reflectors occur on both sides of the PR zone. These reflectors, at SP  275-300 (Fig. 3.11, and reflector 'q', Fig. 3.8) and 500-600 (reflector V , Fig. 3.8),  are well focussed at high stacking velocities, do not appear to crosscut the basal reflection and  seem to be continuous with the basal reflection on the outer sides of the PR zone. At  the 85-09 crossing of PR 4, we interpret a similar dipping reflector as the western limit of the zone of crustal underplating, which extends east beyond the seamount and  is bounded at  about SP 940 by a similar westward dipping reflector (Calvert et al., 1990). These features  1  km  Figure 3.11: Dipping reflector east of pseudofault trace of propagating rift 7 on 85-07 (dots); CMP stack is f-k filtered between apparent velocities of ±2000 m/s and bandpassed 5-35 Hz. W: water bottom multiple; PT: pseudofault trace.  Chapter 3. OCEANIC CRUSTAL STRUCTURE OBSERVED ON MCS DATA  78  will be discussed further in Section 3.4.  Reflection coefficient (RC) estimates Estimation of RCs at the basement surface and the top of the layering beginning at 2 s TWT sub-basement have been made at several locations along 85-09, in an effort to constrain the material properites across the 2 s TWT layering. The technique employed followed that of Anstey (1977) for calibration of the RC to that of the seafloor: first, the RC of the seafioor was estimated by comparing the measured amplitude of the seafloor and itsfirstmultiple on a number of near offset traces resampled to 2 ms, corrected for geometrical divergence at the water velocity and stacked (4-fold) to improve the signal/noise. (The RC of the seafloor is the negative ratio of the first multiple to seafloor amplitudes). The data were then corrected for geometrical divergence and transmission loss using reflection velocity estimates. Reflection coefficients of the basement surface and the 2 s TWT layering were derived by comparison of reflection amplitudes with that of the water bottom. Computations of attenuation loss in the sediments assuming a dominant frequency of about 30 Hz for a range of values reported in the literature (70, 150, and 300) were included in the RC estimates, but resulted in <1% difference in the results for the largest RC at the 2 s layering. Computations of the acoustic impedance for a 2-layer case were carried out using different possible rock density contrasts as measured by Salisbury and Christensen (1978) from the Bay of Islands ophiolite (Fig. 3.12). Both crustal and mantle density values were used in calculation of the lower layer velocities. The lowest V\ values correspond to the average velocity through the crust at that location, and thus provide minimum bounds. A value of 7.7 km/s was chosen as the maximum expected value for crustal velocities (Figs. 3.9 and 3.12). The velocity contrast was also calculated for a Vi of 7.3 km/s, corresponding to the average lower crustal velocity from analysis of the reflection data at the eastern end of 85-07 (Fig. 3.7).  Chapter  3. OCEANIC  UTHOLOCT  CRUSTAL  STRUCTURE  OBSERVED  ON MCS  q/cc 26  1  28 3.0 ' I'  DATA  79  Vp , km/»«c 32 3.4  1 '1 '1 1  50  60  ' | ' ' l I ' i t i i [ i T  • i  70  80  i I i r i i l [ i I I I l i I I I | I I I II  \\  B«CC. DIKES SMCCTtO  -OtKES  (WM0OLUlTEl  _ Wfi. MfTMAMMO  N  — s^i  moxoic SiBBHO  £5 CL  tu  OUVIKt  o  \  i*88K>  \  \ \  t*4 6 -rwcrouTt  JL  " UURA-  10  i  . i i  i  , i 1J  1  »' ' ' ' i  i  t i i i I  i  i i  i • i i i i  I  i i , .  . i i i  ,  I  i i  11  i  Figure 3.12: Envelopes of compressional wave velocity (Vp) and density (p) versus depth for the Blow-Me-Down massif from Salisbury and Christensen (1978). Heavy curves represent best fit to data. Dashed lines between 2.6 and 3.8 km indicate discontinuous velocity inversions. Velocities shown at depths less than 1.3 km represent maximum velocities.  Chapter 3. OCEANIC CRUSTAL STRUCTURE OBSERVED ON MCS DATA  80  TSN  r'5.5  6o0  6o5  6a0  6  D  5  7  D  0  Figure 3.13: Examples of data used for RC estimates, line 85-09. a) SP 222-270: RC measurement was made at SP 247 (arrow). Traces are stack of inside traces of 6 adjacent CMPs, resampled to 2 ms, and corrected for spherical divergence and transmission loss to the basal reflector at 6.15 s (bullseyes). TSN = trace sequntial number, b) SP 1361-1412: RC measurements made at SP 1382-1384 (arrow). Basal reflector is at 6.35 s; display as in a). This is the location of the highest amplitude basal reflector on both MCS lines.  Chapter  3. OCEANIC  CRUSTAL  STRUCTURE  OBSERVED  Table 3.1 shows a range of values for the RC  ON MCS  DATA  81  of both the top of the basement and the  basal reflector (see Fig. 3.13 for examples of the deeper section of the data). Although the accuracy of the estimates is probably ±30%,  there is a clear variability in the results.  Estimates were taken at SP 399-400 to get an idea of the RC in what were assumed to be optimal conditions for this technique: the flat sediments and basement reflector would provide the least diffracting surfaces on the line, without complicated ray  paths. The average  of the results is about 4.5%. The velocities for the lower layer shown in Table 3.1 are all less than 7.6 km/s, i.e., lower than mantle velocities, assuming a mantle-type density of 3300 kg km . -3  mantle (Fig.  If we assume a layer density of 3000 kg km , -3  which is unreasonably low for the  3.12), we require Vi < 7.3 km/s to produce crustal velocities (i.e., < 7.8 km/s) in  the lower layer. This suggests that velocities on the order of 7.4-7.8 km/s are possible in the deeper underplated zone (-6.5-7 s; see Section 3.4). Although suitable traces were not found upon which to measure the RC of the underplated zone, Calvert et al. (1990) report RCs  of  about 1-2% for these features, which would indicate velocity increases on the order of 100 m/s. Although the resolution of these data is limited, the RC results support the hypothesis that the basal reflector arises from layering at the base of the crust, with the increase in velocity to mande value occurring in stepwise fashion rather than as a single discrete jump. The  sub-'Moho' reflectivity in this location could delineate the required velocity increase to  mantle values. SP 247 was chosen for RC  estimation because of the relatively high amplitude reflector  at the base of the crust (Fig. 3.2). The top of the crust is quite diffractive, with a measured RC of only 20%. Crustal velocities are about 5800 m/s and  the RC at the base of the crust is  7-8%. The results in Table 3.1 indicate that assuming a density of 3300 kg km , -3  typical of  the shallow mande (Fig. 3.12), it is possible to produce a velocity as great as 7.9-8 km/s in the lower layer if the lower crustal velocity is 7.3-7.7 km/s. At this location, the transition from crustal to mantle velocities may occur as a discrete jump.  82  Chapter 3. OCEANIC CRUSTAL STRUCTURE OBSERVED ON MCS DATA  Table 3.1: RC estimates (%) from line 85-09. Shot Point  RC RC (Basement) ('Moho')  247  20  6.2 s: 7-8  399  32  6.2 s: 4  400  38-43  6.2 s: 4.5  1382  62  6.3 s: 11  1383  60  6.3 s: 14  1384  50  6.3 s: 10  Vi (km s~ ) V (km s- ) V (km s- ) p=2.8-2.95 p=2.9-3.0 p=3.3 5.8 6.4-6.6 5.7-6.0 7.3 8.1-8.3 7.6-7.9 7.7 8.6-8.8 7.5-8.0 6.1-6.2 5.8 5.3-5.6 7.3 7.6-7.8 6.7-7.1 8.1-8.2 7.7 7.1-7.5 5.8 6.1-6.3 5.4-5.7 6.8-7.2 7.3 7.8-7.9 7.7 8.2-8.4 7.2-7.6 • 6.1 7.4-7.5 6.5-6.8 7.3 8.8-9.0 7.7-8.1 7.7 9.3-9.4 8.2-8.6 6.1 7.8-8.0 6.9-7.2 7.3 9.3-9.5 8.2-8.7 7.7 9.9-10.0 8.7-9.1 6.1 7.2-7.3 6.3-6.7 7.3 8.6-8.8 7.6-8.0 7.7 8.0-8.4 9.1-9.3 1  l  2  1  2  Estimated values of RCs used to derive bounds on velocity contrast across layering at the base of the crust. V i is the estimated velocity through the oceanic crust: the lowest values correspond to stacking velocities from the reflection data; intermediate and highest velocities correspond to the interval velocity of the lowest crustal layer derived from reflection data at the eastern end of line 85-07 and the maximum crustal velocity value measured on rock samples from the Bay of Islands ophiolite (Fig. 3.12), respectively. Acoustic impedance values for the crust above the basal reflector, computed for likely bounds of crustal density values (p=2.8 and 2.95; see Fig. 3.12), were used with RCs to derive lower layer acoustic impedances. V 2 is the estimated velocity range for both crustal and mantle density values (p=2.9-3.0 and 3.3, respectively; see Fig. 3.12).  Chapter The  3. OCEANIC  CRUSTAL  remaining RC  estimates were made at the location of the highest amplitude basal  reflector. RCs  STRUCTURE  OBSERVED  ON MCS  DATA  83  of 10-14% were measured at adjacent shot point locations. Table 3.1 shows  that to arrive at mantle velocity values of around 8.2 km/s, assuming a density of 3300 kg km , -3  the lower crustal velocities must be between 7300-7700 km/s. If the upper layer has  a density of 2800 kg km (about the minimum possible value for the crust (Fig.  3.12)), the  -3  density of the lower layer must also be greater than about 3100 kg km  -3  to produce plausible  velocity-density relations, indicating a strong density contrast at the base of the crust. This would require juxtaposition of less mafic material with ultramafic material at the base of the crust. In the more likely event that the upper layer density is about 3000 kg km , -3  a density  of at least 3200 kg km would be necessary in the lower layer to produce velocities < 8.2 -3  km/s,  suggesting that the lower layer is comprised of mantle material (Fig. 3.12). Because of the difficulty of identifying the basal reflector on unstacked data, the RC  estimates were made at the locations of highest amplitude basal reflections except for the SP 399-400 locations. Since only the locations of highest amplitude reflectors can provide the acoustic impedance jumps required for a discrete jump in velocity from crustal to mantle values, it follows that in general, the crust-mantle boundary in this area occurs in a stepwise fashion, supporting the proposition that the 'Moho' reflection arises, in general, from layering at the base of the crust (Collins et al., 1986).  3.3  Seamount  A small, previously unmapped seamount was discovered at about SP 700 on line 85-09 (Fig.  3.2 and 3.14). It rises about 0.5 km above the surrounding seafloor and 1.1 km above  the basement, assuming velocities of 1.46 and 1.8 km/s for the water and sediment layers respectively, and is about 8 km wide at its base. An 800 m diameter depression of about 50-100 m relief at the top of the seamount is typical of the calderas of young seamounts,  Chapter and  3. OCEANIC  CRUSTAL  STRUCTURE  OBSERVED  ON MCS  DATA  84  suggests a volcanic origin (Fig. 3.15) (Fornari et al., 1988). The smoothness of the  summit plateau may be due to sheet lava flows. A package of subhorizontal reflectors about 0.25 s thick and 2 km long directly beneath the peak of the seamount at 4.5 s TWT  (about  3 km sub-basement), at a similar depth to the high velocity anomaly detected beneath Jasper seamount by Hildebrand et al. (1989), is interpreted as a possible cumulate sequence formed at the base of a magma chamber (Calvert et al., 1990). The  basal reflector is continuous from the pseudofault zone at SP 300 to SP 740 on the  northeast side of the seamount (Figs. 3.4 and 3.14). The apparent break in reflectivity at SP  670 is due to the velocity pull-up effect of the near-surface topography which results  in a 1 s shallowing of the reflectivity. Sinceflexureof the crust is not observed, the basal reflector is probably horizontal in depth; the observed pull-up would then require a velocity of about 2800 m/s within the volcanic edifice if the normal crustal thickness exists beneath the seamount (equating one-way travel times we have: 1100 m/(seamount velocity) + 0.25 s (velocity pull up) = 0.34 s (water) + 0.30 s (sediment), giving the result). Between SP 740-820 the subhorizontal reflectivity is absent. Instead, southwest dipping reflectors, present at SP 750 (6.2 s) and SP 770 (7.0 s) to SP 830 (6.4 s) (Fig. 3.14), are interpreted as the eastern termination of the underplated region which has its western limit at PR 4. These features are discussed in more detail in the next section.  3.4 The  Crustal structure of pseudofault traces of propagating rifts (PR) MCS  lines cross PR's 4 and 7 as identified in the model of Wilson et al. (1984) (Fig.  1.9). PR 4 is crossed once, by line 85-09, at about SP 300. PR 7 is crossed twice, by 85-09 at about SP 1380 and 85-07 at about SP 350 (Figs. 3.1-3.2). The crustal ages across PR  4 differ by about 1 My; across the 85-07 and 85-09 crossings of PR 7, they differ by  about 0.5 and 0.25 My,  respectively. The PR zones are remarkable in every case: on 85-09,  Figure 3.14: Stacked section, migrated at water velocity, of seamount on line 85-09. Beneath seamount at 4.5 s, a 2 km long package of subhorizontal reflectors (FMC) may be a fossil magma chamber. Crust-mantle transition zone (CMT) is "pulled up" beneath seamount due to velocity effects associated with it. Dots delineate westward-dipping reflectors discussed in text.  X  Ui  Chapter  3. OCEANIC  CRUSTAL  STRUCTURE  OBSERVED  ON MCS  DATA  87  PR 4 is associated with the boundary between a region with an anomalously flat basement surface to the east from 'normal' crust to the west; at PR 7 on 85-09 we have the highest amplitude basal reflection (Section 3.2.4); near PR 7 on 85-07 the crust is highly reflective, the sediments are disturbed and there is a high degree of topographic relief. At all locations there are indications of reflectivity below the apparent 'normal' base of the crust. At the 85-09 crossing of PR7,  at about SP 1380, the basement reflection is very bright,  with a reflection coefficient as high as 63%, implying p-wave velocities of 5.5-6 km/s, which probably indicates a massive lava flow. The basement is also topographically relatively high, about 125 m above the surrounding top of the crust. A feature at SP 1350 may be a volcanic flow disrupting the surface of the crust (Fig.  3.16). The  basal reflector is also of anomalously  high amplitude (10-15%) below the high amplitude area of the basement, and is 2.2 s  TWT  sub-basement, indicating a crustal section slightly thicker than average. Velocities through the crust to the basal reflector are about 6100 m/s, within the expected range. Within the limits of resolution of these data (about + 500 m/s at the base of the crust), it is not possible to determine whether the apparent crustal thickening is real and not the effect of slower intracrustal velocities. However, the flow features and high topographic relief are consistent with magma generation after crustal emplacement, which could increase the crustal thickness. As  discussed in Section 3.2.4, the high amplitude basal reflector indicates the presence of  high density material at the base of the crust, suggesting a first order transition to mantle material. However, at about the western limit of the high amplitude basal reflector, an event continuous with the basal reflector dips west between SP 1400-1333. Between SP 1165-1230 a lenticular body appears to be delineated at the base of the crust (Fig.  3.16). These events  may be due to subcrustal reflectivity since they stack in at high velocities (e.g., 3500 m/s), and  do not appear to crosscut the base of the crust (at SP 1250, an event which does seem to  crosscut the basal reflector does so at a greater angle than the other events, and with dipping noise from the basement surface).'  is continuous  Chapter At  3. OCEANIC  CRUSTAL  STRUCTURE  OBSERVED  ON MCS  DATA  88  the 85-07 crossing of PR7 more dipping events occur at the base of the crust, as  discussed in Section 3.2.4 (reflectors 'q' and r \ Fig. c  3.8). These events have high stacking  velocities, around 4000-4500 m/s, giving interval velocities of about 7000 m/s from intermediate crustal reflectors, and do not appear to crosscut the base of the crust. The deeper crustal structure below the disrupted sediments appears lenticular in cross-section (Fig. 3.8). Looking carefully at the section, one may discern possible reflectivity between about 6.0-6.25 s and again between about 6.8 and 7 s at SP 350 (events 'p' and 't', Fig. 3.8). Velocity resolution of the shallow event is very poor (±20%), however, the best estimate (i.e. 6900 m/s) is somewhat lower than the velocity estimates from further east (> 7000 m/s), suggesting that the lower event is a more likely candidate for the base of the crust. The  85-09 crossing of PR 4, at about SP 300, coincides with a transition from relatively  thick crust (2.2 s TWT)  with topographic variation on the order of 200 ms on die southwest,  to anomalously smooth (< 30 ms variation over about 20 km), slightly thin (1.9 s) and deep crust to the northeast (Fig. 3.2). The basal reflector is disrupted over about 5 km at the PR location (SP 295-360). However, over that same range, a dipping reflector is imaged which is continuous with both the basal reflector on the young side of the PR (6.3 s at SP 250) and the base of a zone of reflectivity below the normal base of the crust (7.1 s at SP 360). This subcrustal reflectivity extends about 40 km to the east beyond the seamount to SP 900 where corresponding subcrustal reflections dipping in the opposite sense are observed (Fig.  3.4 and  3.17). The continuity of the dipping reflector at SP 250-360 suggests that the subcrustal complex was formed at the same time as the crust-mantle transition on the young side of the PR zone, implying a large volume of melt was supplied to the PR tip, forming sills both ahead and off-axis of the rift. Subcrustal reflectivity associated with seismic velocities of 7.4-7.8 km/s, detected around Oahu in a coincident reflection and refraction experiment, was interpreted as a sill complex underplated to the oceanic crust as a result of hotspot  Chapter  3. OCEANIC  CRUSTAL  STRUCTURE  OBSERVED  ON MCS  DATA  89  volcanism (ten Brink and Brocher, 1987). Since the reflectivity at -6.2 s between SP 360600 appears typical of the normal basal sequence, the deeper reflectivity may be analogous to the structure observed in Oahu, and  hence due to extensive underplating of the preexisting  crust associated with both the PR trace and the seamount. Correlation of apparent massive lava flows (reflection coefficients of 40-50%, implying velocities of 4.5-5 km/s) with the region of underplating suggests that surface volcanism also occurred as a result of the large magma supply (see Section 4.3.2). Coherency filtered stacks (Figs. 3.3-3.4) show other dipping events elsewhere on both lines, particularly on the western half of 85-07. However, in all locations, they have greater dips than the events associated with the PR traces. It appears likely that these events are diffractions, probably from out of the plane of the section. Furthermore, the basal reflector is assumed to correspond to a significant compositional boundary. It is unlikely that reflectivity due to composition changes or pervasive faults would crosscut it. The  dipping reflectors at the base of the crust associated with the PR zones may be  symmetrically disposed about the PR zone (85-07 crossing of PR 7, Fig. 3.8) or may coincide with the PR crossing (both PR zones on 85-09, Figs. 3.16-3.17), dipping towards the younger (PR  7) or older (PR  4) crust. In all cases, there appears to be some thickening or  underplating of the crust in the vicinity of the PR trace: at PR 4, this underplating is about 0.5 s TWT,  equivalent to about 1.75 km if the velocity is 7.3 km/s, a substantial crustal  thickening if the average thickness is 6-7 km. zone extends for about 40 km,  If our  interpretation is correct, the underplated  and therefore indicates a substantial change in the supply of  magma to the spreading ridge, even if it has only a limited out-of-plane extent. In the other PR  locations, such thickening is not as extensive^ At PR 7 on 85-07, the crust may be as  thick as in the PR 4 area if the base is at 7 s (i.e., about 1.75 km), but it persists for about 18 km and is approximately centered on the PR trace. The PR 7 crossing on 85-09 is more difficult to interpret, because of high amplitude diffractions from the top of the basement;  Chapter 3. OCEANIC CRUSTAL STRUCTURE OBSERVED ON MCS DATA  90  if the low angle dipping events beneath the crust are reflectors, a less extensive amount of thickening is indicated, perhaps as much as 1.5-1.75 km, but occurring as lenticular features 5 km wide. There is an interesting correspondance between the reflection results and the refraction velocity model from further north in the basin (Figs. 2.1 and 3.9). The refraction line crosses the PR 7 trace near its interpreted termination (Wilson et al., 1984), where the age difference across the PR trace would be about .75 My. Near the same location the refraction model indicates a thickening of oceanic crustal Layer 3B (White and Clowes, in preparation), consistent with the interpretation of the MCS lines presented here. Dipping reflectors at the base of the crust and crustal underplating have been observed in other locations in association with new crustal formation. Mutter et al. (1989) observed a reflector dipping towards the ridge near the base of young oceanic crust at the East Pacific Rise, which they interpret to map the region of mass deficiency in the uppermost mantle. The initiation of rifting in the Atlantic appears to have been accompanied by extensive volcanism at the base of the crust (Mutter et. al., 1988). Groschel-Becker and Rosendahl (1990) report upper mantle-level low angle dipping reflectors in association with oceanic fracture zones. White et al. (1990) observe material with refraction velocities of 7.3-7.6 km/s beneath the interpreted crustal section at the Blake Spur Fracture Zone. The association of subcrustal reflectivity and PR pseudofault traces will be discussed further in Chapter 4.  3.5  Summary  The two MCS lines presented in this research compare young crust of 1.4-5.3 My in age formed at different locations along the Juan de Fuca spreading axis. The most obvious feature of the seismic images is the high degree of variability in reflectivity within the crust; the eastern third of 85-07 is highly reflective with at least two intracrustal reflectors, interpreted  Chapter  3. OCEANIC  CRUSTAL  STRUCTURE  OBSERVED  ON MCS  DATA  91  here as the Layer 2-3 and 3A-3B boundaries. In general, the rest of the line lacks intracrustal reflectors. This is true for most of 85-09, although possible dipping intracrustal reflectors are observed near the eastern ends of both lines. The basal reflector is strongly developed along most of 85-09, in contrast to 85-07, where it is often difficult to trace, particularly on the western half of the line. Although in general, the crustal thickness is fairly constant, events are observed beneath the normal occurrence of the basal reflector which are interpreted to indicate large changes in crustal thickness in the vicinity of PR traces. In all cases, the PRs  are associated with distinctive structures. These issues will be considered further in the  following chapter.  Figure 3.16: a) Stacked section of PR 7 (PT, at about SP 1390), line 85-09. Crustal ages are about 4.25 My to the northeast and 4.0 My to the southwest (young side) of PR trace. Highest amplitude basal reflection occurs at depth within approximate limits of resolution of pseudofault location from surface magnetic data (+1.5 km); subcrustal reflector continuous with high amplitude basal reflector dips southwest toward younger crust; other dipping reflectors (arrows), as well as subhorizontal subcrustal reflectivity delineate subcrustal thickening. PT : pseudofault zone; W : water bottom multiple, b) Detail of base of crust (following page). Dots delineate subcrustal dipping reflectors discussed in text.  Figure 3.17: Stacked seismic data across PR4, at about SP 300, 85-09. The crustal age increases from 2 My on the southwest to 3 My on the northeast side of the pseudofault zone. Southwestward termination of smooth basement topography (4.3 s) occurs at SP 350. Reflective crust-mantle transition zone (CMT) is broken, or weakened across pseudofault. Deeper reflection (dots) dips beneath older crust from 6.5 s at SP 280 to 7.1 s at SP 360 and then becomes horizontal about 0.7 s beneath top of original crust-mantle transition.  Chapter 4  DISCUSSION AND CONCLUSIONS  4.1  Introduction  The data involved in this research furnish new information on two areas of oceanic crustal research. First, they provide an example of the seismic character of young oceanic crust derived from an intermediate rate spreading ridge. Little MCS data have been acquired in oceanic basin settings, and any new information is therefore of great interest. Second, these are probably the first MCS data collected across propagating rift zones. As such, they provide an image of the initiation of the oceanic rifting process, as well as the most detailed subsurface picture yet available of rift propagation. The following sections will discuss the implications of both these aspects of the data.  4.2  General crustal characteristics  Recent studies reviewed in Chapter 1 indicate that the oceanic crust, rather than having a one-dimensional structure, is highly variable. Crustal generation is differentiated in terms of both distance from the spreading axis (crustal age) and location with respect to the ends of spreading center segments. The seismic data examined in this study support this depiction of the oceanic crust. Although the base of the crust is generally located at about 2 s TWT sub-basement, the crustal thickness varies significantly, and may be as much as 25% greater than 2 s. Explanation of this order of variation by a change in crustal velocity structure would require the assumption of unreasonable velocity values, indicating that the supply of magma to  95  Chapter  4. DISCUSSION  AND  CONCLUSIONS  96  the ridge can vary considerably. Variability in the processes of formation and modification of the crust is expressed in a wide range of reflectivity, both throughout the crust and in the basal reflector. The reflection coefficient of the basement surface varies from 20-63%, indicating a range of upper crustal velocities from 2.5-6 km/s, characteristic of material ranging from highly porous, altered pillow basalts to massive flows. There is no obvious periodicity in the eruptive style. The typical diffractive basement surface is interrupted for any distance in only two cases by materials with different characteristics: the apparent massive lava between SP  350-810 on 85-09 which appears related to coincident crustal underplating, and at SP  1600-1830 on 85-09, characterised by a hummocky surface with an average of 25 m relief. For  the most part, the crust is transparent, indicating that transitions in physical properties  or composition within the crust are mainly gradual at reflection wavelengths. However, the eastern section of 85-07 is highly reflective. Where data characteristics enable its calculation, the reflection coefficient of the basal reflector (crust-mantle transition) varies by a factor of three, from 4 to 14%. This data set is somewhat inconvenient to interpret because of its oblique orientation to the spreading axis, so that the lines do not provide a direct comparison of either crust formed at different times from the same point on the spreading center axis, or crust formed at the same time from different locations on the axis. However, it is evident from comparison of the two lines, mainly formed from different ranges or segments of the spreading axis, that crustal structure also varies along-axis. One indication of this variation is the difference in the clarity of the basal reflection between 85-07 and 85-09, particularly in crust less than 2 My old OPig- 3.1-3.2). Furthermore, crust formed near the ends of the spreading axis at propagating rift tips may be as much as 25% thicker than 'average' crust formed elsewhere on the axis, whose thickness varies only about 10%.  Chapter 4.2.1  4. DISCUSSION  AND  CONCLUSIONS  97  Basal reflection  As noted in Chapter 3, the reflection at the base of the crust is remarkably continuous along both lines and is disrupted only in extreme situations, e.g., at the locations of two of the PR traces, and  at the eastern end of the seamount. The  reflection waveform varies from 200-500  ms in extent and is doubled or tripled, with reflection coefficients estimated at from 4-14%. It was shown in Section 3.2.4 that although the basal reflector may at times result from a discrete jump from crustal to mantle velocity values, the transition from crustal to mantle rock types is probably most often accomplished in more than one velocity increase, as the synthetic models derived from rock sample velocities and structural relations of the Bay of Islands ophiolite suggest (Collins et al., 1986). The  fact that the basal reflector is virtually continuous across crust of 1.4-5.3 My supports  the concept of a crust-mantle boundary which is formed early in the cooling history of the crust, and arises from a fundamental boundary in the oceanic lithosphere [Rohr et al. (1988) observed 'Moho' reflections within 12 km of the Juan de Fuca ridge axis on 85-03]. The  most intriguing aspect of the seismic images is the occurrence of several examples of  dipping sub-'Moho' reflectors which are continuous with subhorizontal reflectivity below an already apparently 'typical' and continuous basal reflection. These features are associated with pseudofault traces of propagating rifts, and  will be discussed extensively in Section 4.3.  4.2.2 Intracrustal reflectivity Although there are several instances of possible intracrustal reflectivity on line 85-09, the only well-established, continuous reflectors are imaged at the eastern end of 85-07. This intracrustal reflectivity is also observed on the western end of line 85-02, the landward extension of 85-07 (Fig. 2.1), for at least 10 km (Hyndman et al., 1990; Yorath et al.,  Chapter  4. DISCUSSION  1987). The  AND  CONCLUSIONS  98  intracrustal reflectivity is somewhat discontinuous, with reflectors at different sub-  basement TWT.  Because of the lack of coincident refraction, we lack the velocity resolution  to discern whether this variation is due to changing velocity structure in the crust, or to changing thicknesses of the crustal layers. However, the reflectivity appears to comprise a number of bodies which are variable in size and lenticular in section (Fig. 3.8), similar to a model of crustal structure comprised of different magma bodies intruded into previously frozen crust, leading to complex structural relations between them (Fig. 4.1) (Malpas et al. (1990). Reflectivity from interfaces between the bodies might arise at many different depths in the crust while this variability could remain undetected at refraction wavelengths if the intrusions had similar velocities. The  interpretation of the reflectors presented in Chapter 3 indicates variation in thickness  of both Layers 2 and 3 near the PR zone. Layer 2 may vary by 0.5 km over a distance of several km.  The apparent crustal thickening at the 85-07 PR 7 crossing may be accomplished  by thickening of the cumulate Layer 3B in a lenticular body approximately symmetrical about the surface location of the PR trace (Fig. 3.1, 3.3 and 3.8). The crust on the old side of the PR zone from SP 240-300 appears to have been rotated upward from east to west toward the PR  zone. The intracrustal reflectivity at 5.7-6 s follows this trend while deeper reflections  from 6-7 s between SP 240-300 dip to the west, hence the interpreted thickening of Layer 3B. The  basement surface appears to have a number of step-like discontinuities, (e.g. at SP 250,  275, 290, 310 and 320) along which crustal uplift might have occurred. The flat basement surfaces between these discontinuities may denote the occurrence of surface volcanism in the form of massive lava flows. This may be an example of reactivation of old abyssal structures by the PR which is seen in the Galapagos PR zone (Kleinrock and Hey, 1989). The  area of increased crustal reflectivity on 85-07 is associated with the PR 7 pseudofault  trace. Although this may be simply a coincidence since increased reflectivity is not obvious at the other PR locations, it may indicate a difference in the process of crustal generation or  Chapter  4. DISCUSSION  AND  CONCLUSIONS  99  PILLOW LAVAS  Figure 4.1: Crustal model based on Troodos ophiolite structural relations (CY-4: drill hole penetrating 2263 m from sheeted dikes to Layer 3-type crustal structure; CY-4 does not penetrate to deformed tectonites of upper mantle. These are observed at surface in a structurally uplifted part of the ophiolite complex). The lower crust is comprised of various magmatic bodies which have intruded into previously frozen crust. (From Malpas et al., 1990).  Chapter  4. DISCUSSION  AND  CONCLUSIONS  100  alteration at this location due to the PR process. If older structures have been reactivated in this area, there may be a higher degree of fracturing within the crust and hence greater circulation of hydrothermal fluids, leading to alteration and reflectivity at this location.  4.3  Propagating rifts (PR)  4.3.1 Characteristics As noted in Section 3.4, the pseudofault traces of PR crossed by the MCS  lines are associated  with anomalous basement surface topography. In two of three PR locations, the normal basal reflector is disrupted. However, the most intriguing aspect of the PR zones is their association with both reflectors which dip below the normal base of the crust and reflectivity. These events suggest that the areas of PR  subhorizontal subcrustal  traces may in general be associated  with increased crustal thickening or subcrustal underplating. In a detailed study of the bathymetry and physiography of the Galapagos PR, Kleinrock and  Hey (1989) identified the sequence of tectonic activity which accompanies the develop-  ment of such seafloor spreading. The length of the Galapagos PR from initial faulting and disruption of the crust to the initial occurrence of full rate spreading at the axis is about 21 km. This zone measured parallel to the rift becomes the pseudofault zone as it moves away from the axis. Initial surface volcanism occurs about 6.5 km behind the initiation of rifting as basalt erupts through faults andfissuresassociated with extensiortal faulting which creates a rift tip depression 500 m deeper than the 'normal' rift behind it. The high degree of differentiation and diversity of composition found in many basalts erupted at the PR tip have been interpreted as requiring numerous single episodes of mixing in isolated magma chambers (Christie and Sinton, 1986). Although the magma compositions and the rift tip depression have been attributed to a limited supply of magma to the PR tip (Christie and Sinton, 1981; Kleinrock and Hey,  Chapter  4. DISCUSSION  AND  CONCLUSIONS  101  1989), recent evidence, including the crustal images from this research, indicates otherwise. First, Christie and Sinton (1986) find that the magmatic sources for the Galapagos PR tip are shallow (25-35 km) compared to normal rift sources, indicating an enhanced magma supply or an increased thermal budget, perhaps associated with the Galapagos hotspot. Second, the rift tip depression has been explained as the result of viscous head loss due to inhibition of upward asthenospheric flow through the colder lithosphere beneath the PR tip (Phipps Morgan and Parmentier, 1985). However, inversions of gravity data suggest that the crustal thickness is constant over the PR, implying that the topography is dynamically created and thenflexurallysupported as it freezes into the lithosphere (Phipps Morgan and Parmentier, 1987). Finally, the seismic images of PR pseudofaults suggest a thickness on the young side of PR's 4 and 7 on 85-09 at least as great (2.2 s TWT)  as that of the undisturbed crust.  Although the location of the basal reflector at PR 7 on 85-09 is not clear, possible positions place it at least 2 s TWT  sub-basement. Furthermore, in all PR locations dipping reflectors  beneath the normal base of the crust may delineate much greater thickening. At PRs  4 and 7  on 85-09, dipping reflectors continuous with the basal reflector coincide with the PR location, while the subcrustal thickening is offset from the PR trace; at PR 4, the crust is underplated for about 40 km  to the east of the PR trace, while at PR 7, thickening may occur for 17 km  to the west. Thickening at the base of the crust at the time of initiation of rifting seems plausible, since the upwelling magma is probablyfirstlimited to the base of the crust, before persistent heating has fractured the crust sufficiently to allow extrusion of material to the surface. The Galapagos PR shows substantial vertical motion ahead of the PR tip, with basalts erupted through reactivated cracks some 15 km ahead of the full spreading rift (Kleinrock and Hey, 1989). Garmany (1989) interpreted high amplitude phases on near-ridge refraction data as an indication of magmatic material intruded at the base of the crust as far as 20 km  from the  spreading axis. The possibility of magma underplating the crust for some distance ahead of  Chapter  4. DISCUSSION  AND  CONCLUSIONS  102  the point at which it first manages to reach surface seems plausible, if we imagine a plane of weakness between the crust and the upper mantle along which it might travel. The base of the crust is the most likely candidate, a significant compositional discontinuity which would deflect ascending melt. Malpas et al. (1989) examined the contact relations of the Troodos ophiolite, and observed evidence for substantial deformation at the base of the crust, which could provide a line of weakness for magma to follow. Further from the ridge, the base of the crust might be cold and brittle enough to allow fracturing of the crust and surface volcanism (e.g., seamount formation). It may be that the PR has a role in localization of subcrustal melts, since such reflectivity appears only in association with PR traces. If this interpretation of the data is correct, there is a wide range in the amount of magma underplated in association with rift propagation. Possible causes of such variability in magma production at the ridge will be examined in section 4.3.2. An interesting feature of the 85-09 crossing of PR 4 is the surface volcanism consisting of apparently massive lavaflowswhich coincides with the interpreted subcrustal underplating. A slight thinning of the crust which may indicate extension also coincides with the surface volcanism (Fig. 3.2). Kleinrock and Hey (1989) note the occurrence of a possibly analogous 'lava lake' associated with extensional faulting at the axial tip of the Galapagos PR.  The onset  of volcanism at the Galapagos PR is associated with the boundary of the magnetic source as determined by an inversion of deep-tow magnetic data (Miller and Hey, 1986). However, the magmatic source boundary as determined from surface measurements is located about 10 km behind this point, somewhat behind the beginning of the well-developed spreading axis. This discrepancy may be due to the influence on the magmatic signature of thick recent lava flows which overlie old crust formed at the failed spreading center (Kleinrock and Hey, 1989). The PR locations used in this study are the interpreted positions based on surface data (Wilson et al., 1984). This suggests that the hypothesis of significant magmatism for 40 km ahead of the PR's position as located by surface magnetic data might be verified by a  Chapter  4. DISCUSSION  AND  CONCLUSIONS  103  deep tow magnetic survey in this area. Even if the massive lavas are not sufficiently thick to modify the magnetic response of the crust, the seamount should provide a strong magnetic signature.  4.3.2  Magma sources  Possible sources of the increased magma supply to the ridge necessary to produce the apparent underplating or thickening of the crust include the influence of the Cobb Hotspot and convective partial melting associated with the initiation of rifting. Both the proximity of hotspots to PR in a number of locations (e.g., the Juan de Fuca, Galapagos and southern Atlantic regions) and the tendency of active PR to propagate away from hotspots have been cited in explanations of the dynamics of rift propagation: either propagating rifts are driven by sublithospheric flow of hotspot-derived material beneath the ridge axis (Hey and Vogt, 1977) or excess magma supply to the ridge as a result of hotspot activity produces elevated ridge axis topography, leading to horizontal gravitational stresses which induce fracturing and extension at the PR tip Q?hipps Morgan and Parmentier, 1985). On one hand the hotspot is a magmatic source, while on the other, only the effect of the nearby heat source is required to induce rift propagation. In order to evaluate these hypotheses, Karsten and Delaney (1989) have calculated the propositions of the Cobb Hotspot and associated PR in the NE Pacific over the last 7 My,  using the PR reconstruction of Wilson  et al. (1984). They provide estimates of the relative volumes of hotspot magma output over the sametimeperiod, based on the volume of the seamounts of the Cobb-Eickelberg chain, the surface expression of the Cobb Hotspot (Fig. .4.2; Fig. 2.1 shows the location of the Cobb-Eickelberg seamount chain). There is a general correlation between the different apparent volumes of crustal underplating imaged on the seismic sections at the PR zones and increased magma production  Chapter  4. DISCUSSION  AND  104  CONCLUSIONS  MEAN SECTOR CRUSTAL AGE (Ma) 0  «  D  C  4  0  )  I  O  I  O  O  <  O  O  O  T  O  i  n  o  O  O  <  D  MEAN AGE OF HOT SPOT ACTIVITY WITHIN SECTOR (Ma)  Figure 4.2: Plot of relative crustal volume for a corridor outlining the Cobb-Eickelberg seamount chain out to about 9 My (see Fig. 2.1 for location). Volumes are determined by integrating the bathymetry within each of 16 equal-area sectors within this corridor oriented parallel to the short axis of the box. The integration is taken relative to a baseline varied to account for thermal subsidence in older crust by using a square root of time conductive cooling model. Sedimentation effects are ignored. The age of hotspot activity within each of the sectors is based on the reconstruction model shown in Fig. 4.3, assuming a hotspot origin for each seamount. Mean crustal ages have been determined from magnetic anomalies and trie time scale in Fig. 1.9. (After Karsten and Delaney, 1989). at the hotspot.  Fig. 4.3 shows the Karsten and Delaney (1989) reconstruction from 4-2  My. The MCS line crossings of the pseudofaults occur at about 4, 3 and 2 My (85-09 and 85-07 crossing of PR 7, and 85-09 crossing of PR 4, respectively). At about 2 My, the high volume of hotspot production corresponds to the inferred time of significant excess magma production at the 85-09 crossing of PR 4. Conversely, at about 4 My, where there is a much smaller amount of underplating at the 85-09 crossing of PR 7, there is relatively small magma production at the hotspot. However, at 3 My, the time of the 85-07 PR 7 crossing,  Chapter  4. DISCUSSION  AND  CONCLUSIONS  i 136°  t  j  105  j  i  124°  Figure 4.3: Schematic diagrams reconstructing the propagation and ridge crest migration history for the Juan de Fuca Ridge relative to a fixed mantle reference frame. Melting anomaly positions are given by the solid squares, with HS identifying the Cobb Hotspot. Each panel depicts a time period and the orthogonal distance between the Cobb Hotspot and the nearest ridge segment (parentheses). Bold lines depict ridge axis locations. Thinner solid lines indicate pseudofault traces of propagating rifts identified in Fig. 1.9; stippled portions indicate lithosphere transferred from one plate to another by propagation. Circular features identify seamount locations, abbreviated: Eick., Eickelberg seamount; H, Hoh seamount; A, Anger seamount; G, Gluttony seamount; C, Corn seamount; BFZ, Blanco fracture zone; SFZ, Sovancofracturezone. 85-09 crossing of PR 7 coincides with PR tip at 4 My; 85-07 crossing of PR 7 coincides with PR tip at 3 My; 85-09 crossing of PR 4 coincides with PR tip at 2 My. (After Karsten and Delaney, 1989).  Chapter  4. DISCUSSION  AND  the production is closer to that at 4 My,  CONCLUSIONS and  106  is perhaps less than might be expected given the  relative amounts of underplating for the two crossings of PR 7. Desonie and Duncan (1990) explain the chemical trends of the Cobb-Eickelberg seamount chain by progressively greater mixing of plume and ridge mantle sources as the Juan de Fuca ridge has approached the hotspot. Although Karsten and Delaney (1989) had mixed success in correlation of parameters such as hotspot proximity to the ridge and productivity (from surface measurements) with propagation rates, excess magma due to hotspot activity may be more available to the ridge as the hotspot draws closer, producing the greater volume of thickening at the 85-07 crossing of PR 7. We also have no information regarding the out-of-plane extent of the thickening, which may vary between the different PR. Two  factors support the hypothesis that the hotspot role at the PR is simply that of a  driving force and does not supply magma to the PR zone. In general, the crustal thickness along the two lines does not appear to be related to the outputfromthe Cobb Hotspot (Figs. 3.1, 3.2 and 4.2). This may indicate that the PR plays a role in the localization of excess hotspot-derived magma at the ridge, but one might expect some variation in crustal thickness if there is substantial along-axis flow. Recent penological data show that the shallow (25-35 km)  magma sources which supply the tip of the Galapagos PR maintain their identities along  the length of the rift but are not found in the lavas at the failing rift, even in rocks erupted from the failing rift before it was overtaken by the PR.  This indicates that the sources must  somehow be related to rift propagation or to the Galapagos hotspot. However, the chemical characteristics of the erupted lavas suggest that sources separate from the hotspot supply the PR  (Christie and Sinton, 1986). Generation of greater volumes of magma at the ridge may also be explained by PR-  induced convective partial melting. Mutter et al. (1988) suggest that the thick sequences of basalt found at the point of initiation of rifting at passive margins may be due to generation of partial melts in which lateral temperature contrasts due to rifting and asthenospheric upwelling  Chapter  4. DISCUSSION  AND  CONCLUSIONS  107  drive convection in the upper mantle. This mechanism could explain the differences in relative amounts of crustal underplating as the result of the age-dependent cooling of the older lithosphere across the rift zone: older and cooler lithosphere would develop a larger lateral temperature gradient when juxtaposed with the thin, hot lithosphere of the PR, inducing greater convective partial melting than younger, hotter lithosphere.  4.3.3 Location of crustal underplating or thickening Another feature which differs between the PR crossings is the location of the subcrustal material with respect to the PR trace. Although we have only 3 examples, it appears that the position of the underplating and the age of the crust intersected by the PR trace may be related: the underplating occurs in progressively younger crust as the age difference across the PR trace diminishes. The  age difference across PR 4 is 1 My,  and the underplating occurs  exclusively in the older crust; at the 85-07 PR 7 crossing the age difference is 0.5 My and the thickening occurs symmetrically about the PR position inferred from surface magnetics; across the 85-09 crossing of PR 7 the ages differ by 0.25 My and the underplating appears on the young side of the PR trace. The location of the underplating may be controlled both by the rheological characteristics of the crust as it cools and  the development of a convective  partial melting system. Where the PR rifts older crust, the stronger temperature contrast across the rift would create a more robust convective system, generating more magma. The base of the crust would be colder and  better developed, tending to force upwelling magma to  flow horizontally into the older crust (Garmany, 1989). In younger crust, less of a convective system would be established. The base of the crust would be more ductile and less well developed, providing no path for localization of magmatic flow, and the smaller amount of magma generated might simply pond at the base of the crust. For very young crust, a convective partial melting system might not be established at all, because of the lack of  Chapter 4. DISCUSSION AND CONCLUSIONS  108  temperature contrast. This might explain the apparent thickening on the young side of the PR 7 trace on 85-09: a convective partial melting system may have begun to establish itself at about the time the crust here was rifted (hence the apparent continuity of the subcrustal dipping reflector and the basal reflector). As the PR ahead and to the west of this location rifted older preexisting crust, more melt would be generated, which might migrate along-axis behind the rift tip creating the observed subcrustal reflectivity.  4.3.4  Crustal structure of transform fault zones  Transform fault zones may be considered as the limiting case of PR zones in which no motion occurs across the transform boundary. Current evidence suggests that the crust in transform fault zones is either very thin (~3 km) and lacking Layer 3, or somewhat thicker than average oceanic crust, with a very thick Layer 3B. The former viewpoint is based on refraction surveys of fracture zones (e.g., White et al., 1984) which show material with p-wave velocities of 7.3-7.6 km/s from 3-6 km sub-basement, interpreted as serpentinized upper mantle material (White et al., 1990). The latter viewpoint is based on studies of the Bay of Islands ophiolite, which suggest that the depth of the crust-mantle boundary does not change substantially across fracture zones, but that the basal magmatic cumulate layer thickens markedly, and is responsible for the anomalous seismic response (Karson and Elthon, 1987). Thisfindinginvites comparison with the interpreted increased thickness of Layer 3B at PR 7 on 85-07 (and the refraction results from the PR crossing shown in Fig. 3.9 (White and Clowes, in preparation) which suggest thickening of Layer 3B). The Bay of Islands ophiolite chemistry also indicates that the amount of partial melting, rather than decreasing or remaining constant, might increase towards fracture zones. Our interpretation of the PR pseudofault traces is consistent with the model of increased partial melting at transform-type offsets, where hot and cold materials are juxtaposed. Convective partial melting might supply  Chapter  4. DISCUSSION  AND  CONCLUSIONS  109  the mechanism for constant or greater crustal thicknesses at fracture zones. 4.4  Conclusions  From the foregoing discussion it may appear that this study raises more questions than answers about the nature of the oceanic crust, particularly in PR zones. Many results and hypotheses arising from other surveys, particularly other MCS surveys, in other parts of the world and for ridges spreading at different rates, have been confirmed for crust generated at the Juan de Fuca intermediate rate spreading center (e.g., the variability of the oceanic crust, the persistence of the basal reflector). However, the lack of other geophysical information, particularly seismic refraction, along the lines makes interpretation of the PR zones difficult. However, a number of conclusions can be drawn from the dataset: 1. The base of the oceanic crust can be traced for more than 100 km across the Juan de Fuca plate on two separate seismic lines, indicating the continuity of this interface in oceanic crust generated at an intermediate rate spreading ridge. 2. Oceanic crustal thickness may vary by hundreds of meters awayfromthe ends of spreading center segments, and by one to two thousand meters at propagating rift traces. 3. The reflective zone at the base of the crust generally rangesfrom200-500 ms in time thickness, but may be as thick as 1 s in isolated instances. 4. Intracrustal reflectivity is highly variable: although generally there is a lack of coherent reflectors, strong intracrustal reflectors are imaged over 40 km on line 85-07. The most prominent and continuous reflector coincides with the interpreted oceanic Layer 3A/3B boundary from a nearby refraction survey. The Layer 2/3 boundary may be imaged  Chapter 4. DISCUSSION AND CONCLUSIONS  110  as a series of short reflectors; these may indicate a change in Layer 2 thickness of 0.5 km over several kilometers horizontal distance. 5. The outer pseudofault traces of propagating rifts have variable crustal structure, however in all cases they are associated with subcrustal dipping reflectors which appear to delineate zones of subcrustal reflectivity. These features are interpreted to indicate a substantial increase in the magma supply to the ridge at the time of rift propagation. 6. A small seamount was discovered at ~SP 700 on 85-09. It is interpreted as arising from excess magmatism associated with rift propagation.  4.5  Future work  The quality of information that can be derived from this survey is limited because of the random orientation (relative to isochrons or flowlines) of the seismic lines. Because of this, it is difficult to arrive at conclusions regarding, e.g., age dependence of crustal structure, since the complex spreading history of the N E Pacific precludes easy comparison of crust formed at the same point on the spreading axis. 3D effects can also be expected because of the line orientation. The dipping events are particularly problematic, because they may arise from out-of-plane scatterers. Crossing lines would be required to confirm these events. Furthermore, although the MCS lines are nearly orthogonal to the PR traces, we might still expect 3D effects in these locations because of the changing nature of the crustal structure at the PR tip which is rafted away as the outer pseudofault zone. However, other factors have led to a high quality dataset: the water bottom multiple is below the zone of interest for most of both lines, eliminating a major source of noise in marine data, and the sedimentary sequence provides a buffer so that all of the source energy is not diffracted toward the surface by the rugged basement, but sufficient energy travels to  Chapter  4. DISCUSSION  AND  CONCLUSIONS  the base of the crust. In fact, more MCS (Fig.  data acquired on aflowlineout from line 85-03  2.1) images the shallow reflector and base of the crust observed on 85-03 out to about  3.5 My Q3tohr et al., 1990). The NE and  111  Pacific also has a history of abundant rift propagation,  several outer pseudofault traces were imaged by the survey. Because some of the most interesting results of the research are associated with the PR  traces, future work might be concentrated on attempts to confirm some of the interpreted features of these zones, particularly the nature of the possible subcrustal reflectivity. The ideal survey might concentrate on the seismic character of the area around the Cobb offset of the Juan de Fuca ridge, with both MCS  and  refraction surveys. Expanding spread profiles  orthogonal to the PR traces at specific increments of age differences across the PR zone would be valuable in assessing the possible age-contrast dependence of the subcrustal structure. More MCS  lines in this area would be useful in addressing the same issue. The optimal  survey would include coincident reflection/refraction lines, combining the high resolution of reflection data with the velocity information of refraction data. The  seamount on 85-09 would also be an interesting survey target, particularly for age  dating, to evaluate the hypothesis that it is due to the rifting process. If this is not possible a deep tow magnetics experiment might accomplish the same result.  REFERENCES  Anstey, N., 1977. Seismic interpretation: the physical aspects, IHRDC, Boston, 637 pp. Becker, K., H. Sakai, A.C. Adamson, J. Alexandrovich, J.C.Alt, R.N. Anderson, D. Bideau, R. Gable, P.M. Herzig, S. Houghton, H. Ishizuka, H. Kawahata, H. Kinoshita, M.G. Langseth, M.A. Lovell, J. Malpas, H. Masuda, R.B. Merrill, R.H. Morin, M J . Mottl, J.E. Pariso, P. Pezard, J. Phillips, J. Sparks, and S. Uhlig, 1989. Drilling deep into young oceanic crust, hole 504B, Costa Rica Rift, Rev. Geophys., 27, pp. 79-102. Botros, M. and H.P. Johnson, 1988. Tectonic evolution of the Explorer-Juan de Fuca region from 8 Ma to the present, / . Geophys. Res., 93, pp. 10421-10437. Bratt, S.R., E.A. Bergman and S.C. Solomon, 1985. Thermoelastic stress: how important as a cause of earthquakes in young oceanic lithosphere?, / . Geophys. Res., 90, pp. 10249-10260. Calvert, A.J., E. Hasselgren and R.M. Clowes, 1990. Oceanic rift propagation - a cause of crustal underplating and seamount volcanism, Geol., 18, pp. 886-889. Christie, D.M. and J.M. Sinton, 1986. Major element constraints on melting, differentiation and mixing of magmas from the Galapagos 95.5° W propagating rift system, Contrib. Mineral. Petrol, 94, pp. 274-288.  Collins, J.A., T.M. Brocher and J.A. Karson, 1986. Two-dimensional seismic reflection modeling of the inferred fossil oceanic crust/mantle transition in the Bay of Islands ophiolite, /. Geophys. Res., 91, pp. 12520-12538. Davis, E.E., D.S. Chapman, C.B. Forster and H. Villinger, 1989. Heat-flow variations correlated with buried basement topography on the Juan de Fuca Ridgeflank,Nature, 342, pp. 533-537. Desonie, D.L. and R.A. Duncan, 1990. The Cobb-Eickelberg seamount chain: hotspot volcanism with MORB affinity, / . Geophys. Res., 95, pp. 12697-12711. Derrick, R.S., P. Buhl, E. Vera, J. Mutter, J. Orcutt, J. Madsden and T. Brocher, 1987. Multichannel seismic imaging of a crustal magma chamber along the East Pacific Rise, Nature, 326, pp. 35-41.  112  REFERENCES  113  Ewing, J. and R. Houtz, 1979. Acoustic stratigraphy and structure of the oceanic crust, in Talwani, M., C.G.A. Harrison and D.E. Hayes, eds., Deep drilling results in the Atlantic Ocean: ocean crust, Washington, D . C , A.G.U., Maurice Ewing Series, p. 1-14. Fornari, D.J., W.B.F. Ryan and PJ. Fox, 1984. The evolution of craters and calderas on young seamounts: insights from Sea MARC I and Sea Beam sonar surveys .of a small seamount group near the axis of the East Pacific Rise at ~10°N, / . Geophys. Res., 89, pp. 11069-11083. Garmany, J., 1989. Accumulations of melt at the base of young oceanic crust, Nature, 340, pp. 628-632. Groschel-Becker, H.M. and B.R. Rosendahl, 1990. Structure of oceanic crust and upper mantle in eastern Gulf of Guinea from deep-imaging seismic reflection data, EOS, 71, p. 1615. Hall, J.M., C C . Walls and J-S. Yang, 1989. Constructional features of the Troodos ophiolite and implications for the distribution of orebodies and the generation of oceanic crust, Can. J. Earth Sci., 26, pp. 1172-1184. Hey, R. and P. Vogt, 1977. Spreading center jumps and sub-axial asthenosphere flow near the Galapagos hotspot. Tectonophysics, 37, pp. 41-52. Hildebrand, J.A., L.M. Dorman, P.T.C. Hammer, A.E. Schreiner and B.D. Cornuelle, 1989. Geophys. Res. Lett., 16, pp. 1355-1358. Hyndman, R.D., C J . Yorath, R.M. Clowes and E.E. Davis, 1990. The northern Cascadia subduction zone at Vancouver Island: seismic structure and tectonic history, Can. J. Earth Sci., 27, pp. 313-329. Jones, T.D. and A. Nur, 1984. The nature of seismic reflections from deep crustal fault zones, /. Geophys. Res., 89, pp. 3153-3171. Karson, J.A. and D. Elthon, 1987. Evidence for variations in magma production along oceanic spreading centers, Geol., 15, pp. 127-131. Karsten, J.L., and J.R. Delaney, 1989. Hot spot-ridge crest convergence in the northeast Pacific, / . Geophys. Res., 94, pp. 700-712. Kent, G.M., A.J. Harding and J.A. Orcutt, 1990. Evidence for a smaller magma chamber beneath the East Pacific Rise at 9°30'N, Nature, 344, pp. 650-653.  REFERENCES  114  Kirby, S., C. Mawer, G. Iturrino and N. Christensen, 1988. Ductile shear zones in layer 3 gabbroic rock: their deformation structures, Vp anisotropy and candidacy for the curvilinear reflection structures of the lower crust, EOS, 69, pp. 1402. Kleinrock, M.C. and R.N. Hey, 1989. Detailed tectonics near the tip of the Galapagos 95.5°W propagator: how the lithosphere tears and a spreading axis develops, / . Geophys. Res., 94, pp. 13801-13838. Langmuir, C.H., J.F. Bender and R. Batiza, 1986. Petrological and tectonic segmentation of the East Pacific Rise, 5°30'-14°30' N, Nature, 322, pp. 422-429. Lamer, K., R. Chambers, M. Yang, W. Lynn and W. Wai, 1983. Coherent noise in marine seismic data, Geophys., 48, pp. 854-886. Lumley, D.E., 1989. A generalized Kirchoff-WKBJ depth migration theory for multi-offset seismic reflection data: reflectivity model construction by wavefield imaging and amplitude estimation, M.Sc. Thesis, Univ. of British Columbia. MacDonald, K.C., P.J. Fox, L.J. Perram, M.F. Eisen, R.M. Haymon, S.P. Miller, S.M. Carbotte, M.-H. Cormier and A.N. Shor, 1988. A new view of the mid-ocean ridge from the behaviour of ridge-axis discontinuities, Nature, 335, pp. 217-225. MacDonald, K.C., 1982. Mid-ocean ridges: fine scale tectonic, volcanic and hydrothermal processes within the plate boundary zone, Ann. Rev. Earth Planet. Sci., 10, pp.155-190. McCarthy, J., J.C. Mutter, J.L. Morton, N.H. Sleep and G.A. Thompson, 1988. Relic magma chamber structures preserved within the Mesozoic North Atlantic crust? G.SA. Bull., 100, pp. 1423-1436. Malpas, J., T. Brace and S.M. Dunsworth, 1990. Strustural and petrological relationships of the CY-4 drill hole of the Cyprus Crustal Study Project, in Cyprus Crustal Study Project: Initial Report, Hole CY-4, LL. Gibson, J. Malpas, P.T. Robinson and C. Xenophontos, eds., Geol. Survey of Canada Paper 88-9. Mankinen, E.A. and G.B. Dalrymple, 1979. Revised geomagnetic polarity time scale for the interval 0-5 My B.P., / . Geophys. Res., 84, pp. 615-626. Miller, S.P. and R.N. Hey, 1986. Three-dimensional magnetic modeling of a propagating rift, Galapagos 95°30'W, J. Geophys. Res., 91, pp. 3395-3406. Morton, J.L., N.H. Sleep, W.R. Normark and D.H. Tompkins, 1987. Structure of the southern Juan de Fuca ridge from seismic reflection records, J. Geophys. Res., 92, pp. 1131511326.  REFERENCES  115  Musgrove, L.A. and J.A. Austin, 1983. Intrabasement structure in the southern Angola Basin, Geol, 11, pp. 169-173. Mutter, J.C, W.R. Buck and C M . Zehnder, 1988. Convective partial melting 1. A model for the formation of thick basaltic sequences during the initiation of spreading, / . Geophys. Res., 93, pp. 1031-1048. NAT Study Group, 1985. North Atlantic transect: a wide-aperture, two-ship multichannel seismic investigation of the oceanic crust, / . Geophys. Res., 90, pp. 10321-10341. Orcutt, J.A., B. Kennett, L. Dorman and W. Prothero, 1975. A low velocity zone underlying a fast spreading rise crest, Nature, 256, pp. 475-476. Phipps Morgan, J. and E.M. Parmentier, 1987. A three-dimensional gravity study of the 95.5°W propagating rift in the Galapagos spreading center, Earth Planet. Sci. Lett., 81, pp. 289-298. Phipps Morgan, J. and E.M. Parmentier, 1985. Causes and rate-limiting mechanisms of ridge propagation: a fracture mechanics model, J. Geophys. Res., 90, pp. 8603-8612. Purdy, G.M., 1987. New observations of the shallow seismic structure of young oceanic crust, / . Geophys. Res., 92, pp. 9351-9362. Purdy, G.M., 1983. The seismic structure of 140 Myr old crust in the western central Atlantic ocean, Geophy. J. R. Astr.Soc, 72, pp. 115-137. Purdy, G.M. and J. Ewing, 1986. Seismic structure of the ocean crust, in Vogt, P.R. and E. Tulcholke, eds., The Geology of North America, Vol. M„ The Western North Atlantic Region, Geol. Soc. Am., pp. 313-330. Raitt, R.W., 1963. The crustal rocks, in Hill, M.N., ed., The Sea, Vol. 3, pp. 85-102, Wiley, New York. Riddihough, R., 1984. Recent movements of the Juan de Fuca plate system, / . Geophys. Res., 89, pp. 6980-6994. Rohr, K.M.M., E.E. Davis and R.D. Hyndman, 1990. Structure of the upper oceanic crust on the eastern flank of the Juan de Fuca ridge: constraints from new multichannel seismic data, EOS, 71, p. 1566. Rohr, K.M.M., B. Milkereit and C J . Yorath, 1988. Asymmetric deep crustal structure across the Juan de Fuca ridge, Geol., 16, pp. 533-537.  REFERENCES  116  Salisbury, M.H. and .1. Christensen, 1978. The seismic velocity structure of a traverse through the Bay of Islands ophiolite complex, Newfoundland, an exposure of oceanic crust and upper mande, / . Geophys. Res., 83, pp. 805-817. Sellevol M.A. and M. Mokhtari, 1988. An intra-oceanic crustal seismic reflecting zone below the dipping reflectors on Lofoten margin, Geol., 16, pp. 666-668. Sempere, J.-C, and K.C. MacDonald, 1987. Marine tectonics: processes at mid-ocean ridges, Rev. Geophys., 25, pp. 1313-1347. ten Brink, U.S. and T.M. Brocher, 1987. Multichannel seismic evidence for a subcrustal intrusive complex under Oahu and a model for Hawaiian volcanism, / . Geophys. Res., 92, pp. 13687-13707. Toomey, D.R., G.M. Purdy and S.G Solomon, 1989. Three-dimensional seismic structure of the East Pacific Rise at 9°30'N, EOS, 70, p. 1317. White, D.J. and R.M. Clowes, 1990. Shallow crustal structure beneath the Juan de Fuca ridge from 2-D seismic refraction tomography, Geophys. Jour. Int., 100, pp. 349-368. White, R.S., 1984. Atlantic oceanic crust: seismic structure of a slow-spreading ridge, in L.G. Gass, S.J. Lippard and A.W. Shelton, eds., Ophiolites and Oceanic Lithosphere, Geol. Soc. Spec. Publ. 13, pp. 101-111. White, R.S., R.S. Derrick, J.C. Mutter, P. Buhl, T. Minshull and E. Morris, 1990. New seismic images of oceanic crustal structure, Geol., 18, pp. 462-465. White, R.S., R.S. Derrick, M.C. Sinha and M.H. Cormier, 1984. Anomalous seismic crustal structure of oceanic fracture zones, Geophy. J. R. Astr. Soc., 79, pp. 779-798. White, W.R.H. and R.M. Clowes, 1987. Juan de Fuca plate crustal structure: results from ocean bottom seismograph studies, EOS, 68, p. 1371. Wilson, D.S., R.N. Hey and C. Nishimura, 1984. Propagation as a means of reorientation of the Juan de Fuca ridge, / . Geophys. Res., 89, pp. 9215-9225. Yilmaz, O., 1987. Seismic Data Processing, S.E.G., Tulsa, 526 pp. Yorath, C.J., R.M. Clowes, R.D. MacDonald, C. Spencer, K. Rohr, J. Sweeney, R. Currie, E.E. Davis, J.F. Halpenny and D.A. Seeman, 1988. Marine multichannel seismic reflection, gravity and magnetic profiles - Vancouver Island continental margin and Juan de Fuca ridge, Sidney, British Columbia, Geol. Survey of Canada Open File Report 1661.  


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



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"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items