Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Shallow crustal structure of the Endeavour Ridge segment, Juan de Fuca Ridge, from a detailed seismic… Cudrak, Constance Frances 1988

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

Item Metadata

Download

Media
831-UBC_1988_A6_7 C82.pdf [ 9.11MB ]
Metadata
JSON: 831-1.0052628.json
JSON-LD: 831-1.0052628-ld.json
RDF/XML (Pretty): 831-1.0052628-rdf.xml
RDF/JSON: 831-1.0052628-rdf.json
Turtle: 831-1.0052628-turtle.txt
N-Triples: 831-1.0052628-rdf-ntriples.txt
Original Record: 831-1.0052628-source.json
Full Text
831-1.0052628-fulltext.txt
Citation
831-1.0052628.ris

Full Text

S H A L L O W C R U S T A L S T R U C T U R E O F T H E E N D E A V O U R R I D G E S E G M E N T , J U A N D E F U C A R I D G E , F R O M A D E T A I L E D S E I S M I C R E F R A C T I O N S U R V E Y B y Constance Frances C u d r a k B. Sc. (Geophysics), Queen 's Univers ity, 1985 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE i n THE FACULTY OF GRADUATE STUDIES DEPARTMENT OF GEOPHYSICS AND ASTRONOMY We accept this thesis as conforming to the required s tandard THE UNIVERSITY OF BRITISH COLUMBIA October 1988 © Constance Frances Cud rak , 1988 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department The University of British Columbia Vancouver, Canada •ate f V _ y ^ W r \H?f ( q g g -DE-6 (2/88) Abstract T h e Endeavour R idge is a segment of the J u a n de Fuca R idge, an active spreading centre wh i ch lies off western No r th A m e r i c a between the Pac i f i c and J u a n de Fuca plates. Th i s segment is a bathymetr ic h igh and a site of hyd ro the rma l a c t i v i t y — b o t h characterist ics suggest an under ly ing heat source such as an ax i a l m a g m a chamber wh ich is associated w i t h crusta l generation. To investigate the creat ion and evo lut ion of oceanic crust, a deta i led refract ion survey was carr ied out over the Endeavour R idge in the fa l l of 1985. A s a component of this survey, a d iamond-shaped array consist ing of eight O B S along a 2 0 - k m l ine across the ridge and two O B S p laced along i t at distances of 10 k m on either side of the cross-ridge l ine was deployed to define the shal low crusta l structure near and beneath the ridge, especially the possible existence of an ax ia l magma chamber. A i r g u n shots at 0.2 k m intervals along ~ 3 0 0 k m of profiles prov ide convent ional reversed and unreversed refraction lines as wel l as mu l t ip le f u l l a z imutha l coverage of the region. Trave l - t ime and ampl i tude da ta f rom fifteen in- l ine a i rgun profiles recorded on the inner array were forward model led using an a l go r i thm based on a symptot i c ray theory w i t h a s tart ing mode l obta ined f rom a concurrent study. Two-d imens i ona l models were constructed and then combined to obta in the three-d imens ional s t ructure of the region. These models consist of four layers, w i th the average mode l correlat ing wel l to the classic mode l of oceanic crust. Layer 2 A averages 0.40 k m i n thickness and has velocities of 2.6 km/s and 2.8 km/s at the top and b o t t o m of the layer, respectively. To achieve such a low velocity, Layer 2 A must consist of h igh ly f ractured vesicular basalts. A sharp veloc ity increase to 4.8 km/s marks the t rans i t ion to Layer 2B. Th i s velocity d i scont inu i ty is also v is ib le as a reflector on a. mu l t i channe l reflection l ine obta ined through the centre i i of the s tudy region and is caused by an abrupt decrease i n porosity. Layer 2 B averages 0.67 k m i n thickness, has a veloc ity of 5.4 km/s at its base and consists of less f ractured p i l low basalts and sheet flows. T h e Layer 2 B - L a y e r 2C interface is a velocity increase to 5.8 km/s and is the p i l low basalt-sheeted dike contact. A sma l l velocity increase f rom 6.3 to 6.5 km/s delineates the base of the 0.95 km-th ick Layer 2 C wh ich is the bounda ry between the sheeted dikes and cumulate gabbros i n Layer 3. Layer 3 has the lowest veloc ity gradient (0.30 s _ 1 ) and a veloc ity of 7.3 km/s at 4.65 k m below the seafloor, the m a x i m u m depth constrained by the mode l l ing . L a t e r a l heterogeneities on the scale of 2-3 k m are super imposed on this bas ic velocity structure. These heterogeneities are effects of poros i ty changes, differential pressure changes, and a l terat ion caused by hyd ro the rma l c i rcu lat ion. Layer 2 A thins and increases i n velocity away f rom the ridge; r idge-paral le l cracks create a velocity anisotropy of ~ 1 0 -2 5 % , the faster d i rect ion para l le l to the ridge. Velocit ies w i t h i n Layers 2B and 2 C also increase by 0.1 km/s away f rom the axis of the ridge. Layer 3 velocities decrease by 0.1 km/s for arrivals t ravel l ing under the ridge. Increased Layer 2 velocities at the ridge crest reveal h igh latera l velocity constrasts i n very young crust, but w i t h i n 0.03 M a the oceanic crust at the ridge has matu red to the off-ridge structure. N o f i rm evidence exists for a large m a g m a chamber under Endeavour R idge. A l t h o u g h the ba thymet r i c h igh and h igh-temperature hydro therma l discharges are evidence for a m a g m a chamber, the lack of recent sheet flows at the ridge crest and the presence of a rift a long the crest ind icate the m a g m a chamber is wan ing and must be of a size (<1 k m in w id th ) not resolvable by seismic refract ion data. in Table of Contents Abstract ii List of Tables vii List of Figures viii Acknowledgements xi 1 INTRODUCTION 1 1.1 T h e Scientif ic E xpe r iment 1 1.2 T h e Natu re of Oceanic C ru s t 2 1.2.1 T h e Seismic S t ruc ture of Oceanic C ru s t . 2 1.2.2 Character i st ics of Ocean Spreading Centres 7 1.2.3 M a g m a Chamber s 10 1.2.4 Three-d imens iona l i ty of Spreading Centres 16 1.3 T h e Endeavour R idge Segment, J u a n de Fuca R idge 19 1.3.1 Reg iona l S t ructure 19 1.3.2 Geomorpho logy of Endeavour R idge 22 1-3.3 Prev ious Geophys i ca l Studies 24 2 DATA ACQUISITION AND PROCESSING 27 2.1 T h e Refract ion Expe r iment , S E I S R I D G 85 27 2.2 Instruments 30 2.3 D a t a Process ing 31 iv 2.3.1 In i t ia l D a t a Reduc t i on 31 2.3.2 Shot and Receiver Po s i t i on ing 32 2.3.3 B a t h y m e t r y and Inst rument Depths 33 2.3.4 T i m i n g Correct ions 34 3 DATA ANALYSIS 36 3.1 In i t ia l Observat ions and Process ing 36 3.1.1 D a t a Character i s t ics and Qua l i t y 36 3.1.2 T rave l T i m e P i c k i n g 42 3.1.3 O B S Pos i t i on ing F rom Wate r -Wave Ar r i va l s 49 3.2 Interpretat ion Methods 52 3.2.1 In i t i a l Interpretat ion 52 3.2.2 Mode l l i n g A l g o r i t h m 53 3.2.3 S ta r t ing Ve loc i ty M o d e l 54 3.2.4 Mode l l i n g P rocedure 58 4 INTERPRETATION 62 4.1 Off-ridge L ines 64 4.1.1 L i ne 8 ( O B S 8 and O B S 5) 64 4.1.2 L i ne 5 (OBS 2 and O B S 8) 70 4.1.3 L i ne 6 ( O B S 2 and O B S 9) 78 4.1.4 L i ne 7 ( O B S 9) 84 4.1.5 L i ne 9 ( O B S 2) 87 4.1.6 L i n e 10 ( O B S 5) 90 4.2 Cross-r idge Lines 93 4.2.1 L i ne 2 ( O B S 12 and O B S 14) 93 4.2.2 L i ne 3 (OBS 8) 101 v 4.2.3 L i ne 4 ( O B S 9) 104 4.3 A long-r idge L i n e — L i n e D5 ( O B S 9 and O B S 8) 107 4.4 F i n a l M o d e l of Region 114 5 DISCUSSION AND CONCLUSIONS 125 5.1 Compar i s on of One-d imens iona l M o d e l w i t h Other Refract ion Surveys . . 125 5.2 Layer 2 127 5.2.1 Layer 2 A 127 5.2.2 Layer 2 B 139 5.2.3 Layer 2C 140 5.2.4 Genera l Character i s t ics of Layer 2 141 5.3 Layer 3 143 5.4 A x i a l Ref lector 145 5.5 Conclus ions 147 References 149 vi List of Tables 1.1 Hou t z and E w i n g (1976) mode l of oceanic crust 3 2.1 Tape-head-skew values for the O B S 34 3.1 T i m i n g and distance corrections made to the O B S pos it ions 51 5.1 Compar i s on of results w i t h other refract ion studies 126 vii List of Figures 1.1 Genera l i zed velocity structure and geology of oceanic crust 4 1.2 Topography and tectonic structure of ocean spreading ridges 8 1.3 A x i a l m a g m a chamber models 14 1.4 T h e J u a n de Fuca R idge system 20 1.5 Nor thern J u a n de Fuca R idge 21 2.1 Loca t i on and conf igurat ion of explosives array 28 2.2 Loca t i on and conf igurat ion of a irgun array 29 3.1 Frequency content of noise and signal 38 3.2 Trave l - t ime branches for L ine 9 - O B S 6 39 3.3 Low-ve loc i t j T B r anch 1 arr iva l 41 3.4 Effect of f i l ter ing poor qual i ty da ta 43 3.5 Effect of f i l ter ing a noise-contaminated trace 44 3.6 W h a l e sound 45 3.7 Pos i t i on of first break and representative wavelet 46 3.8 P i c k i n g arrivals f rom good qual ity da ta 47 3.9 P i c k i n g arrivals f rom poor qualit)- da ta 48 3.10 Repos i t i on ing an O B S 50 3.11 S ta r t ing mode l and refracted arrivals 55 3.12 S ta r t i ng model and reflected arrivals 56 3.13 Ve loc i ty -depth funct ion of start ing mode l 57 vin 4.1 Loca t i on of a i rgun lines model led 63 4.2 L i ne 8 - O B S 8 data and mode l 65 4.3 L i ne 8 - O B S 5 data and mode l 68 4.4 L i ne 5 - O B S 2 data and mode l 72 4.5 L i ne 5 - O B S 8 data and mode l 74 4.6 R a y paths for secondary arrivals „ 77 4.7 L i ne 6 - O B S 2 data and mode l 79 4.8 L i ne 6 - O B S 9 data and mode l 82 4.9 L i ne 7 - O B S 9 data and mode l 85 4.10 L i ne 9 - O B S 2 data and mode l 88 4.11 L i ne 1 0 - O B S 5 da ta and mode l 91 4.12 L i ne 2 - O B S 12 data and model l ing. . 95 4.13 Line 2 - O B S 14 data and mode l l ing 97 4.14 Ve loc i ty mode l for L i ne 2 100 4.15 L i ne 3 - O B S 8 data and mode l 102 4.16 L i ne 4 - O B S 9 data and mode l 105 4.17 L i ne D 5 - O B S 9 da ta and mode l 109 4.18 L i ne D 5 - O B S 8 da ta and mode l I l l 4.19 Seafioor bathymetry 116 4.20 Th ickness of Layer A 117 4.21 Dep th to Layer A - L a y e r B interface 118 4.22 Th ickness of Layer B 120 4.23 Dep th to Layer B - L a y e r C interface 121 4.24 Th ickness of Layer C 122 4.25 Dep th to Layer C -Layer D interface 123 4.26 F i n a l average one-dimensional mode l 124 i x 5.1 P-wave ve loc i t j f as a funct ion of poros i ty 130 5.2 Compar i son of reflectors and refract ion interfaces for cross-ridge l ine. . . 134 5.3 B u l k poros i ty and permeabi l i ty i n D S D P Hole 504B 136 x Acknowledgements I wi sh to thank my thesis advisor, Dr. R .M. Clowes, for his gu idance and encouragement dur ing m y three years as his student. M y most heart fu l thanks goes to D o n W h i t e (BOSS! ) w i thout w h o m I wou ld never have had any data to interpret. His organ iz ing of the f ield program, d ig i t i z ing of the data, and answering of my my r i ad of questions are greatly appreciated. Numerous people assisted i n obta in ing the data. I acknowledge C a p t a i n A l a n C h a m -ber la in and the officers and crew of the CSS Parizeau for their cooperat ion and assistance. A l t h o u g h the explosives da ta are not analyzed i n this thesis, I express my thanks to Ch ie f A l Woods . Les Rourke, and D a n Desjardins of the Fleet D i v i n g Un i t , C F B E squ ima l t , for the i r safe and careful hand l i ng of the explosives under d i f f icult sea condit ions. Ben C i amma i che l l a , J o h n Bennest, and B o b M e l d r u m must be commended for their s ta lwart efforts on the technica l side lead ing to fourteen successful ocean bo t t om seismograph deployments and recoveries and an overal l successful program. Spec ia l thanks to the geo-physics graduate students, Sonya Dehler, Chr i s P i ke , and A n d y Bo l and , who vo lunta r i l y commi t ted themselves to three weeks on the Pac i f i c Ocean i n late Oc tober -November . T h a n k you again to Don W h i t e and Dr. R o n Clowes who p l anned and organized the fu l l f ield p rog ram and who were also at sea. Th i s da ta wou ld not have been dig i t ized w i thout the electronic w i za rd ry of Bob M e l d r u m and Ben C i amma i che l l a and the p rog ramming skil ls of D o n Wh i t e . J o h n Hole and B o b M e l d r u m spent many months obta in ing a i rgun clock dr i ft corrections and I thank them for their efforts. xi Special thanks to Colin Zelt for his expertise in refraction modelling, computer pro-gramming, and cutting and pasting and his constant support and understanding. Thank you to my office mate, Stan Dosso (STAN!), who knows more about seismic refraction modelling than he wishes he did and to Andrew Calvert for information on the seismic reflection data. A "global" thank you to Dr. D.W. Oldenburg who tolerated my teasing and thankfully did not mark my thesis. The careful reading and thoughtful comments of my thesis readers, Drs. M.J. Yedlin and R.L. Chase are appreciated. Thank you to all the members of the Geophysics and Astronomy Department not specifically mentioned; you made my stay in Vancouver an enjoyable experience. I wish to express my thanks to Ron and Lorraine MacLeod and the Zelts (Harold, Joan, Mike and Barry) who pulled me away from the computer terminal for food, cards, .and Pictionary. A final thank-you to nry parents, brother, and sisters for their love and understanding over the years. Financial assistance for the data acquisition was provided by DSS Contract No. 11SB-23227-5-0209 (1985) through the Pacific Geoscience Centre (GSC). Research support was provided by EMR Research Agreement No. 206 (1987) through the Geological Survey of Canada and Infrastructure Grant No. 0855 and Operating Grant A7707, both from the Natural Sciences and Engineering Research Council, Canada. xii Chapter 1 I N T R O D U C T I O N 1.1 The Scientific Experiment Since the establ ishment of the seafioor spreading hypothesis i n the 1960's, the concept that oceanic crust is formed at spreading ridges by magmat i c in t rus ion has been widely accepted. Fundamenta l questions s t i l l exist, however, about the processes by wh ich oceanic crust is formed and how i t changes as i t moves away f rom the ridge. To answer some of the questions concerning crusta l creation and evo lut ion, a detai led seismic refrac-t ion survey ( S E I S R I D G 85) was carr ied out over an act ive spreading ridge, the J u a n de Fuca R idge ( J D F R ) wh ich is located i n the northeast Pac i f i c Ocean west of the Canad i an landmass. T h e specific site chosen for the s tudy was centred on the m idd le of Endeavour Ridge, a segment of the J D F R . Three ma i n exper imenta l objectives were used to p lan the present stud}': 1. Dete rminat i on of the regional crusta l s t ructure across and along an act ive spreading ridge w i t h part icu lar emphasis on the role of the crust -mant le boundary, the Moho. 2. Invest igat ion of the detai led structure of the shal low oceanic crust of the ridge and sur round ing areas, especial ly the evo lut ion of crusta l structure f rom the ridge to older off-ridge parts of the newly formed plate. 3. Invest igat ion of the existence of a shal low m a g m a chamber beneath the ridge crest. 1 Chapter 1. INTRODUCTION 2 Th i s thesis presents results that fu l f i l l the second exper imenta l objective. Determi -na t i on of the regional crusta l s t ructure w i l l be the basis of a separate stud}'. A computer a l go r i thm for tomograph ic invers ion of the seismic refract ion data has been developed by W h i t e (1988) to s tudy the th i rd object ive, the del ineat ion of a possible m a g m a chamber. Results f rom this work (Wh i t e and Clowes, 1988) w i l l be summar ized for completeness. 1.2 The Nature of Oceanic Crust T h e fo l lowing sections describe the basic s t ructure of the oceanic crust, the geomorphol -ogy of oceanic spreading ridges, and how oceanic crust is bel ieved to be formed. Th i s i n fo rmat ion provides a rat ionale and background for the exper imenta l objectives and the exper iment locat ion and w i l l be referred to in the discuss ion of the data. 1.2.1 The Seismic Structure of Oceanic Crust T h e basic structure of oceanic crust is known f rom seismic refract ion and reflection s tud-ies, Deep Sea D r i l l i n g Project ( D S D P ) and Ocean D r i l l i n g P r o g r am ( O D P ) sampl ing and well- logging, dredging of samples f rom the ocean b o t t o m , submers ible observations, and the geological stud}' - of ophio l i te sequences on land. O f these different survey techniques, seismic refraction has been the most impo r tan t i n obta in ing the structure of the oceanic crust because of the relative ease i n obta in ing refract ion data. T h e interpretat ion procedures used on seismic refraction data obta ined in surveys i n the 1970's usua l ly assumed constant-veloc i ty layers because the density of the da ta obta ined was inadequate to resolve details such as velocity gradients; however, the ter-mino logy used to labe l these layers is st i l l in use, even though gradient models have su-perseded the constant-veloc ity layered models ( Spud ich and Orcut t , 1980a). T h e Hou t z and E w i n g (1976) model for oceanic crust i n the Pac i f i c Ocean is shown in Tab le 1.1. Chapter 1. INTRODUCTION 3 P-wave Velocity Thickness Layer (km/s) (km) 2 A 3.33 ± 0 . 1 0 0.74 ± 0.23 2B 5.23 ± 0.44 0.72 ± 0.26 2C 6.19 ± 0 . 1 6 1.83 ± 0 . 7 5 ° 3 6.92 ± 0.17 b a: C u m u l a t i v e thickness of Layers 2B and 2C. b: M a n t l e and Layer 3 B general ly not detected. Tab le 1.1: Hou t z and E w i n g (1976) layered ve loc i ty mode l for 5 Ma -o l d Pac i f i c oceanic crust der ived f rom sonobuoy refract ion data. T h e mant le and Layer 3B were generally not detected because of the short shot-receiver offsets for these data. T h e determinat ion of rock types for the different seismic velocity layers is established by compar i son w i t h oph io l i te sequences (segments of oceanic crust emplaced on land) and petrologic studies of dredged samples, since i t is not possible to define unambiguous ly geologic s t ructure f rom compress ional velocity alone. A geologic mode l for the oceanic crust is presented in F igure 1.1. E a c h of the oceanic crustal layers w i l l be discussed, as wel l as some of the problems associated w i t h their interpretat ion. T h e velocity and thickness of Layer 2 A is poorhy constrained because the configura-t ion of surface shots and receivers used i n a t yp i ca l refraction exper iment results in the direct water-wave arr iva l mask ing arrivals f r om the uppermost crust. B o t t o m receivers (ocean b o t t o m seismographs or O B S ) have helped somewhat, but even so a velocity is u sua l l y assumed or ext rapo lated for this layer (e.g., E w i n g and Purdy , 1982). The average velocity for this layer determined by H o u t z and E w i n g (1976) is 3.33 km/s (Ta-ble 1.1) w i t h the lowest velocities recorded at the axis of ridges (2.8 to 3.7 km/s) and the veloc ity increas ing to 4.5 km/s w i t h distance away f rom the ridge (Houtz , 1976). Mo re recent studies have defined a steep velocity gradient i n this layer (e.g., Spud ich and Or-cutt , 1980b; Pu rdy , 1987). L abo r a t o r j 7 measurements of basalt samples, however, obta in Chapter 1. INTRODUCTION VELOCITY (KM/S! 0 2 4 6 (xxxxxxxxxxx xxxxxxxxxxx- x ( X X X X X X X X X X X xxxxxxxxxxxx X X X X X X X X X X X X X X X X X X X X X X X ( X X X X X X X X X X X X X X X X X X X X X X X ( X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X ( X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X ( X X X X X X X X X X X X X X X X X X X X X X X ( X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X V X X X X X X / / / / s / / / / / J / / / ' / / /• / / / / / / / / / / / / / / Mils % •"v. basaltic pillows and flows pillows intruded by dikes sheeted dikes isotropic gabbro cumulate gabbro ultramafic harzburgite Figure 1.1: Generalized velocity structure of oceanic crust and inferred geologic interpre-tation. Figure based on results from D S D P Hole 504B (Newmark et al, 1985a; 1985b), refraction studies (Bratt and Purdy, 1984; Purdy and Detrick, 1986), and studies of ophiolite complexes (e.g., Salisbury and Christensen, 1978; Casey et ai, 1981). C h a p t e r 1. INTRODUCTION 5 velocit ies of about 6 km/s (e.g., Chr i s tensen and Smewing, 1981; Chr i s tensen and Salis-bury, 1985), much higher than those obta ined in situ. T h e most l i ke ly cause of this low veloc i ty is high poros i ty caused by fluid-filled vesicles and fractures since geologically, Layer 2 A has been found to consist of a chaotic extrus ive assemblage of p i l l ow lavas, mass ive basalt flows, rubble, and intercalated sediments w i th cracks and fissures that f i l l w i t h low-temperature a l terat ion minerals over t ime (Fox and Stroup, 1981). Hou t z and E w i n g (1976) suggested a systemat ic increase i n velocity of this layer w i t h age, but O rcu t t et al. (1984) disagreed w i t h this conclus ion. Hou t z and E w i n g (1976) a t t r i bu ted the ve loc i ty increase to a decrease i n porosity by the in f i l l i ng of cracks and pores by sedi-ments and a l terat ion products p roduced by hyd ro the rma l c i rcu lat ion, wh i le O rcu t t et al. (1984) stated that the perceived velocity increase may s imply be ascr ibed to regional or even loca l var iat ions in crusta l structure. L i t t l e is known about this layer i n very young oceanic crust, especial ly on s low-spreading ridges l ike the M i d - A t l a n t i c R idge, because of the d i f f icu l ty i n do ing exper iments and interpret ing da ta (i.e., topograph ic corrections; Pu rdy , 1982a) i n the region of r ap id l y changing ba thymet r y found at spreading ridges. Layer 2B velocities are aga in lower than those obta ined in laboratory studies of dredged samples, w i t h the increase i n velocity f rom Layer 2A at t r i bu ted to the progres-sive closing of cracks w i th a l terat ion minerals and l i thostat ic pressure. Th i s layer consists of massive basalts and cross-cutt ing dikes and has its lower boundary at ~1.3 k m below the seafloor. Layer 2C velocities f ina l ly reach those obta ined f rom laboratory studies of uncracked basalt at the appropr iate pressures and temperatures. T h e velocity increase is caused by lower porosit} ' and an increas ing amount of gabbro. Th i s layer is formed by sheeted dikes, ~1 m wide and bound by chi l led margins, w i t h gabbroic rocks at the base of the section at a depth of ~ 2 k m (Casey et al., 1981). Layer 3, the oceanic layer, is ~ 5 k m thick and intrus ive vo lcanic in or ig in, consist ing Chapter 1. INTRODUCTION 6 of p redominant l y unaltered gabbro ic rock w i t h cumulate maf ic gabbro at the base. It is often d iv ided in to 3A and 3B , where 3 A ranges in ve loc i t j ' f r om 6.5-7.1 km/s and is 2-3 k m thick and overlies 3B wh i ch has velocities f rom 7.1-7.7 km/s and is ~ 3 k m th ick (Fox and St roup, 1981). In most seismic refraction surveys, velocities i n this layer are the best constra ined because arr ivals refracting through this layer are cont inuous for over 10 k m of source-receiver distances and the velocity gradient is low, contrast ing marked ly w i t h the steep, well-defined ve loc i ty gradients associated w i t h the upper 2 k m of the oceanic crust. Some surveys f i nd a basal layer of low-velocity i n Layer 3 (Lewis and Snydsman, 1977) wh i le others find a h igh-velocity one, but evidence for a layer w i t h a different ve loc i ty is tenuous ( Spud i ch and Orcu t t , 1980a). T h e depth to the Mohorov i c i c d i scont inu i ty (Moho) is h igh ly var iable. Sometimes the M o h o is a sharp velocity d i scont inu i ty, whi le i n other places the t rans i t ion is much smoother, occur r ing over 1-2 k m . Th i s var iab i l i ty is also found i n the dep th over wh ich the change f rom maf ic to u l t ramaf i c rocks occurs i n ophio l i te suites (Wh i t e , 1984). Ve-locit ies increase across this t rans i t ion to the t yp i ca l values of 8 km/s found i n the upper mant le. Th i s s imple layered structure has worked well i n the past, but the layer ing is p r i nc i -paUy an art i fact of the acquis i t ion procedures, the sparseness of the da ta sets, and the methods used to interpret the d a t a sets (Spudich and Orcut t , 1980a; P u r d y and E w i n g , 1986). T h e resolut ion of layers by refract ion da ta is l im i ted because the thicknesses of the seismic layers obta ined by mode l l i ng are approaching the wavelengths of the seismic waves used to invest igate the c rus ta l structure. Few int racrus ta l reflectors that wou ld corroborate the layered structure have been seen, a l though the values l i sted i n Tab le 1.1 are consistent w i t h substant ia l ref lection coefficients of 0.22, 0.08, and 0.06. Mo re recent surveys us ing the mu l t i channe l seismic technique have been able to resolve some interna l layer ing and var iat ions in crusta l thickness (e.g., Herron et ai, 1978; Ha le et ai, 1982; Chapter 1. INTRODUCTION 7 M u t t e r and N A T S tudy G r o u p , 1985). Deta i l ed refract ion and reflection t ravel - t ime and amp l i tude da ta interpreted us ing state-of-the-art procedures are required to produce models more representative of the actua l var iat ions w i th i n the oceanic crust. 1.2.2 Characteristics of Ocean Spreading Centres One of the most surpr i s ing results of a l l the different types of studies of oceanic crust is the s imi la r i ty of crust i n a l l ocean basins, even though gross morpho log ica l differences exist at spreading ridges, the source of the crust. Ocean spreading ridges are d i v ided into three types based on their fu l l spreading rates: slow, 1.0-5.0 cm/a; intermediate, 5.0-9.0 cm/a; and fast, 9.0—18.0 cm/a (Macdona ld , 1982). Th i s spreading rate appears to contro l the gross morpho logy of the r idge by contro l l ing the magmat i c and the rma l budget. S low-spreading ridges, such as the M i d -A t l a n t i c R idge ( M A R ) , have a 1.5-3.0 k m deep rift valley ma rk i n g the axis plus rough and faulted topography (F igure 1.2). Intermediate-spreading ridges, such as the J u a n de F u c a R idge ( J D F R ) and the East Pac i f i c R i se ( E P R ) at 21°N, have a shal low rift val ley on ly 50-200 m deep super imposed on a b road ax ia l h igh and smooth f lank ing topography. Fast-spreading ridges, such as the E P R at 3°S, have no rift valley, but instead have a t r iangular , semi-circular, or rectangular-shaped ax ia l h igh and relat ively smooth topography w i t h a fine-scale horst and graben structure (Macdona ld , 1982; 1986). E v e n w i t h these large-scale differences, the fine-scale crusta l structure and tectonics of spreading ridges evolve i n a pattern that is largely independent of spreading rate (Macdona ld , 1982). Th i s pat te rn consists of three zones: a neovolcanic zone, a fissured zone, and an active fau l t ing zone (F igure 1.2). Spreading rates p r imar i l y contro l the cont inu i ty in t ime and space of these three zones, but this apparent pa r t i t i on ing of the ridge does not necessarily i m p l y that the volcanic and tectonic processes that are responsible for it are steady-state (Sempere and M a c d o n a l d , 1987). Chapter 1. INTRODUCTION 8 F igure 1.2: H i gh resolut ion deep-tow profiles of the M i d - A t l a n t i c R idge rift valley i n the F A M O U S area (bot tom) w i th profiles of intermediate and fast-spreading parts of the East Pac i f i c Rise for comparison. Neovolcanic zone bracketed by V s , zone of Assur ing by Fs, and plate boundary zone (w id th of active fault zone) by PBs. (F igure and capt ion f rom M a c d o n a l d , 1986.) Chapter 1. INTRODUCTION 9 T h e neovolcanic zone is the region of recent and ongoing vo lcan i sm at the spreading centre. Th i s area is very narrow (1-2 k m wide) and characterized by fresh, glassy lava flows and very l i t t l e sediment cover. T h e cont inu i ty of the volcanoes a long the spreading centre and the type of volcanic extrusives are dependent on the spreading rate. A t s low-spreading rates, the ax ia l zone is marked by a h igh ly discont inuous chain of central volcanoes ~ 2 5 0 m high wh ich consist of p i l l ow basalts. A t intermediate-spreading rates, the higher cont inu i ty of the central volcanoes is i n te r rupted by smal l (<1 km) en echelon offsets; sheet flow basalts are more frequently observed (Francheteau and Ba l l a r d , 1983) and the central volcanoes reach on ly 50 m i n height. A t fast-spreading rates, the central volcano is remarkab ly continuous and is on ly i n te r rupted by t ransform faults. B o t h p i l low lavas and sheet flows are observed. F l a n k i n g the neovolcanic zone, a 1-3-km-wide zone of intense Assur ing occurs. These fissures are 1-3 m wide and extend 10 m to 2 k m along strike, para l le l l ing the ridge (Macdona ld , 1982; M a c d o n a l d and Luyendyk , 1985). A t s low-spreading rates, fissures are also observed along the spreading axis, but are obscured by the centra l volcanoes at faster rates. These fissure systems l ikely prov ide the p r i nc ipa l access for cold seawater to penetrate the young, hot oceanic crust (L ister, 1972; 1974; 1982) and submers ible surveys have found that hyd ro the rma l act iv i ty, i n c l ud i ng h igh termperature vents such as b lack smokers, is usual ly located w i t h i n fissures and along the base of fault scarps ( R yan , 1986). A t a distance of 2-6 k m f rom the spreading axis, some fissures develop large vert ica l offsets by no rma l fau l t ing , w i t h most of the faults d i pp i ng toward the spreading axis (Macdona l d , 1982; Sempere and M a c d o n a l d , 1987). T h e vert ica l throw on these faults is 200 n i or more for s low-spreading ridges, result ing i n the deep rift valley. For intermediate or fast-spreading ridges, the faults have throws of 50 m or less, fo rming a shal low rift val ley or no rift at a l l . Ac t i ve fau l t ing occurs up to 30 k m off-axis in s low-spreading Chapter 1. INTRODUCTION 10 ridges but on ly up to 10 k m off-axis for faster spreading ridges (Macdona ld , 1982). T h e region where fau l t ing stops marks the edge of the p late boundary zone and the beg inn ing of the rigid plate. 1.2.3 Magma Chambers T h e as sumpt ion of magmat i c intrus ions at spreading centres has been extremely useful i n exp la in ing the format ion of the pseudo-layered structure shown in F i gu re 1.1. In this hypothesis, c rus ta l volcanic and p lu ton ic rocks are formed by f ract ionat ion of mant le-derived parent magmas i n a shal low crustal m a g m a chamber rather than by in ject ion and erupt ion d i rect ly f rom the mant le (Macdona ld , 1982). T h i s chamber provides the source f rom wh ich the c rus ta l section forms by the extrus ion of mid-ocean ridge basalts through the roof and the format ion of the layered lower crust v i a processes of d i f ferent iat ion and crysta l sett l ing (Orcu t t , 1987). Ev idence for an a x i a l m a g m a chamber has been obta ined f rom analyses of ophiol ites ( Cann , 1974; Pa l l i s ter and Hopson, 1981), petro log ic studies of ocean rocks ( B r y an and Moore, 1977; Stakes et ai, 1984), therma l models (Sleep, 1975; 1978; Sleep and Rosendahl , 1979; Mo r t on and Sleep, 1985a), and seismic experiments (O rcu t t et al, 1975; Hale et ai, 1982; M c C l a i n et ai, 1985). T h e first seismic refraction s tudy that could be interpreted in terms of a magma chamber was that by Orcut t et al. (1975) on the E P R at 8°N. Further experiments on the E P R prov ided more evidence for such a m a g m a body (e.g., Re id et ai, 1977; Herron et ai, 1980; Stoffa et al, 1980; M c C l a i n et al, 1985). T h e strongest evidence for a magma chamber has been obta ined on the E P R at 13°N. A crusta l magma, chamber ~1 .5 k m below the seafloor was postu lated by Orcut t et a/.-(1984) based on the at tenuat ion of refracted arrivals para l le l to the rise crest. U s ing two-d imens iona l ray t rac ing, M c C l a i n et al. (1985) determined the chamber to be ~ 4 k m wide and d iamond-shaped i n cross section. F r om the interpretat ion of fan and obl ique l ines. Bu rnet t et al. (1985) showed the Chapter 1. INTRODUCTION 11 chamber was cont inuous along-strike. Further to the south at 11°20 'N , f rom an analysis of compress ional and shear wave arrivals f rom the R i ve ra Ocean Seismic Exper iment ( R O S E ) , B r a t t and Solomon (1984) inferred that any mo l ten m a g m a chamber must be sma l l and i so lated because the observed arrivals were not s ignif icant ly attenuated wh i le propagat ing across the rise. However, this site is near a ridge axis d i scont inu i ty (an over lapp ing spreading centre) wh ich indicates a compl icat ion relat ive to the n o r m a l spreading process. Ref lect ion da ta also ind icate the presence of ax ia l magma chambers beneath the E P R at depths of 2 to 6 k m . A t 9°N on the E P R , a 4 -km wide ref lect ion at a depth of 1.5 to 2 k m below the seafloor was found to have a po la r i ty reversal, i nd i ca t i ng a l i qu id m e d i u m below the reflector (Herron et al, 1978; Her ron et al, 1980; Ha le et al., 1982). Between 9°N and 13°N, this magma-chamber-roof reflection is continuous a long strike except for large axis offsets (Detr ick et ai, 1987). A t 11°20 'N, the refract ion work of B r a t t and So lomon (1984) constrains the m a g m a chamber to narrow dikes and sma l l pockets w i t h a vert ica l thickness of less than 1 km. Other reflectors wh ich are poss ible ax ia l m a g m a chamber roofs have been found on the Va l u Fau R idge i n the L a u back-arc basin ( M o r t o n and Sleep, 1985b), on the southern J D F R (Mo r t on et ai, 1987), and on the Endeavour Segment of the J D F R (Rohr et al, 1988). In contrast, refraction experiments on the J D F R ( M c C l a i n and Lewis, 1982) and the M A R (Poehls, 1974; P u r d y and Detr ick, 1986) have not found evidence for a m a g m a chamber, even though petrologic da ta require one (Stakes et al., 1984). Reflections f r om w i t h i n the inferred p luton ic section of the oceanic crust in the No r t h A t l an t i c Transect ( N A T ) s tudy have been interpreted to be a remnant of the or ig ina l magma chamber on the M A R , but the spreading rate i n the past may have been faster (Mu t te r and the N A T S tudy G r o u p , 1985; N A T S tudy G roup , 1985). Moores et al. (1986) suggest that these reflections may be a basal detachment caused by l i s t r ic faults. Chapter 1. INTRODUCTION 12 Indirect evidence for ax ia l m a g m a chambers is obta ined f rom bathymet r i c data, heat flow data, and submersible observations. Francheteau and Ba l l a rd (1983) postu lated that regions of bathymetr ic highs are i sostat ica l ly up l i f ted by the presence of a magma reservoir i n the crust. Th i s m a g m a reservoir then migrates away f rom the topographic h igh i n pulses to form an arcuate along-axis ba thymet r i c high. Fresh lava flows and a higher p ropor t ion of sheet l ava flows (f lu id lavas) as observed i n submers ib le studies occur at these topographic highs and ind icate recent vo lcan i sm (Francheteau and Ba l l a rd , 1983) . T h e h igh water temperatures obta ined f rom geothermal measurements over ridges require a steady heat source such as a m a g m a chamber, wh i le large sulf ide deposits such as those found in the E P R ax i a l valley also require a long- l ived heat source (Francheteau and Ba l l a r d . 1983). T h e detect ion of ax ia l m a g m a chambers on the E P R but not on s lower-spreading ridges has led to questions about the l i fet ime of a m a g m a chamber (e.g., Stakes et ai, 1984) . T h e seismic refraction evidence suggests that m a g m a chambers are steady-state features beneath intermediate and fast-spreading centres, but they may vary i n size and shape w i t h t ime (Macdona ld , 1986). Some the rma l and petrologic studies concur w i t h this assessment, others suggest steady-state m a g m a chambers under a l l ridges, whi le others suggest no steady-state magma chambers at a l l . F r om considerat ion of the cool ing effects of h yd ro the rma l c i rcu lat ion th rough fissures, some the rma l models predict a temporary m a g m a chamber at s low-spreading rates, but predict a steady-state magma chamber at intermediate- and fast-spreading rates (Sleep, 1975; 1978; 1983; Sleep and Rosendah l , 1979). L i s ter (1983), however, pred icted that hyd ro the rma l c i rcu lat ion th rough sub-vert ical cracks should cool a m a g m a chamber so qu ick ly that no steady-state m a g m a chamber could exist, whi le K u z n i r (1980) obta ined a steady-state model even at s low-spreading rates. These differences stem f rom poor ly constra ined estimates for the magn i tude and areal d i s t r i bu t ion of hyd ro the rma l heat loss Chapter 1. INTRODUCTION 13 at spreading centres and ignorance of the rate and depth of crusta l Assur ing (Macdona ld , 1986). T h e Sleep (1975) steady-state mode l is supported by studies of the morpho logy of the neovolcanic zone (Macdona ld , 1986). T h e ax i a l shield volcano at fast-spreading rates is remarkab ly continuous and at un i f o rm depth, a morpho logy that wou ld be diff icult to ma in t a i n w i thout a steady-state magma chamber and frequent eruptions. In con-trast, s low-spreading ridges have h igh ly d iscont inuous ax ia l volcanoes of vary ing height and morphology (Macdona ld , 1982). T h e en echelon character of the central volcanoes i n intermediate-spreading ridges also suggests an under l y ing cont inu i ty i n the m a g m a source. Petro log ic evidence suggests tha t the ax ia l m a g m a chamber is quasi-steady state, even at s low-spreading rates (B ryan and Moore, 1977; Stakes et al., 1984). Basa l t samples f r om the M A R show l i t t l e compos i t iona l var iat ion; i f m a g m a chambers were not steady-state, a m u c h greater range of basalt compos i t ions wou ld exist since very p r im i t i ve basalts wou ld be obta ined at the beginning of a m a g m a chamber ' s l i fet ime and h igh ly f ract ionated basalts at its end. T h e contradict ions between seismic, petrologic, and t he rma l results, especially on s low-spreading ridges, have led to many different shapes and sizes of ax ia l magma cham-bers (F igure 1.3). T h e infinite onion mode l of C a n n (1974; F i gu re 1.3b) i n wh ich the edges of a steady-state ax ia l m a g m a chamber successively freeze, brought, accord ing to M a c d o n a l d (1986), "tears to the eyes of seismologists who had found no evidence of an ax ia l m a g m a chamber anywhere a long the M A R " . N i sbet and Fowler (1978) proposed the infinite leek mode l (F igure 1.3c) i n wh ich se ismical ly-undetectable m a g m a chambers are repeatedly formed and crysta l l ized; however, the h igh ly f ract ionated lavas that wou ld be p roduced upon freezing have not been found i n detai led petrologic sampl ing programs. F r om petrologic data. Stakes et al. (1984) have proposed a funnel mode l (F igure 1.3a) Chapter 1. INTRODUCTION 14 A. FUNNEL—SHAPED SAMAIL OPHIOLITE MODEL V E = |:| Dikes 11111' 11 ± L S HJ N 111111 -Gobbros V tV t+X\AWtM 1 3 Km C. "INFINITE L E E K " MODEL Temporary High Level Chamber .+,+ .•+ •+ + fW> + + + + + + + ,+,+ +, , 1—' + + + ,+ + B. " INFINITE ONION" MODEL D. ALONG STR IKE VIEW OF MAGMA CHAMBER FOR MODELS A,B Transform Foulf Primitive or Fract ionated Basal ts ji Homogeneous Basal ts Axial Nebvolcanic Zone - Ephemeral Magma Chamber Steady State Mogmo Chamber Inf lated S ta te Regions of S m a l l , Ephemera l Magma C h a m b e r s I— 1 ~ 5 0 Km F igure 1.3: Three of the possible models for ax ia l m a g m a chambers for s low-spreading ridges. ( A ) M o d e l based on observations in the Samai l oph io l i te in O m a n after Pa l l i s ter and Hopson (1981), (B ) the infinite onion mode l after C a n n (1974) and B r y a n and Moo re (1977), ( C ) the infinite leek mode l after N i sbet and Fowler (1978). T h e Sama i l and infinite onion models predict a steadj ' -state m a g m a chamber (S.S.M.C.). A s drawn, bo th chamber models should be seismical ly detectable, bu t not i f they are less than 1-2 k m wide. M o d e l (C ) predicts a non-steady state m a g m a chamber; see text for details. Poss ib le subsurface extensions of no rma l faults in to the dike and gabbro layers are not shown for s impl ic ity. (D ) shows a possible along-str ike section for the stead}'-state m a g m a models. A l o n g s low-spreading ridges such as the M A R , a m a g m a chamber may be conf ined on ly to ax ia l highs away from t rans form fault intersections. (F igure and capt ion f rom Macdona l d , 1986.) Chapter 1. INTRODUCTION 15 i n wh ich steady-state replenish ing and cont inued m a g m a m i x i n g account for the un i f o rm basalt compos i t ion. B y shr ink ing this chamber to only 1-2 k m wide, it wou ld not be detectable by seismic experiments. Stakes et al. (1984) also proposed that the ax ia l m a g m a chamber on ly exists away f rom t rans form faults i n regions of elevated regions of the inner f loor of the rift val ley (F igure 1.3d); this explains the discont inuous nature of the neovolcanic zone at s low-spreading rates. T h e shape of this magma chamber is very s imi lar to that obta ined by M c C l a i n et al. (1985) f rom mode l l i ng of refract ion da ta obta ined on the E P R . T h e size and shape of the ax i a l m a g m a chamber are impo r tan t contro l l ing parameters of the petrology, structure, and s t rat ig raphy of the oceanic crust. T h e uppermost layer of extrus ive volcanics is most l i ke ly formed by erupt ive cycles of sheet flows fo l lowed by p i l low lavas. These cycles are p robab l y of short durat ion (1-100 a), separated by long periods of quiescence (Macdona ld , 1982); evidence for these cycles exists f rom analogy w i t h terrestr ial eruptions and the thickness of crusta l units i n D S D P holes (Macdona ld , 1986; A d a m s o n , 1985). T h e frequency of these cycles increases as approx imate ly the square of the spreading rate (Macdona ld , 1982). T h e to ta l thickness of the volcanic section (sheet flows and p i l low lavas) is control led by the depth to the magma chamber, whi le the extent of the p lu ton ic gabbro section is determined by the thickness of the m a g m a chamber. T h e sheeted dikes are formed by inject ion of m a g m a that fails to reach the surface. M a g m a chamber so l id i f icat ion apparent ly involves " p l a t i n g " of gabbros down f rom the roof wh i le gabbroic and u l t ramaf ic cumulates deposit progressively upward f rom the floor (Macdona ld , 1982). T h e downward and upward-so l id i fy ing parts of the chamber meet i n a " sandwich zone" of the most compos i t iona l l y evolved minerals of the melt. Chapter 1. INTRODUCTION 16 1.2.4 Three-dimensionality of Spreading Centres A l t h o u g h man}' studies of oceanic crust near spreading centres have been interpreted assuming one-dimens ional structure, this a s sumpt ion is not va l id . T h e assumpt ion that spreading centres are continuous along strike wou ld produce a mode l that is two-d imens iona l : however, three-dimens ional effects such as morpho log ica l changes in the neovolcanic zone, ridge offsets, changes in crusta l thickness, and veloc ity anisotropy have a l l been observed. T h e neovolcanic zone, especial ly on s low-spreading ridges, often has gaps w i t h no vo lcan ic edifices. H yd r o t he rma l ac t i v i t y may consist of vigorous h igh- temperature vents, low-temperature c i rcu lat ion, or no th ing at a l l (Macdona ld , 1982). T h e shape of the rift val ley can vary a long strike due to changes i n fau l t ing . Some regions of the neovolcanic zone are covered w i t h p i l low lavas, others w i t h sheet flows. M a n y of the observations can be accounted for by a mode l i n wh ich the frequenc}' of vo lcanic and associated hyd ro the rma l episodes is dependent on the spreading rate (Macdona l d , 1982). T h e presence of t ransform faults is an obvious cont rad ic t ion of the assumpt ion of two-d imens iona l spreading centres. C ru s t near fracture zones has been found to be anomalous ly t h i n , miss ing the Layer 3 refractor, and has lower crusta l velocities because of fau l t ing and pervasive hyd ro the rma l c i rcu lat ion. T h e existence of this anomalous crust is independent of the amount of offset across the f racture zone (Detr ick and Purdy . 1980; W h i t e et al., 1984; Ab r ams et ai, 1988). H igh-reso lut ion mapp ing of mid-ocean ridges by swath inst ruments such as Seabeam and S e a M A R C I and II has revealed the existence of new families of r idge axis d iscont inu-ities wh i ch can accommodate smal l offsets (0-25 km) "a long the strike of spreading centres. These discontinuit ies inc lude propagat ing rifts and over lapp ing spreading centres; they are prevalent on fast and intermediate-rate spreading ridges, but not on s low-spreading Chapter 1. INTRODUCTION 17 ridges (Macdona ld , 1983). A l ong w i t h t ransform faults, smaller ridge axis d iscontinuit ies may be the surf ic ia l expression of the segmentat ion of mid-ocean ridges into geochemical ly d i s t inct spreading cells (Sempere and M a c d o n a l d , 1987). P ropaga t i ng rifts are spreading centres wh ich propagate along str ike th rough plates, creat ing a V - shaped wake in the i r pa th . T h e y offer an exp lanat ion for the relocat ion and propagat ion of spreading centres and prov ide a mechan i sm for spreading centres to change az imuth i n response to changes i n the or ientat ion of plate mot ions (Hey et ai, 1980). P ropaga t i ng rifts have been used to exp la in the obl ique magnet ic anoma ly offsets and p late boundary reorganizations i n the northeast Pac i f i c (Hey et ai, 1980; Hey and W i l s o n , 1982). T h e largest of the non-transform offsets of the neovolcanic zone are cal led over lapping spreading centres (Macdona ld and Fox, 1983) and are ub iqu i tous between 23°N and 23°S on the E P R (Macdona ld and Fox, 1983; M a c d o n a l d et al., 1984; Lonsdale, 1983; 1985). The}- have a character ist ic geometry wh i ch consists of two over lapping curv ing ridges separated by an elongate depression. T h e overlap bas in, wh ich can be up to 600 m deep, contains no evidence of str ike-sl ip fau l t ing that cou ld accommodate the offset in a classical way (Sempere and M a c d o n a l d , 1987). M a c d o n a l d et al. (1984) suggested that over lapp ing spreading centres are the surf ic ia l expression of the interact ion between magmat i c pulses propagat ing beneath spreading centres wh i ch fa i l to meet head on. A s well as determin ing the depth to possible m a g m a chambers, mu l t i channe l seismic techniques are also able to map w i t h h i gh resolut ion the structure of oceanic crust such as i n te rna l layer ing and variat ions i n crusta l thickness (Herron et ai, 1978; N A T S tudy G r o u p , 1985; R o h r et al, 1988). Ref ract ion surveys a long the E P R have determined tha t the veloc ity structure of the oceanic crust is la tera l ly homogeneous on the scale of 2-3 k m but the thickness of the upper vo lcanic layer varies by hundreds of metres over distances of several tens of k i lometres (Purdy, 1982b; Get t rus t et al., 1982; B r a t t and Chapter 1. INTRODUCTION 18 Pu rdy , 1984). An isotropy, bo th of the oceanic crust and of the uppermost mant le, has been observed i n many seismic refract ion surveys. Results f r om special ly-designed borehole seismic refract ion exper iments reported by Stephen (1985; 1986; 1988) i n the western A t l a n t i c and in the eastern Pac i f i c ( D S D P Hole 504B) show az imutha l seismic anisotropy i n the upper oceanic crust. P-wave velocity var iat ions f r om 4.0-5.0 km/s were observed, w i t h the fast d i rect ion perpend icu la r to the fossil spreading d i rect ion and para l le l to large-scale cracks and fissures. F r o m the examina t i on of bo th the P-wave and the shear wave data f rom a seismic refract ion survey i n the south Pac i f i c wh ich used bo th O B S and borehole receivers, Shearer and Orcu t t (1985; 1986) found evidence for a z imutha l anisotropy in b o t h the upper crust and the upper mant le, w i t h the fast d i rect ion in the crust orthogonal to the fast d i rect ion in the mant le. T h e upper crust P-wave velocity var iat ions were about 0.2 to 0.4 km/s whi le the upper mant le velocities varied between 8.0 and 8.5 km/s. An i so t ropy has also been documented i n analyses of samples f rom ophiol i tes (Chr i s tensen and Smewing, 1981; Smewing , 1981; Chr i s tensen, 1984a); upper mant le anisotropy is caused by preferential a l ignment of o l iv ine crystals. T h e contrad ict ion between the t ime-dependent three-d imens ional view of spreading centres and the one-dimens ional c rusta l s t ructure usua l ly assumed near a spreading cen-tre can only be solved by deta i led areal surveys of a smal l region. On l y by mapp ing systemat ic changes i n structure can i n fo rmat ion about crustal processes be obta ined. To prov ide i n fo rmat ion on c rus ta l creat ion, the locat ion , size, and shape of any crustal m a g m a chambers must be determined. Var ia t ions i n crusta l s t ructure perpendicu lar to the r idge suggest evolut ion of the crust w i t h age whi le var iat ions para l le l to the ridge show changes a long strike. T h e detai led three-d imens ional survey over the J D F R wh ich is described in this thesis was designed to obta in i n fo rmat ion about these variations. Chapter 1. INTRODUCTION 19 1.3 The Endeavour Ridge Segment, Juan de Fuca Ridge 1.3.1 Regional Structure T h e J u a n de Fuca R idge system is a group of act ive spreading centres wh i ch extends f rom Cape Mendoc i no off nor thern Ca l i f o rn ia to the southern t ip of the Queen Cha r l o t te Islands and consists of three sections: the G o r d a R idge bound ing the G o r d a P l a te , the J u a n de Fuca R idge bound ing the J u a n de Fuca P l a te , and the Exp lo re r R idge bound ing the Exp l o re r P l a te (F igure 1.4). T h e 500-km-long J u a n de Fuca R idge ( J D F R ) is located between 44°N and 49°N and has an intermediate fu l l spreading rate of 6 cm/a ( R i d d i -hough, 1984), wh i le i n the hotspot frame of reference i t is mov ing west at 2 cm/a (Davis and Ka r s ten , 1986). T h e J D F R is the boundary between the large Pac i f i c P l a t e to the west and the much smaller J D F P l a t e to the east wh ich is subduct ing beneath western N o r t h Amer i ca . A l t h o u g h its spreading is s ymmetr i c as defined by magnet ic anomaly patterns, majo r a symmetry exists across the J D F R (Rohr et al., 1988). T h e J D F P l a te is heavi ly sedi-mented by Ple istocene turb id i tes , whi le the much larger Pac i f i c P l a te has l i t t l e sediment and many seamounts and seamount chains (Davis and Kar s ten, 1986). T h e spreading r idge presently consists of several r idge segments, i nc lud ing propagat-ing rifts and over lapp ing spreading centres, each approx imate ly 100 k m long (Ka r s ten et ai, 1986). T h e tectonic evo lut ion of the J D F R is the direct result of rift p ropagat ion wh ich caused the reorganizat ion of the Pac i f i c -Fara l l on p late boundary in the last 9 M a (W i l s on et ai, 1984). One of these propagat ing rifts is the C o b b offset (F igure 1.5) wh ich has evolved by a series of no r thward and southward propagat ing events (Johnson et ai, 1986). Net propagat ion has been to the no r th and appears to be discont inuous and relat ively r ap i d compared to the spreading rate (Johnson et ai, 1983). F igu re 1.4: T h e J u a n de Fuca R idge system. Ar rows ind icate relat ive directions of mot i on . Sma l l letters ind icate locations of exper iments for wh i ch results w i l l be discussed i n this thesis: (a) Dav i s et ai, 1976; (b) A u and Clowes, 1982; (c) M c C l a i n and Lewis, 1982; and (d) this study. Chapter 1. INTRODUCTION 21 131°W 130°W 129°W 128°W 127°W F i gu re 1.5: T h e northern J u a n de Fuca Ridge. Shaded contours ind icate pos it ive mag-netic anomaly patterns (Raff and Mason , 1961). O p e n s t ipp l i ng identif ies those anomalies inferred to have been produced by spreading on the Endeavour Segment, wh ich extends f rom the C o b b Offset to the Sovanco Fracture Z o n e - N o o t k a Fau l t t r ip le j unc t i on . T h e E n -deavour Segment is d iv ided into two sections b}' the Endeavour Offset. B o l d north- south lines represent the current pos i t ion of the spreading axis, w i t h the fo l lowing sections label led: NSR -Northern Symmet r i ca l R idge, E R - Endeavou r R idge, W V -Wes t Val ley, and M V - M i d d l e Valley. Dashed lines ind icate the trace of the pseudofault produced by the no r thward -mig ra t ing C o b b P ropagato r (Hey et al., 1980). T h e major seamounts in the r e g i o n — t h e Heck Seamount cha in , the Heck le Seamount chain, and the Spr ingf ie ld, Exp lorer , and C o b b Seamount s—a re ind icated by the i r bathymetry. T h e d i amond shape over the Endeavour R idge is the explos ion array for the S E I S R I D G exper iment shown in F i gure 2.1. (F igure and capt ion f rom Ka r s ten et ai, 1986.) Chapter 1. INTRODUCTION 22 1.3.2 Geomorphology of Endeavour Ridge A majo r offset in the J D F R occurs at the C o b b Offset, separat ing the Endeavour Segment to the nor th f rom the rest of the ridge. Th i s offset also defines two d i s t inct m a g m a sources: the basalts i n the Endeavour Segment are potass ium-enr iched whi le those to the south are potass ium-def ic ient (Delaney et ai, 1986). T h e Endeavour Segment is itself subd iv ided at the Endeavour Offset i n the v i c i n i t y of Endeavour Seamount, the easternmost por t ion of the Heck Seamount chain (Ka r s ten et al., 1986). South of Endeavour Offset, the spreading centre consists of the shal low Endeavour R idge segment wh ich deepens into two broad grabens to the nor th and s o u t h — N o r t h Endeavour Va l ley and South Endeavour Valley. N o r t h Endeavour Va l l e j ' overlaps the West Va l ley segment whi le South Endeavour Va l ley overlaps the Nor thern S ymmet r i ca l segment, bo th offsets being over lapping spreading centres. T h e shallowest por t ion of the ridge lies at the projected intersect ion of the Heck le Seamount cha in w i t h the spreading axis near 47° 58 'N (Kar s ten et ai, 1986) and is a known site of hyd ro the rma l ac t i v i t y (T i vey and Delaney, 1986). T h e geomorphology of the Endeavour Segment has been examined by a number of geophys ical methods i nc lud ing Seabeam, h igh frequency (100 k H z ) side-scan sonar, deep-tow camera/video (a l l used by Ka r s ten et al., 1986), S e a M A R C I (C rane et ai, 1985), S e a M A R C II (Davis and Cur r ie , 1985), and single-channel seismic reflection (Johnson et al., 1983). H i gh sediment accumula t ion due to the p rox im i ty to No r t h A m e r i c a causes some diff iculty i n interpretat ion of these records. F r o m these surveys, the ridge axis along the entire Endeavour Segment has been found to consist of a relat ively narrow (<10 km) crestal ridge super imposed on a 30 to 40-km-wide broader swell of young oceanic crust ( K a p p e l and R y a n , 1986). Th i s broader swell suggests that this part of the ridge axis has had long-term enhancement of its m a g m a supply relat ive to other parts of the ridge (Kar s ten et ai, 1986). A n inner rift Chapter 1. INTRODUCTION 23 or elongate summi t depression ( K a p p e l and R y a n , 1986) of variable w i d t h (~1 km) and depth (~ 100 m) bounded by steep, inward-fac ing scarps exists at the s ummi t of the ridge. These scarps are interpreted as traces of no rma l faults (Kars ten et ai, 1986) and form as a result of tectonic stretching w i t h wan ing vo lcan i sm (F rank l in and K a p p e l , 1986). T h e topography through wh ich this rift extends varies d ramat i ca l l y a long str ike f rom a narrow 4-km-wide, ax ia l h igh at Endeavour R idge (shallowest depth 2.1 k m ) to the b road (10 k m wide), deep (2.7-3.0 km) basins of N o r t h Endeavour Va l ley and Sou th Endeavour Va l ley (Kar s ten et al., 1986). T h e morphology of the Endeavour R idge at its shallowest po r t i on is s imi lar to that at ~ 1 3 ° N on the E P R , a region w i t h a fu l l - spreading rate of 11-12 cm/a (Francheteau and B a l l a r d , 1983). Na r row ridges and valle} rs s ymmetr i ca l l y f lank the ax ia l r idge at intervals of approx imate ly 5 km; K a p p e l and R y a n (1986) suggest that these, ridges are remnants of former ax ia l ridges t ransported re lat ive ly intact f rom the spreading axis. Delaney et al. (1986) state that these structures represent a per iod of near ly a m i l l i on years dur ing wh i ch the output of the Endeavour R idge m a g m a source was higher. T h e ax ia l val ley of the Endeavour R idge is wel l defined by a narrow zone of cont inuous faults and Assuring. Th i s is the zone of Assuring discussed by M a c d o n a l d (1982). T h e dens i t j ' of these r idge-paral le l cracks in the rift val ley is measured to be greater than one every 50 m, w i t h the cracks fo rming an en echelon pattern ( K a p p e l and R y a n , 1986). N o r t h of 47° 56 'N , the fissures die out and the floor is not wel l l ineated, i nd i ca t ing larger-scale vo lcanic roughness (Crane et ai, 1985); deep-tow photograph ic coverage near 48° 0 0 ' N shows that the ax ia l val ley is dominated by l ight ly -sed imented, young, lobate and p i l low flows. Th i s sediment cover dates the flows i n the ax ia l valle3 T at ~ 1 0 000 a (Davis et ai, 1984b). To the south, the fissured floor dominates the tectonics of the region. T h e zone of recent vo lcan i sm, the neovolcanic zone, appears to be restr icted to the floor of the narrow inner r ift (Kar s ten et ai, 1986), a l though there has been l i t t le Chapter 1. INTRODUCTION 24 vo lcan ic ac t i v i t y since the crestal r idge was bu i l t ( F rank l in and K a p p e l , 1986). Ben th i c communi t ies and sulf ide boulders associated w i t h hyd ro the rma l vents are also present (T i vey and Delaney, 1986; Ka r s t en et al, 1986) and appear to He at the j u n c t i o n of the fissured and unf issured rift val ley floor (Crane et al, 1985). T h e flanks of the crestal r idge consist of over lapp ing asymmetr ic bu lbous mounds capped w i t h p i l l ow lavas that have an inter -p i l low sediment cover. F lows w i t h a surface area of more t han 4 k m 2 , interpreted to be f rom a single erupt ive event, have been mea-sured on these flanks. T h e almost to ta l lack of observable f ractur ing ( K a p p e l and R y a n , 1986; T i v e y and Johnson , 1987) suggests that the ax ia l r idge is a volcanic edifice; the fissures that wou ld have been created here have been covered by the effusive vo lcanism. F rac tu r ing again occurs more than 5 k m f rom the inner rift, at the base of the ax ia l ridge. These faults may be continuous for tens of k i lometres along strike and some have vert ica l d isplacements of more than 100 m. Th i s is the zone of active fau l t ing descr ibed i n Sect ion 1.2.2. 1.3.3 Previous Geophysical Studies T h e J D F R and Pac i f i c and J D F Plates are some of the most studied areas i n the ocean because of their p rox im i t y to N o r t h Amer i ca . T h e pioneer ing magnet ic s tudy of Raf f and Mason (1961) was done i n this area and many other geophysical studies have been carr ied out since then. T h e exper iments that are impo r tan t to the present s tudy area are ment ioned below, but any relevant in fo rmat ion determined f rom them w i l l be discussed later in the context of the present study. Seismic refract ion surveys have been carr ied out in a number of areas on the J D F P l a t e and the J D F R (F igure 1.4). No r t h of Endeavou r Segment at the Sovanco Trans-fo rm, Dav i s et al. (1976) d i d one of the first refraction surveys us ing ocean b o t t o m seismographs, f i nd ing low velocity mant le beneath the ax ia l valley of the ridge. Fur ther Chapter 1. INTRODUCTION 25 south, us ing deep-tow reflection da ta and refract ion data, M c C l a i n and Lewis (1982), found no evidence for a low veloc ity zone. Other surveys have been done on older crust i n the area, i nc l ud ing one carr ied out near the N o o t k a Faul t Zone by A u and Clowes (1982; 1984). M c M a n u s et al. (1972) undertook one of the first reflection surveys on the northern J D F R . M u c h more recent surveys inc lude those described by R o h r et al. (1988), in the centre of the present study area and p lanned w i t h knowledge of the scheduled refraction exper iment, and M o r t o n et al. (1987) to the south, bo th of wh ich obta ined reflections f rom a poss ible ax ia l magma chamber. T h e m a g m a chamber reflection obta ined by M o r t o n et al. (1987) is only 1-2 k m wide and 2.3-2.5 k m beneath the seafloor and wou ld p robab l y not be detected by a seimic refract ion exper iment such as that described by M c C l a i n and Lewis (1982). Heat flow7 anomalies have been found on the Endeavour R idge by a number of authors (De laney et al, 1984; C rane et al., 1985; T i vey and Delaney, 1986) and on the northern J D F R (Davis et al., 1980). These heat flow anomal ies inc lude hot springs w i t h water temperatures in excess of 400°C at 47° 59 'N , 129° 4.9 'N (Delaney et ai, 1984; T i v e y and Delaney, 1986). Geotherma l fields found by C rane et al. (1985) are located along major along-axis bathymetr i c highs. Ev idence for hyd ro the rma l act iv i ty f r om photographs and sulf ide deposits has been obta ined using submersibles and deep-tow inst ruments ( John-son and Delaney, 1984; Ka r s ten et al., 1986; T i v e y and Delaney, 1986). These large sulf ide deposits also require a heat source to susta in a hyd ro the rma l system (F rank l in and K a p p e l , 1986). H i g h resolut ion S e a M A R C and Seabeam da ta also exist in the s tudy area ( Ea r th Phys ic s B r anch , 1984; Davis et al., 1984a; Dav i s and Cur r ie , 1985) thereby enhancing the ab i l i t y to relate surface features to seismic velocity anomalies. G r a v i t y da ta (Yora th et ai, 1988) have also been obta ined over Endeavour R idge, but a. p re l im inary interpretat ion Chapter 1. INTRODUCTION (Rohr et al, 1988) indicates that no large density anomalies are present. Chapter 2 DATA ACQUISITION AND PROCESSING 2.1 The Refraction Experiment, SEISRIDG 85 F i e l d work for S E I S R I D G 85 was carr ied out i n October and November 1985 on the CSS Parizeau by the Un iver s i ty of B r i t i s h C o l u m b i a i n co l laborat ion w i t h the Pac i f i c Geoscience Centre. T w o d iamond-shaped areal arrays of ocean b o t t o m seismographs ( O B S ; Figures 2.1 and 2.2) were used to achieve the three ma i n objectives descr ibed i n Sect ion 1.1 by p rov id ing convent iona l reversed refract ion lines as wel l as mu l t i p l e fu l l a z imutha l coverage. A n outer array w i t h four outer O B S and s ix inner O B S w i t h a m a x i m u m d imen-sion of 60 k m along the ridge was deployed to determine the regional c rusta l s t ructure (F igure 2.1). L ines i n the outer array were prof i led us ing 153 90-kg explosive charges detonated at 2 k m intervals along a to ta l d i stance of 310 km. A l l but 80 k m of the outer array was also prof i led w i t h a 32 L a i rgun w i t h a shot spacing of approximately-0.125 km. These lines were or ig ina l ly scheduled to be shot w i t h two 32 L airguns as the source, but sea condit ions were too rough to deploy bo th airguns s imultaneously. T h e four outer O B S were then recovered and redeployed along the cross-ridge l ine to fo rm a dense l inear array of eight O B S across the ridge (F igure 2.2) so that sufficient ray coverage wou ld be obta ined at the depth of a possible m a g m a chamber for use in a tomograph ic invers ion scheme (Wh i te , 1988; W h i t e and Clowes, 1988). T h e lines i n this array were prof i led w i th a 32 L a i rgun w i t h a shot spacing of 0.200 k m w i th 27 Chapter 2. DATA ACQUISITION AND PROCESSING 28 F igu re 2.1: Conf i gu ra t ion of O B S and explos ion lines for the outer array. A s i n F igure 1.5, the open s t ipp l i ng identifies those magnet ic anomalies inferred to have been produced by spreading on the Endeavour Segment whi le bo ld lines again represent the current pos i t ion of the spreading axis. Explos ives were shot on a l l the lines shown; a irgun prof i l ing was done on those port ions of the explos ion lines tha t have been emphasized by bo lder lines. Chapter 2. DATA ACQUISITION AND PROCESSING 29 129 20 15 48 10H t-10 48 00 + 55 4-50 + 47 45 129 00 55 128 50 1 1 f-48 10 E n d e a v o u r A i r g u n L i n e s O B S 10 Kilometres 15 129 20 15 10 + 20 48 00 47 45 129 00 55 128 50 F i gu re 2.2: O B S posit ions and airgun lines for inner array. T h e four outer O B S f rom F igure 2.1 were p laced on the cross-ridge l ine. Chapter 2. DATA ACQUISITION AND PROCESSING 30 the except ion of L i ne 4 wh ich was shot at 0.125 k m spacing. T h e to ta l length of lines prof i led was 400 k m . Th i s conf igurat ion of O B S and transect lines produced reversed and unreversed lines obl ique to the ridge (L ines A G - 5 , A G - 6 , A G - 8 , and A G - 7 ) , reversed and unreversed lines perpend icu lar to the ridge (Lines A G - 2 , A G - 3 , and A G - 4 ) , unreversed lines para l le l to the ridge (Lines A G - 9 and A G - 1 0 ) as wel l as numerous fan shot lines. T h e oldest crust sampled i n this inner array is ~0.40 M a (cf. R o h r and M i l ke re i t , 1988) and has l i t t le or no sediment cover. Nav iga t ion dur ing the exper iment was obta ined us ing B I O N A V (Bedford Inst i tute of Oceanography Integrated Nav i ga t i on System; G ran t , 1980) wh ich integrates passive rang ing (p — p L o r a n - C ) and Trans i t Satel l i te Nav i ga t i on ( SatNav) ; Sa tNav fixes are used to update the less accurate but more frequently obta ined L o r a n - C posit ions. Nav i ga t i ona l pos i t ions were logged onto disks on a mic rocompute r every minute; satell ite pos it ions were also recorded for later use. Abso lu te Sa tNav posit ions have an accuracy of ± 1 2 0 m (Ea ton et ai, 1976) whi le absolute Lo ran -C posit ions are accurate to ± 2 7 0 m w i t h no satell ite fixes and to ± 1 8 0 m w i t h satell ite fixes (Grant , 1973). W i t h i n a l ine, the relat ive error between a i rgun shots is est imated to be ± 1 5 m, whi le between lines, the relat ive error is ± 3 0 m (Grant , 1973). Ba thymet r i c records were obta ined concurrent ly in analogue form us ing - the ship 's 12 k H z depth sounder. 2.2 Instruments S ix Un iver s i ty of B r i t i s h C o l u m b i a O B S w i t h t imed release and four A t l a n t i c Geoscience Cent re O B S w i t h acoustic release were used as receivers (Hefner and Ba r re t t . 1979). These axe cont inuous ly-recording analogue systems w i t h the data being recorded onto cassette tapes at a slow speed (~0.008 inches per second (ips)). Each tape records four Chapter 2. DATA ACQUISITION AND PROCESSING 31 channels: signals f rom a 4.5 H z vert ica l seismometer, a 4.5 H z hor i zonta l seismometer, a hydrophone, and a crystal osc i l lator in terna l clock wh ich t ransmits a t ime code r i d i ng on a 10 Hz ca l ib ra t ion signal. A 32 L a i rgun manufactured by B O L T Associates was used as the source for the closely-spaced shots. M a x i m u m pressure (13.79 M P a or 2000 psi) was ma inta ined at a l l t imes to control the repeatabi l i ty of the source. T h e airgun was towed at a nom ina l depth of 30 m. T h e airgun shots were t imed by a c lock-control led f i r ing un i t wh ich was per iod ica l l y rated against a master clock and W W V B . 2.3 Data Processing Since only in - l ine a i rgun data w i l l be interpreted i n this thesis, on l y processing done to those da ta w i l l be considered. 2.3.1 Initial Data Reduction Due to the slow speed at which the O B S tapes were recorded, a series of re-recordings was required to prepare the data for analogue-to-d ig i ta l conversion. T h e slow speed cassette tapes f rom the O B S were t ranscr ibed onto 1/4" F M tape using a T E A C M o d e l R -70A tape transport at a speed of 1 7/8 ips, ~ 2 4 0 times faster than the rate at which the d a t a were recorded. If this tape were played in to the dig it izer at this speed, however, the transmiss ion rate of the data wou ld be too fast for the d ig i t i z ing sj ' stem; more recording steps are needed to slow down the t ransmiss ion rate. To accompl i sh this, these da ta were re-recorded onto 1/4" F M tape using a R A C A L tape recorder at 30 ips, thereby obta in ing a speed-up factor of 15 t imes for the or ig ina l O B S data. T h e F M tapes were then played into a PDP-11-ba.sed d ig i t i z ing system using an H P recorder at a tape speed of 3 3/4 ips, for a f ina l speed-up factor of 30. Because of the large Chapter 2. DATA ACQUISITION AND PROCESSING 32 amount of da ta recorded, au tomated d ig i t i z ing software was developed for the d ig i t i z ing system ( G M A C o m p u t e r Technology, L t d . , 1986). To account for tape speed var iat ions, the dig it izer was synchronized to the 10 Hz t ime code to control the d ig i t i za t ion interva l , i.e., each 0.1 s i n terva l was d i v i ded into equal increments on the basis of t ime, not tape length. Th i s technique l imits the t im i ng errors to less than one sample po int (8 ms). T h e da ta were d ig i t i zed at 120 samples per second and recorded onto 1600 B P I tapes. For the single a i rgun data, 25 seconds of da ta were d ig i t ized. T h e O B S da ta were then demul t ip lexed on the U B C ma in f rame computer ( A m d a h l 5860) and reformatted. 2.3.2 Shot and Receiver Positioning In order to determine the shot and i n i t i a l receiver postions, the nav igat iona l da ta had to be processed and corrected for errors. T h e receiver ( O B S ) locations were later altered to ma tch the water waves w i t h i n 2 k m of the O B S (Section 3.1.3). R a w B I O N A V Lo r an -C files logged on the mic rocomputer were copied f rom diskette to the ma in f rame computer and then edited to check for t ime j umps and to interpo late missed points. T h e B I O N A V da ta were smoothed us ing a 3-point moving-average f i l ter to remove sharp j umps i n pos i t ion. Such f i l ter ing caused the movement of shot posit ions by up to 50 m but most posit ions were moved less t han 20 m. T h e Lo r an -C da ta were then adjusted us ing satel l ite fixes since L o r an -C contains range biases wh ich are ma i n l y due to errors i n synchron izat ion between the t ransmit ter and receiver a tomic clocks (Grant , 1973). A l inear ly -vary ing pos i t ion correction w i t h t ime was used to correct this dr i ft of the L o r an -C between satell ite fixes. Because of the large absolute error of L o r an -C pos it ions, this correct ion was up to 250 m, but the correction is s imp ly a stat ic shift of al l L o r a n - C p o s i t i o n s — t h e relative error between a i rgun shots is s t i l l ± 1 5 m. T h e la t i tude and long i tude determined were corrected for the distance f rom the ship 's stern to the L o r a n - C antenna and for the distance f rom the airgun to the stern of the ship. Th i s Chapter 2. DATA ACQUISITION AND PROCESSING 33 antenna-a i rgun offset was est imated to be 75 m. 2.3.3 Bathymetry and Instrument Depths Wate r depths were determined f rom the paper analogue 12 k H z depth sounder records obta ined dur ing the cruise. These depths are p lo t ted f rom the travel t imes assuming a constant water speed of 1500 m/s. T h e or ig ina l records were photocop ied and then d ig i t i zed on a Cybe rg raph d ig i t i z ing table at 1 minute intervals. W W V B times were assigned to d ig i t i zed depths and lat i tude- long i tude points assigned us ing the nav igat ion data. T h e ba thymet r y depths were reconverted us ing a constant water speed of 1485 m/s for use i n the d a t a interpretat ion, where 1485 m/s is the average veloc ity for a sounding obta ined from a water co lumn i n the area during^ the mon th of October (Tabata and Peart , 1985). Errors in the bathymet ry are due to measurement f rom the bathymet r i c records and the d ig i t i za t ion interva l . Some bathymet r i c profiles were of poor qua l i ty wh ich made d ig i t i za t ion di f f icult ; the m a x i m u m error i n the depths caused by this is p robab ly ± 1 0 m, wh i ch when converted to t ime, is a m a x i m u m of ± 7 ms. In regions of rap id bathymet r i c changes, the d ig i t i za t ion interva l of 1 minute also creates error; this error is <10 m. T h e O B S depths were determined dur ing deplo3'ment us ing the depth sounder. T h e stat ionar i ty of the ship over the O B S deployment site enabled the determinat ion of these depths to ± 1 0 m. T h e airgun dep th was var iable because the conf igurat ion used to pu l l the a i rgun beh ind the ship changed. A constant depth of 30 m was chosen w i t h the estimate of the depth error ± 1 0 m. Th i s nom ina l depth of 30 m wras used for al l lines and subtracted f rom the ba thymet r i c depths so that the airgun depth, not sea level, was used as the zero da tum. Chapter 2. DATA ACQUISITION AND PROCESSING 34 OBS Head Skew (ms) 2 60 ± 3 5 89 ± 4 8 6 ± 2 9 2 ± 1 12 - 3 1 ± 2 14 18 ± 1 Table 2.1: Tape-head-skew values for the O B S . 2.3.4 Timing Corrections T i m i n g corrections were made for O B S clock dr i f t , tape skew, and dr i ft of the a i rgun shot clock. T h e O B S clocks were rated against W W V B using a master clock and then deployed; on recover}7, the drift was determined by again rat ing the O B S clocks against the master clock. T h e rate of dr i ft was assumed l inear; such an as sumpt ion is jus t i f ied because of the constant temperature env i ronment on the sea bo t tom. Because of imper -fect a l ignment of the recording heads i n the O B S tape recorders, the d a t a channel was delayed relat ive to the t ime channel. Th i s head skew was determined f rom large-scale plots of the raw digit ized da ta on an electrostat ic p lotter by us ing a ca l ib rat ion s ignal recorded on a l l four channels. Th i s correction was as h igh as 89 ms but was usual ly much less than this (see Table 2.1). T h e clock that tr iggered the a i rgun at one or two minute intervals was also rated against W W V B v i a the master clock; the a i rgun clock t ime code was recorded on F M tape along w i t h W W V B and later rated to determine the drift. T h e or ig in t ime of the airgun clock has an uncerta inty of ± 3 ms after the dr i ft correction. For one of the " doub le " a irgun l ines, L i ne D5 , the shot clock had large nonl inear t im ing shifts as wel l as poor qual i ty concurrent ly-recorded W W V B . T h e dr i f t for this l ine was determined by comparison w i t h the W W V B s ignal us ing a s tat i s t ica l app roach—seve ra l seconds of W W V B were stacked and cross-correlated w i t h a step funct ion to obta in t imes Chapter 2. DATA ACQUISITION AND PROCESSING 35 at 15 s intervals. T imes were then interpo lated between these 15 s intervals. A s a result, the t im i ng errors for L i ne D5 are larger ( ± 1 0 ms) than for other lines (J . A . Hole, pers. comm, 1988). Chapter 3 DATA ANALYSIS O n l y in - l ine a i rgun da ta recorded on the inner array w i l l be considered i n this study. T h e explos ion data , further offset a i rgun data , and fan shot da ta are being considered i n separate studies. F i f teen sections recorded on ten lines are ana lyzed in this thesis. Most of these lines were recorded on four O B S — O B S 2, 5, 8, and 9 (F igure 2.2). 3.1 Initial Observations and Processing 3.1.1 Data Characteristics and Quality T h e vert ica l component of the a i rgun da ta recorded on O B S 5 and 9 is generalh r of high qua l i ty (e.g., F i gure 3.2). F i r s t arrivals are v is ib le for the complete length of the lines and have a h igh signal-to-noise rat io. Ve r t i c a l component da ta recorded on O B S 2 and 8 are much poorer; first arr ivals are not easily v is ible along the section and are contaminated w i t h h igh amp l i t ude noise (e.g., F i gure 3.4). A n unexp la ined observation is that the high qua l i ty da ta were recorded on inst ruments on the J u a n de Fuca ( J D F ) P l a te , wh i le poor qua l i ty da ta were recorded on inst ruments located on the Pac i f i c P late. For tunateh r , i t was possible to use the hj ' d rophone data to extend the distances to wh ich arrivals could be picked (e.g., F i gure 3.9). T h e hor i zonta l component of the da ta was not used i n mode l l i ng , a l though it was examined for a few lines, pa r t i cu la r l y w i t h respect to secondary arrivals. F r o m power spectra, the frequency content of the vert ica l component data ranges 36 Chapter 3. DATA ANALYSIS 37 f rom 4-15 H z (F igure 3.1) w i t h a strong peak at ~ 1 0 Hz. Mos t of the noise is of higher frequency than the data. For seismic velocities of 3 to 6 km/s, the wavelength at 10 H z is 300 to 600 m. If the resolut ion of a seismic wave is assumed to be 1/4 of its wavelength, the m i n i m u m resolut ion for this s tudy is 75 to 150 m. To enable easier discussion of the interpretat ion of the da ta presented i n Chap te r 4, one seismic section w i l l be descr ibed in deta i l to int roduce the terminology that w i l l be used and the t ravel - t ime and amp l i t ude characterist ics of the da ta (F igure 3.2). Unless otherwise noted, a l l da ta have been f i ltered f r om 0-15 H z us ing an 8-pole zero-phase Bu t t e rwo r th bandpass f i lter and are p lot ted w i t h a reduc ing veloc ity of 6.0 km/s. T h e section shown i n F i gu re 3.2 has not been f i ltered to emphasize the h igh qua l i ty of this data. To show the relative amp l i t ude var iat ions, the trace ampl i tudes are mu l t i p l i ed by a factor p ropor t i ona l to distance to present t rue relative ampl i tudes. Four different t rave l - t ime branches w i l l be discussed. T h e start of each branch is i nd i ca ted by a t r iangle wh i ch is located at the first break for the marked trace. T h e travel t ime of a l l arrivals is very topography-dependent, w i t h the first arr ivals advanced in regions of shallower ba thymet ry and vice-versa. T h e rap id var iat ions i n ba th j 'met ry also cause fluctuations i n the observed ampl i tudes due to focussing and defocussing effects. T h e strong water wave occurs as the first a r r i va l for the first 2 k m on either side of the O B S . Its relat ive amp l i tude is low due to the distance scal ing factor. T h e h igh amp l i t ude of this a r r iva l overloaded the seismograph system and causes r ing ing after the first ar r iva l ; however, these h igh amp l i tude arrivals are not c l ipped because the nonl inear amp l i t ude response of the seismograph system causes a soft c l ipp ing. In this figure, the complete water-wave ar r i va l has been shown so that the r ing ing may be seen. O n a l l other data plots, the water-wave arr ival has been t runcated at later times since it provides no useful i n fo rmat ion . Chapter 3. DATA ANALYSIS 38 F i gu re 3.1: (a.) Pe r i odog ram of 2 s of noise, (b) Pe r iodogram of 2 s of s ignal. T h e trace used to obta in these figure is f rom F igure 3.2 and is marked by arrows. DISTPNCE (KM) Figure 3.2: True relative amplitude section for Line 9-OBS 6. The data are of excellent quality and have not been filtered. Branches described in the text have been labelled on the section. The triangles indicate the start of each branch and are located at the first break of the indicated traces. The prominent travel-time undulations are primarily caused by bathymetric variations. The arrows mark the trace whose power spectra is shown in Figure 3.1. C O C O Chapter 3. DATA ANALYSIS 40 Emerg ing tangent ia l ly f rom the water wave is a low-velocity crusta l a r r iva l wh ich w i l l be referred to as B ranch 1. T h i s low veloc i ty phase (~2.5 km/s) exists as a first arr iva l for approx imate ly 1 km; it is observable as a later arr iva l on some sections beyond this distance wh i l e on other sections it d isappears w i t h i n h igh -ampl i tude reverberations. F i gu re 3.3 is a po r t i on of F i gu re 3.2 that shows this b ranch emerging f rom the water wave. T h e distance at which B ranch 1 is first observed is dependent on the depth of the i n s t r u m e n t — w h e n the inst rument is shallower, i t takes the arrivals a shorter distance to emerge f rom the water wave. Th i s exper iment was successful i n ob ta in ing arrivals f rom this layer because of the relat ively shal low depth of Endeavour R idge and the smal l shot spac ing used i n the acquis i t ion of the a i rgun data . A f te r B r anch 1, there is a rap id breakover to a h igh -ampl i tude ar r iva l w i t h a velocity of ~ 5 km/s (F igure 3.3). Th i s ve loc i ty est imate was determined f rom other sections where variat ions i n bathymet ry were less severe. Th i s phase, B r anch 2, extends out to ~ 6 k m (F igure 3.2). E x t end i n g out to 11 k m is a branch w i t h a velocity of ~ 6 km/s. T h e ampl i tudes of B r anch 3 f luctuate bo th w i th i n a section and between sections; the general amp l i tude t rend is sometimes stronger than that of B r a n c h 2, but it is more often weaker. O n the longest-offset sections there is a fourth branch wh ich extends to the l im i t of the distances achieved for this survey (26 km) ; i t has a velocity of ~6 .7 km/s. T h e lower amp l i t ude of the B r anch 4 arrivals is p robab ly due to a decrease i n velocity gradient, and causes dif f iculty i n p ick ing the first breaks w i t h i n the ambient noise. Later arrivals are vis ible on man}- sections, most of which appear to be P -wave mu l -tiples since they m i m i c the trend of the p r imary arrivals. These arrivals sometimes have higher ampl i tudes than the p r ima ry arrivals, a phenomenon observed i n other mar ine refract ion data sets (Baggeroer et al, 1986). Coherent noise on some sections is due to s imultaneous use of a smal ler airgun to DISTANCE (KM) Figure 3.3: Portion of Figure 3.2 showing the low-velocity arrival from the uppermost oceanic crust. The position of the water-wave arrival is shown by the dashed curve and the first breaks for the Branch 1 arrivals are indicated by arrows. Arrivals from Branch 2 with an apparent velocity of ~4.8 km/s are also marked by a dashed line. Chapter 3. DATA ANALYSIS 42 ob ta i n vert ica l inc idence reflection da ta (F igure 3.4). Its use was te rminated because of poor sea condit ions and through concern that i t was causing noise on the O B S . A noisy trace f rom F igure 3.4 is presented i n F igure 3.5 to show the effect of zero-phase bandpass f i l ter ing f rom 0-15 Hz . Some sections are contaminated by high frequency "sausage" waves w i t h a frequency of 20 H z (F igure 3.6). These are believed to be whale sounds because at least one whale was sighted i n the area whi le the experiment was being done and signals Like these have been recorded by others ( G . M . Pu rdy , pers. comm, 1988; Jacobson et ai, 1987); f i l ter ing f r om 0-15 H z removes most of these signals. 3.1.2 T r a v e l T i m e P i c k i n g Choos i ng the pos i t ion of the first break can be a diff icult procedure i f the da ta qua l i ty is poor or i f the phase of the waveforms varies a long the branch. T h e repeatable a i rgun source and the high densit} ' of arr ivals used i n the acquis it ion of these da ta made p ick ing of the first breaks relat ively stra ightforward. F r o m v iewing the best qua l i ty data, the first break was chosen at the pos i t ion shown i n F i gu re 3.7. T h e general cont inu i ty of the waveform on a pa r t i cu la r section made it poss ible to use a transparent template to p ick arrivals. A " c l e an " unf i l tered arr iva l t yp i ca l of most of the first arrivals on the section was chosen and this waveform was fit by eye to a l l the arrivals us ing the template. Th i s kept the p ick ing consistent and overcame smal l changes i n waveshape and the loss of the first break i n the precursory noise at greater distances. Trave l t imes were picked on unf i l tered common -max imum-amp l i t ude vert ica l -component da ta p lot ted at a reduc ing velocity of 6 km/s; us ing a l ight table, this section was p laced over a 0-15 H z bandpass-f i l tered version of the same data. T h e f i ltered sec-t i on was p lot ted i n variable area format (e.g., F igure 3.8) wh i ch made it easier to see the cont inu i ty of arrivals. These were large-scale pen plots on l i ned paper w i t h a t ime Chapter 3. DATA ANALYSIS 43 - L I N E 5 - O B S 8 in whale sound small airgun DISTANCE (KM) Figure 3.4: Common maximum amplitude plots for Line 5 -OBS 8, a section with poor data quality, (a) Unfiltered: These data are contaminated by noise from a small airgun and from whale sounds. Traces marked by arrows are shown in Figures 3.5 and 3.6. The large arrowhead indicates water-wave arrivals from a small airgun. (b) Filtered: Filtering these data from 0-15 Hz improves the data quality by removing high frequency noise such as whale sounds (Figure 3.5). Chapter 3. DATA ANALYSIS 44 Figure 3.5: Noise-contaminated arrival (a.) before and (b) after filtering from 0-15 Hz. (c) Power spectra of arrival before and after (dashed line) filtering. Trace is marked in Figure 3.4. Chapter 3. DATA ANALYSIS T I M E ( S ) o Figure 3.6: (a) Whale sound and (b) power spectrum. Arrival is from Figure Chapter 3. DATA ANALYSIS 46 2.0 2.1 2.2 2.3 2.4 TIME (S) F igure 3.7: Po s i t i on of first break ind icated by the large arrow. Th i s a r r i va l is i nd i ca ted i n F i gure 3.2. T h e source wavelet used later in the synthet ic sections extends between the arrows. scale of 0.25 s/in. and a distance scale of 1 k m / i n . For O B S 2 and 8, arrivals at greater distances were clearer on the hydrophone; f i l tered sections of the hydrophone channel were used as an underlay. A sma l l t ime shift exists between the hydrophone and vert ica l seismometer da ta because of the different sensors used, i.e., pressure and displacement, respectively. On l y t imes f rom vert ica l da ta were used; readings f rom hydrophone chan-nels were adjusted for the t ime shift. For a l l reversed lines, t ravel times were picked to ensure reciprocity at the OBS- separat ion offset. F i gure 3.8 shows the two different formats used for a h igh-qua l i ty da ta set and the resolut ion poss ible in the picks, wh i le F i gure 3.9 shows the three different plot formats used to pick a poor qua l i ty data set. W h e n travel - t ime p icks were made, on ly the under lay was p lotted in var iable area format; a l l sections i n Figures ' 3.8 and 3.9 have been p lotted i n var iable area format so that the cont inu i ty of arr ivals may be better observed. It must be emphas ized that the data shown in Chap te r 4 are not p lot ted in the same format that Chapter 3. DATA ANALYSIS 47 D I S T R N C E (KID F igu re 3.8: P l o t formats used to pick good qua l i t y data. B o t h sections are com-mon -max imum-amp l i t ude plots of the vert ica l component data, (a) Unf i l te red da ta for a po r t i on of L i ne 9 - O B S 6 (F igure 3.2). Locat ions of first a r r iva l picks are shown as dots, (b) F i l t e red version (0-15 Hz) of (a). T h e la tera l cont inu i ty of arrivals is clear for this section. Chapter 3. DATA ANALYSIS 48 DISTANCE F igu re 3.9: P l o t formats used to p ick poor qua l i ty data, (a) Unf i l tered common m a x i m u m amp l i t ude da ta f rom the vert ical seismometer for a por t ion of L i n e 5 - O B S 8 (F igure 3.4). Locat ions of t ravel - t ime picks are shown as dots, (b) F i l te red version (0-15 Hz ) of (a), (c) F i l te red version (0-15 Hz) of the hydrophone data, recorded for the same line. A t these distances, the first arrivals are clearer than those on (b). T h e pair of lines on each plot j o i n points of constant phase f rom the hydrophone first arr iva l . Chapter 3. DATA ANALYSIS 49 was used for p ick ing the travel t imes. In many cases, the travel -t ime picks were made after cons iderat ion of many different plot formats and reducing velocities and considerat ion of the pos i t ion of later phases. A l l first arr ivals and some later arrivals were picked on a l l sections; picks were made to the nearest 5 ms, the resolut ion of the t ime lines on the p lo t ted section. A p ick ing error of ± 1 0 ms is est imated for the best qua l i ty da ta w i t h the average error ± 1 5 ms and the largest error ± 2 5 ms for the noisiest da ta at the farther offsets. T h e error for the later arrivals is at least ± 2 5 ms for a l l sections since the p ick ing of these arrivals is more uncerta in . These errors are va l id assuming the correct cycle of the wavelet is p icked; i f this a s sumpt ion is incorrect, the error can be up to ± 1 0 0 ms. T h e water-wave arrivals were picked w i t h a different template, and these travel times were later used to constra in the O B S pos i t ion. T h e relat ively h igh spat ia l density of the d a t a suggested that stacking and f-k filtering may be appropr iate for sections that were contaminated by noise. Travel - t ime f luctua-tions due to changes i n topography as well as uneven shot spacing caused problems w i t h these techniques, so they were not considered further. 3.1.3 OBS Positioning From Water-Wave Arrivals T h e travel t imes of the water-wave arrivals about the O B S pos i t ion were often not sym-met r i ca l as expected, so the O B S was repos it ioned to correct this. Us ing the O B S deployment depths and only a l lowing movement para l le l to the l ine, water-wave arrivals f rom w i t h i n ~ 2 k m of the O B S were fit to the picked travel t imes us ing a least-squares fo rmulat ion (F igure 3.10). D i s tance shifts d i d not completely account for the mi smatch so t ime shifts were also made (Table 3.1). These t ime shifts account for off-l ine pos i t ion-ing of the O B S , airgun depth changes, and res idual errors in bathymetr} ' , clock drift and tape-head skew. T h e consistency among t i m i n g corrections for different lines recorded Chapter 3. DATA ANALYSIS 50 F i gu re 3.10: Repos i t ion ing an O B S f rom match ing the water-wave arrivals, (a) Theoret i -cal water-wave arrivals (curve) compared to those picked f rom the observed d a t a (boxes), (b) Theoret i ca l water-wave arr ivals compared to observed picks after the O B S has been moved para l le l to the l ine and a t ime shift made. T h e R M S error between the observed and ca lcu lated water-wave times has decreased f rom 61.8 to 8.7 ms. Chapter 3. DATA ANALYSIS 51 OBS Line Time Shift Distance Shift R M S Error (ms) (ms) (m) Before After 5 -10 - 2 14.2 10.5 2 6 - 8 73 25.5 7.6 9 -17 27 19.2 6.9 5 8 -42 7 44.3 12.3 9 -38 - 4 38.1 12.1 3 -48 38 46.8 8.4 8 5 -58 -40 57.4 6.8 8 -59 53 61.8 8.7 D5 -72 -147 41.7 7.7 4 -48 -34 49.7 7.6 9 6 -41 11 42.9 8.6 7 -46 -124 58.0 8.1 12 2 - 4 232 48.1 9.3 14 2 0 -237 58.0 9.8 Tab le 3.1: T i m e and distance shifts made to travel times and O B S posit ions f rom water wave match ing. R M S error refers to difference between picked travel times and calculated travel times before and after O B S repos i t ion ing. Chapter 3. DATA ANALYSIS 52 on the same O B S shows that these t ime shifts are va l id . T h e ba thymet ry was then rep lotted w i t h reference to the new O B S pos i t ion, shot-receiver ranges were changed, and t ime shifts made to the travel-t ime picks. A f te r reposi-t i on ing the O B S separately for each l ine, the adjusted O B S posit ions were not consistent f rom l ine to l ine, but most are w i t h i n the L o r a n - C absolute nav igat ion error of 180 m. T h e O B S posit ions are not consistent because the O B S were not al lowed to move off-line i n the water-wave f i t t ing procedure. T h e R M S travel -t ime error obta ined after f i t t i ng the water-wave arrivals is a measure of the water-wave p ick ing error. It ranges f rom 5 ms (the resolut ion of the travel - t ime picks) to 12 ms. These errors are less than those for the refracted arr ivals because of the ease i n choosing a first break on the impu l s i ve near-offset water-wave arrivals (e.g., F igu re 3.3). 3.2 Interpretation Methods 3.2.1 Initial Interpretation A n i n i t i a l one-dimensional mode l was obta ined by stra ight- l ine fits to the travel times us ing formulae for the source and receiver at different depths (Dehler, 1986). These s imple models were extremely gross i n structure because rap id t ravel - t ime f luctuat ions caused by the topograph} 7 made f i t t i ng a straight l ine diff icult. A s an independent check, wavefield cont inuat ion (P ike, 1986; C l a y t on and M c M e c h a n , 1981) was also t r ied, but poor contro l on the upper crustal velocities and the large topographic corrections required for a one-d imens ional interpretat ion made this me thod unsuccessful. A two-d imens iona l forward mode l l ing scheme was then used to analyze the in- l ine arrivals, w i t h a number of different s tart ing models examined (Cud rak et al, 1987). T h e start ing mode l used for the models shown i n this thesis is descr ibed i n Sect ion 3.2.3. Chapter 3. DATA ANALYSIS 53 3.2.2 Modelling Algorithm T h e two-d imens iona l ray t rac ing a l go r i thm described by Zelt and E l l i s (1988a), based on a symptot i c ray theory ( A R T ) (Cerveny et ai, 1977), was used to forward mode l the refract ion data. M o r e accurate methods, such as f inite-difference a lgor i thms, are avai l -able for forward-model l ing i n two- and three-dimensions, but are not computa t i ona l l y p rac t i ca l for this study. T h e Zelt and E l l i s (1988a) a lgor i thm was chosen because of its p ract i ca l i t y and efficiency wh ich a l lowed for rap id forward mode l l i ng of travel t imes and ampl i tudes (e.g., Zelt and E l l i s , 1988b). A n impo r tan t aspect of this a lgor i thm is its au tomat i c determinat ion of r a j ' groups. For the Zelt and E l l i s (1988a) a lgor i thm, the mode l is parameter ized i n terms of quas i -hor izonta l layers, each layer separated by a boundary consist ing of stra ight- l ine segments (F igure 3.11). In this parameter izat ion, ba thymet ry can be i npu t as a single interface w i t h the seawater as the first layer: topograph ic corrections are not required. W i t h i n each layer, the P -wave veloc i ty s t ructure is defined by specify ing a single upper and lower layer veloc ity for each stra ight- l ine segment of the upper layer boundary. T h e a lgo r i thm then div ides these layers into blocks based on changes i n slope of the upper and lower boundaries. Since the thickness of layers can vary, the velocity gradient can change laterally. P inch-outs or i so lated bodies can also be inc luded. In add i t ion to the standard reflected and refracted rays (F igure 3.11 and 3.12), mult ip ly - ref lected and/or converted rays may also be model led. A m p l i t u d e calculat ions are based on zero-order A R T wh i ch models wave phenomena that can be predicted by geometr ica l opt ics, i.e., d iffractions cannot be model led. T h e Zelt and E l l i s (1988a) a l go r i thm can also calculate the ampl i tudes of common-receiver profiles i n the correct d i rect ion; i.e., most ray t rac ing programs assume that the ca lcu lat ion of ampl i tudes is rec iproca l so that the O B S is mode l led as the shot and the airgun shots as the receivers. Chapter 3. DATA ANALYSIS 54 but th is as sumpt ion is not va l i d , especial ly for locat ions close to the O B S ( C A . Zelt, pers. comm., 1988). 3.2.3 Starting Velocity Model T h e start ing velocity model used (F igure 3.11) consists of five l a y e r s — t h e water and four crusta l l a y e r s — a n d is a sl ight mod i f i cat ion of the start ing mode l used by W h i t e and Clowes (1988). Other s tar t ing models were used (e.g., C u d r a k et ai, 1987) and were relat ively successful, but the W h i t e and Clowes (1988) mode l was consistent w i t h the invers ion procedure (Wh i te , 1988) and thus independent of interpreter biases. F i g -ures 3.11 and 3.12 are two-d imens iona l versions of the start ing mode l wh i le F i gure 3.13 is a ve loc i ty -depth funct ion that better shows the changes i n gradient. A velocity of 2.5 km/s and a veloc ity gradient of ~0 .5 s _ 1 was used for the f irst layer. W i t h this gradient, the velocity at the b o t t o m of the 0.45 km- th i ck layer is 2.7 km/s. Since a constant thickness was used i n the s tar t ing model , the b o t t o m of this layer m im icked the seafloor topography. Th i s layer w i l l be referred to as Layer A . T h e second layer, Layer B, had a velocity of 4.80 km/s at the top of the layer and a velocity of 5.30 km/s at the b o t t o m of the layer. A gradient of ~0.75 s _ 1 was created by a layer thickness of 0.65 k m . T h e b o t t o m of this layer also fol lowed the seafloor topography. T h e velocities of Layer C were 5.80 and 6.30 km/s for the top and the b o t t o m of the layer, respectively. The bo t t om of this layer was chosen to he at a depth of 4.35 km/s; the layer was of var iable thickness (~1.0 km) and gradient (~0.5 s _ 1 ) . T h e b o t t o m layer, Layer D, extended f rom 4.35 k m to 7 k m , the depth l im i t of a l l the models. W i t h a velocity of 6.5 km/s at the top of this layer and 7.3 km/s at the bo t t om, the gradient is ~0.30 s - 1 . -15 -13 -11 DISTANCE (KM) -5 -3 -1 1 " i n — CM ^ CD -Q O I C M " Branch 1 Branch 4 Branch 3 ooooo 0 oooooo0000 0 =oo 0 oooo0 0 0 o 0 o o o 7 o o o o o ~ e^-- X&^>> 2 ° o o o o o o g * . B r a n c h _ 2 / £ . « * Branch 3 i — i — r i i l i — r 7 Branch 4 ~~i i i i r 15 -13 -11 i~~i r~ -9 -7 i r -5 n Figure 3.11: (a) Starting model for Line 10-OBS 5 (Figure 4.1) showing the layer bound-aries and the model parameterization into blocks. The Layer A-B and Layer B-C in-terfaces follow the seafloor bathymetry. The four different refracted arrivals used in the modelling procedure have been traced on this model, (b) Comparison of calculated travel times (lines) from refracted arrivals for the starting model to the observed travel-time picks (symbols). This comparison shows that the starting velocities are correct because the slope of the calculated branches agrees with the observed travel times, but Layers A and B are too thick because arrivals from Branches 2, 3, and 4 are delayed. The final model for this line is shown in Figure 4.11. Figure 3.12: (a) Reflections from crustal interfaces for Line 10-OBS 5 starting model. This is the only figure that exclusively shows the reflected arrivals, (b) Comparison of calculated reflected arrivals (lines) to observed travel times (symbols). Reflection arrivals were not used in the travel-time modelling; this figure shows the relationship of the reflected arrivals to the travel times so that the reflected arrivals may be compared to Figure 3.11. The cusp at 0 km is caused by plotting the calculated data with a reducing velocity. Chapter 3. DATA ANALYSIS 57 F i gure 3.13: Ve loc i t y -depth funct ion of s tart ing model . T h e velocities and depths of layer boundar ies are label led. Chapter 3. DATA ANALYSIS 58 3.2.4 Modelling Procedure Two-d imens i ona l mode l l ing of refract ion da ta is done us ing a tr ia l -and-error approach. A s tar t ing mode l is chosen, rays are traced through this mode l and their t ravel times compared to those observed on the da ta sections. Mod i f i ca t ions are then made to the mode l to improve the fit between the model led and observed travel times. Trave l times f rom progressively deeper layers are matched so that the upper crusta l s t ructure is first determined, and i f a l ine is reversed, bo th sets of arrivals are matched s imultaneously. Synthet ic seismograms are also ca lcu lated and compared to the observed da ta to better est imate the vert ica l velocity gradients i n the model . On l y the general amp l i t ude char-acterist ics are matched because rap id A c t u a t i o n s are due to small-scale heterogeneities not incorporated i n the ra.y theory- model l ing . T h e synthet ic and observed ampl i tudes are on ly compared by eye so that t ravel -t ime f i t t i ng provides the p r imary source of mode l constraints. Intersection points w i t h other models are also used to constrain the velocities and layer thicknesses of the mode l . S ince layer thicknesses, layer velocities, and velocity gradients a l l have an effect on t ravel t imes, adjustments to different parameters can be used to match the same travel-t ime data. For the models shown in this thesis, layer velocities were not adjusted unless necessary; changes i n layer thicknesses were used to match the travel -t ime data. Layer velocit ies were only changed when reversed da ta or constraints f rom intersect ing lines wou l d not al low the adjustment of the layer thickness, or when layer thickness changes were required that were judged unreasonable. Th i s method thus presumes that the dominant effects on travel t imes are s t ructura l and that latera l velocity var iat ions are a secondary effect. Ve loc i ty gradients were adjusted when synthet ic seismogram mode l l ing was done since the ampl i tudes of refracted arr ivals are par t i cu la r l y dependent on velocity gradients. In order that the t ravel - t ime match wou ld be altered as l i t t le as possible, the Chapter 3. DATA ANALYSIS 59 ve loc i ty gradient was adjusted so that the average veloc ity i n a b lock remained constant; the veloc ity at the top of the block was increased and the velocity at the bo t tom of the block decreased to decrease the velocity gradient and vice-versa for an increase i n ve loc i ty gradient. Th i s mode l l i ng procedure produces crusta l models that have sma l l ve loc i ty differences but larger thickness var iat ions when compared to the s tart ing model . To calculate synthet ic seismograms, some phys ica l constants must be assumed. Po is -son's ratios of 0.50 and 0.30 were used for the water co lumn and first layer (Layer A ) , respectively, and 0.25 was used for the other c rus ta l layers ( Spud ich and Orcu t t , 1980b). T h e veloc ity-dens i ty re lat ionsh ip used was obta ined f rom da ta presented i n L u d w i g et al. (1970). To compensate for anelast ic at tenuat ion (Zelt and E l l i s , 1988a), the ampl i tudes of the synthet ic seismograms were calculated assuming a Q of 25 and 200 for Layer A and B, respectively (Jacobson and Lewis, 1988), and 450 for Layers C and D (Spud ich and O rcu t t , 1980b). These O values were not determined f rom the amp l i tude mode l l i ng of the da ta set but were incorporated in the amp l i t ude calculat ions to ob ta in more rel iable estimates of vert ica l velocit} ' gradients (Zelt and E l l i s , 1988b). T h e p r i nc ipa l ray groups used i n mode l l i ng were refracted rays through each of the four layers (F igure 3.11). Ref lected rays f rom the three interfaces were traced when amp l i t ude mode l l ing was done (F igure 3.12). T h e t ravel times of precr i t ica l reflections (F igure 3.12b) at the O B S pos i t ion form a cusp because of the reducing velocity used; the earliest arrivals at short offsets are reflections f rom the first i n t e r f ace—the direct water wave was not model led. For some lines, mul t ip les and reflected-refractions were model led. A l l l ines were model led so that the ca lcu lated and observed travel t imes matched to w i t h i n ± 1 5 ms. Th i s average error includes the p i ck i ng error, ba thymet ry error, and the samp l i ng error and is the error est imate for an ar r iva l near the m idd le of a l ine; near the O B S , the error is less, wh i le at the furthest distances, the error is greater. Other errors Chapter 3. DATA ANALYSIS 60 such as those due to nav igat ion are consistent w i t h i n a l ine and do not have an effect on the t ravel t imes. T h e travel t imes of later phases were less closely matched and were on ly used to make m ino r adjustments to the models. T h e most common mu l t ip le was model led as a P - w a v e peg-leg w i t h i n Layer A (RI, F i gure 4.6a). For later arrivals, it was diff icult to determine the ray pa th an ar r iva l had taken, since i n some cases more than one phase could arr ive at the same t ime by different combinat ions of reflections and shear-wave conversions. N o amp l i t ude match ing was done to the secondary phases for this reason, a l though ampl i tudes were generated. Observed secondary arrivals w i t h higher ampl i tudes t han the p r ima ry arr ivals were not successfully repl icated. A more accurate method w i t h respect to f in ite frequency data, such as ref lect iv ity in one d imens ion or finite-difference i n two dimensions, wou ld be more appropr iate to the s tudy of bo th the travel t imes and ampl i tudes of these later phases. A m p l i t u d e var iat ions w i t h i n the observed da ta are often caused by l a tera l changes i n s t ructure or scatter ing by topography, effects that cannot be model led accurately by A R T . Other problems i n the ca lcu lat ion of A R T ampl i tudes i nc lude the amp l i tude of the shallowest tu rn ing ray in a layer and sens it iv i ty to sharp discontinuit ies i n the model . A R T ampl i tudes are not va l id i n the interference zone of the reflected and head waves near the c r i t i ca l po int. Th i s causes an ar t i f i c ia l amp l i tude low for the shallowest refracted rays f rom a layer and creates a pat tern of successive highs and lows on a synthet ic section, the number of lows equal to the number of layers in the velocity model . Inc lud ing the head waves i n the mode l l i ng wou ld increase the ampl i tudes near the c r i t i ca l point, but the amount cannot be determined us ing A R T . Since A R T is an inf in i te frequency app rox imat i on , d iscont inuit ies i n velocity (especially at vert ica l block boundar ies) and veloc ity gradients cause art i f ic ia l highs and lows in the amp l i tude data whereas the finite frequency observed da ta smooth out these discontinuit ies. To par t ia l l y overcome this Chapter 3. DATA ANALYSIS 61 problem, both the layer boundaries and the amplitude-distance curves are smoothed to average out these amplitude artifacts (Zelt and Ellis, 1988a). Since the amplitudes of the data were only approximately matched, these limitations of ART are not important and a more accurate technique is not required. When modelling, velocity variations were made in increments of 0.05 km/s; this amount is an estimate of the uncertainty of the velocities in the models. Thickness changes were dependent on the interface. For the shallowest layer, thickness variations of 0.05 km are significant, while for the deepest interface, variations of 0.10 km are significant. These estimates of the uncertainties were used in matching the models at intersection points with other lines: velocities were matched to ±0.05 km/s and interfaces to ±0.05 km. Quantitative error bounds cannot be assigned to specific model parame-ters because of the forward modelling technique used. The uncertainties in the velocities and interface depths are a minimum estimate of the errors since the simultaneous ad-justment of velocities and depths beyond these uncertainties may create a model that will still match the observed travel times. Doubling the velocity and depth uncertainties to 0.1 km/s and 0.1 to 0.2 km may represent a better general estimate of the errors associated with the trial-and-error procedure which was applied in this study. Chapter 4 INTERPRETATION Th i s chapter presents an in terpretat ion of the fifteen in- l ine seismic sections recorded along ten different shot lines (F igure 4.1). To fac i l i tate the presentation and discussion of these sections and their interpretat ions, the fifteen profiles w i l l be d i v i ded into three groups: off-ridge l ines, cross-ridge l ines, and the single along-ridge l ine. Presentat ion of each l ine w i l l fol low a s tandard format w i t h four diagrams for each seismic section: (a) ob-served seismic da ta section, (b) synthet ic seismic section, (c) ray t rac ing for first arr iva l refract ion phases super imposed on the velocity structure mode l , and (d) compar i son of observed and calculated travel t imes. T h e observed da ta i n part (a) are zero-phase bandpass-f i l tered f rom 0-15 Hz , scaled by a factor p ropor t iona l to distance and p lot ted as true relat ive ampl i tudes w i t h a reducing veloc ity of 6.0 km/s. T h e h igh-ampl i tude, low-velocity water-wave ar r i va l has not been i nc luded in the theoret ical c a l c u l a t i o n s— the low-velocity arr iva l near the O B S i n part (b) is the reflection f rom the base of Layer A. The different branches as descr ibed i n Section 3.1.1 are label led on the observed data; the triangles mark the start of each branch and ind icate the pos i t ion of the first break. T h e wavelet used in a l l s} 7nthetic seismogram plots is that shown i n F igure 3.7. For the mode l in part (c), the layer boundar ies and only some of the blocks used i n mode l l i ng are shown. Ve loc i ty changes of 0.05 km/s or more w i t h i n the mode l are label led; numbers shown represent the veloc ity at the top and b o t t o m of the layer. To show raj 7 coverage and yet ma in ta i n a clear presentat ion, on ly representative rays for the refracted arrivals are traced. Other ray groups such as 62 Chapter 4. INTERPRETATION 63 129 20 15 48 11 48 00 55 50 + 47 45 10 129 00 55 128 50 H r-48 10 Endeavour Ridge Airgun Lines OBS 10 Kilometres 15 129 20 15 10 20 129 00 55 128 50 F igu re 4.1: A i r g u n lines model led i n this study. On l y those port ions of the lines model led are shown (cf. F i gu re 2.2). Chapter 4. INTERPRETATION 64 reflected phases and mult ip les have been used to create the theoret ica l da ta presented i n the synthet ic section (b) and the travel-t ime curves (d), but these ray groups are not shown i n the ray d iagram (c). Other O B S on the same l ine, intersection points w i th other l ines, and the locat ion of Endeavour Pudge are label led. T h e observed travel t imes in part (d) are p lo t ted as symbols and the theoret ica l t ravel times as lines. T h e symbols are ± 1 5 ms i n height, the average t im ing error i n the data. In a l l plots, approx imate or ientat ion of the lines is ind icated (e.g., N-S, E - W ) . For l ines hav i ng more than one O B S , one O B S is always located at 0 km. A l l distances, unless otherwise noted, refer to mode l distance. Shown on each set of plots is a smal l version of F i gure 4.1 w i t h the O B S and l ine be ing model led emphasized. 4.1 Off-ridge Lines Of the six lines i nc luded i n this group, three are reversed. Mos t of the f ina l models obta ined differ l i t t l e f rom the start ing model , but any differences w i l l be noted i n the discuss ion of each model . T h e three reversed lines w i l l be described first. 4.1.1 Line 8 (OBS 8 and OBS 5) L i ne 8 runs northwest to southeast between O B S 8, located just to the west of Endeavour R idge on the Pac i f i c P la te , and O B S 5, 15.2 k m to the southeast on the J D F P l a t e (F igure 4.1). T h e record section f r om O B S 8 is of poor qual i ty (F igure 4.2a); da ta f rom the hydrophone channel were used to extend the t ravel - t ime picks f r om the noisier vert ica l component da ta as described i n Section 3.1.2. Noise f rom the sma l l a irgun and whale sounds have further deteriorated the data qualit} 7. Branches that have been ident i f ied are marked on F igure 4.2a. A mult ip ly- ref lected refract ion has been p icked at 2.25 s between 5 and 9 k m . In contrast to data, recorded on O B S 8, the record section f rom O B S 5 NW LINE 8-OBS 8 SE j Branch -6 -'5 -4 -'3 -2 -1 0 1 2 3 4 rj 7 3 9 l'O l'l 1'2 1'3 1'4 l'5 lfe l'7 l'8 l'9 20 21 DISTANCE (KM) r r e 7 8 9 DISTANCE (KM) Figure 4.2: (a) Observed data for Line 8-OBS 8. (b) Synthetic section for (a). N W L I N E 8-OBS 8 DISTANCE (KM) SE -6 -4 -2 0 2 4 6 8 10 12 14 16 18 20 DISTANCE (KM) Figure 4.2: (c) Refracted rays traced through final velocity model for Line 8-OBS 8. (d) Comparison of calculated (lines) and observed (symbols) travel times. A peg-leg multiple in Layer A has been traced. The travel-time increase at 11 km could not be successfully modelled. Chapter 4. INTERPRETATION 67 has a h igh signal-to-noise rat io (F igure 4.3a); i nd i v i dua l branches have been label led. B a t h yme t r i c effects cause undu lat ions i n the travel - t ime curves and in the magn i tude of the ampl i tudes. A mu l t ip le was also picked for this reverse profi le. T rave l t imes for L i ne 8 - O B S 8 (Figures 4.2c and d) are matched well by the calculated travel t imes except for the region between 10 and 12 k m where the data qua l i ty is poor. A t t e m p t s to improve the fit in this region had a degrading effect on the better da ta of O B S 5, so the discrepancy was left. For L i n e 8 - O B S 5, the travel -t ime match is also very good, even for the mu l t ip le (Figures 4.3c and d). T h e large travel-t ime undulat ions were di f f icult to mode l because the t rave l - t ime shifts are more than those due solely to the vary ing bathymetry. Ad ju s tmen t of the thickness of Layer B and decreasing the velocity i n port ions of Layer C enabled the undu lat ions to be model led. B o t h lines exh ib i t an upward curvature in the travel t imes near 6 k m , a feature wh ich was repl icated by increas ing the thickness of Layer B below the r idge to 1.10 k m f rom its average value of 0.73 k m and by lowering the veloc i ty i n Layer C near O B S 8. F igures 4.2a and b and Figures 4.3a and b show the compar isons between the observed and synthet ic record sections for L i ne 8 - O B S 8 and L ine 8 - O B S 5, respectively. Because of its h igh noise level, the compar ison for L i n e 8 - O B S 8 is dif f icult; however, B r a n c h 1, pa r t i cu la r l y as a later arr iva l and B r anch 2 prov ide reasonable fits. T h e h igh amp l i tude B r a n c h 1 arrivals near the O B S have been c l ipped. T h e synthet ic section for L i ne 8 -O B S 5 compares favourably w i t h the real data. T h e amp l i t ude increase near 6 k m is caused p r ima r i l y by construct ive interference of ampl i tudes f rom refracted and near-c r i t i ca l reflected phases. A t distances greater than - 2 k m , the first arr iva l theoret ical ampl i tudes ar is ing f rom refractions w i t h i n Layer D are too large compared w i t h the observed data. T h e observed amp l i tude decrease and the slight travel-t ime delay in the observed da ta could be caused by s t ruc tu ra l or velocity complexit ies beneath the ridge. T h e extended wave coda on the theoret ical sections arises f rom the inc lus ion of a LINE 8-OBS 5 Figure 4.3: (a) Observed data for Line 8-OBS 5. OBS 5 is located at 15.2 km. (b) Synthetic section for (a). Figure 4.3: (c) Refracted rays traced through final velocity model for Line 8-OBS 5. (d) Comparison of calculated (lines) and observed (symbols) travel times. A peg-leg multiple in Layer A has been traced. Chapter 4. INTERPRETATION 70 mult ip ly - ref lected refraction. T h e synthet ic seismograms for L i ne 8 - O B S 5 appear to be delayed relative to the observed seismograms because of the wavelet used; it was more emergent t han the arrivals i n F i gu re 4.3 wh ich causes the large peaks to arr ive at a later t ime t han i n the observed da ta . Th i s is not a p rob lem because on ly the amp l i t ude character of the synthet ic sections was compared to the observed d a t a — t h e travel t imes were matched separately and are shown i n F igure 4.3d. Th i s apparent delay of the synthet ic da ta is also noticeable on other synthet ic d a t a sections. B o t h lines i l l u s t rate features c ommon to a l l real and synthet ic sections analysed i n this study. T h e calculated ampl i tudes for B ranch 1 near the O B S are always too large when compared w i th the observed data. T h e reason that h igh ampl i tudes near the O B S are not observed i n the da ta is because they have been c l ipped by the seismograph system. A t further offsets, the low 0 value assumed (25) has drast ica l ly a t tenuated the amp l i tude of these arrivals so that they more closely reproduce the observed data . Even so, the ca lcu lated ampl i tudes are often larger than those of the observed data, even when the b ranch has an excellent t rave l - t ime match and terminates at the same distance as on the observed data. Because of the reverberations f r om the water-wave arrivals at near offsets, the precr i t ica l reflections shown on the t ravel - t ime diagrams are not observed on the real data. E v e n the high amp l i t ude reflection f rom the base of Layer A shown i n the synthet ic section is not vis ible. Ca re must be taken not to confuse this reflection w i t h the earl ier water-wave arr iva l wh i ch was not model led. 4.1.2 Line 5 (OBS 2 and OBS 8) L i n e 5 is located over the Pac i f i c P l a t e and extends 13 k m southwest f rom O B S 8 to O B S 2 located 10 k m west of the ridge (F igure 4.1). D a t a recorded at O B S 2 (F igure 4.4a) are also of poor qual ity, a l though not as poor as O B S 8 data. O n O B S 2, water-wave Chapter 4. INTERPRETATION 71 arrivals f r om the smal l airgun are apparent as noise s tart ing at 9 k m w i t h an apparent veloc ity of 1.5 km/s. T h e most not iceable feature of this section is the h igh amp l i tude arrivals s tart ing at 2.5 s; they are not continuous across the section so they p robab ly are a comb ina t i on of many arrivals construct ive ly and destruct ively interfering. S im i l a r to O B S 8, the hydrophone channel was used to help extend the t ravel - t ime picks for the p r imary arr ivals. T h e whale sounds on the da ta recorded on O B S 8 (F igure 4.5a) make a poor qua l i ty section even worse. T h e filtering f r om 0-15 H z d i d not remove a l l the effects of these whale sounds because the lower frequency of the wha le noise is below 15 Hz (F igure 3.6). T h e hydrophone d a t a for O B S 8 and the travel t ime f rom L i n e 5 - O B S 2 at the reversed po int distance of 13 k m were used to determine the t rave l - t ime picks on L i ne 5 - O B S 8 beyond 1 k m where the arrivals are obscured by noise. A mu l t i p l e was p icked for b o t h sections. Because the mul t ip les for O B S 2 are h igh i n amp l i tude , a number of different ampl i tude groups were picked. T h e ca lcu lated first ar r iva l t imes for bo th O B S compare wel l w i t h the observed data (Figures 4.4c and d and Figures 4.5c and d). Var ia t ions i n the observed t ravel t imes for B r a n c h 4 arr ivals are p robab ly due to p ick ing error because of the poor qua l i t y of the da ta at further offsets. T h e arcuate shape of the t rave l - t ime curves is pa r t i a l l y caused by the ba thymet r i c h igh between the O B S . In order to ma tch the t ravel - t ime .delay near 3 k m on O B S 2 and near 1 k m on O B S 8, Layer A at 2 k m has been th ickened to 0.56 k m f rom its i n i t i a l thickness of 0.45 k m . To the southwest, the velocities i n Layers C and D increase s l ight ly and the thickness of the top three crusta l layers decreases to ma tch the higher apparent velocity of the B r a n c h 4 arrivals on O B S 8. Due to the h igh noise level and amp l i tude fluctuations on both sections, i t is d i f f icult to compare the ampl i tudes of the first arrivals for the observed da ta to the ca lcu lated data, especial ly at far shot-receiver offsets (Figures 4.4a and b and 4.5a and b). T h e ampl i tudes of B r anch 1 for O B S 2 compare favourably w i t h the observed da ta but it is d i f f icult to Chapter 4. INTERPRETATION 72 LINE 5-OBS 2 NE -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 ^ 9 10 11 12 13 14 15 16 17 DISTANCE (Krl) ^ I I r S 6 7 8 9 OISTRNCE (KM) F igure 4.4: (a) Observed da ta for L i ne 5 - O B S 2. T h e large arrow marks the pos i t ion of arrivals f rom the smal l a i rgun. T h e high amp l i t ude secondary arrivals at ~2 .5 s are a feature unique to this section, (b) Synthet ic section for (a). M a n y different mult ip les were used in an at tempt to repl icate the h igh amp l i t ude secondary arrivals. Chapter 4. INTERPRETATION 73 Figure 4.4: (c) Refracted rays traced through final velocity model for Line 5 -OBS 2. (d) Comparison of calculated (lines) and observed (symbols) travel times. Many different multiples have been traced. See Figure 4.6 for an explanation of these multiples. Multiple RS is concident with R5 to the northeast. Chapter 4. INTERPRETATION 74 Figure 4.5: (a) Observed data for Line 5 -OBS 8. OBS 8 is located at 13 km. The method by which picks were made for this section is shown in Figure 3.9. (b) Synthetic section for (a). Chapter 4. INTERPRETATION 75 F igure 4.5: (c) Refracted rays t raced through f ina l velocit} ' mode l for L i ne 5 - O B S 8. T h e uneven i&y coverage is caused by the blockiness of the model ; T h e travel -t ime curves i n (d) were obta ined w i t h more even coverage, (d) Compar i s on of ca lcu lated (lines) and observed (symbols) travel t imes. A peg-leg mu l t i p l e in Layer A has been traced. Chapter 4. INTERPRETATION 76 determine where the branch terminates because of the h i gh -ampl i tude secondary arrivals para l le l l ing the first arrivals. T h e h igh ampl i tudes at the end of B r anch 3 on bo th O B S have been reproduced by construct ive interference of Layer C and Layer D refractions. M a n y different combinat ions of mult ip les were used i n an a t tempt to ma tch the travel t imes picked for the h i gh -ampl i tude secondary arrivals on L i n e 5 - O B S 2. T h e different ray paths used are shown i n F i gure 4.6; the ca lcu lated t rave l - t ime branches are marked i n F i gu re 4.4d. T h e travel-t ime compar i son for many of these mult ip les is quite poor for O B S 2; however, p i ck ing of the correct t ravel times is d i f f icult because of the reverbera-tory nature of the seismograms and unknown phase shifts. A r r i va l s for RI were repicked assuming a 180° phase shift (caused by reflection f rom the seafloor) bu t the later phases were not, even though the other combinat ions i n F i gu re 4.6 wou ld also have a phase shift. A 180° phase shift wou ld cause the t ravel - t ime picks for the later mult ip les to be earl ier by ~ 5 0 ms. For the reflected phases, on ly one of the different combinat ions i n each set was t ravel - t ime model led; the two-d imens iona l ray- t rac ing p rogram used is able to f ind automat i ca l l y the comb ina t i on consist ing of the refracted ray fol lowed by the reflection. T h e other combinat ions were not traced because they have to be deter-m ined manual ly. W h e n the ampl i tudes of these secondary arrivals were ca lcu lated, the amp l i t ude obta ined was mu l t i p l i ed by the number of different combinat ions; e.g., for the peg-leg mu l t i p l e shown i n F i gu re 4.6a, the ampl i tudes obta ined were doubled. Th i s is on l y an approx imat ion to the ampl i tudes that wou ld be obta ined i f a l l the different combinat ions were model led, but since these ampl i tudes were i nc luded as an at tempt to repl icate the h igh -ampl i tude wave trains and not to constra in the mode l , the approx i -ma t i on is just i f ied. T h e reflected-refracted phases were found manual ly. For the later mul t ip les (R2 to R5), the t ravel - t ime curves shown have been smoothed before p lo t t i ng since they contained many t r ip l i cat ions ; the R5 t ravel t imes are discont inuous because of the sharp interfaces in the model . Chapter 4. INTERPRETATION 77 DISTANCE (KM) 0 3 6 9 12 IS 18 UJ Q Rl A I I » D a DISKING: (KM) 0 3 6 9 12 15 18 _| i I I OISTANCE (KM) 3 E 9 12 15 18 21 24 o A ^ - \ A A / / » DEPTH S 4 V \ /A/ c DEPTH S 4 D c DISTANCE (KM) 3 6 9 12 15 i l l I DISTANCE (KM) 3 6 9 12 15 18 Figure 4.6: Different ray paths used for multiple calculations. Only representative rays are shown for each group, (a) Rl: peg-leg multiple within Layer A . This was the most common multiple modelled, (b) R2: peg-leg multiple between the seafloor and the Layer B-Layer C interface, (c) RS: peg-leg multiple between the seafloor and the Layer C-Layer D interface, (d) R4: double peg-leg multiple within Layer A . (e) R5: reflected-refraction off the seafloor. Chapter 4. INTERPRETATION 78 Converted phases that consisted of shear-wave paths th rough Layer A and Layer B were also examined. These phases have arr iva l t imes close to those obta ined for the mult ip ly - ref lected group Rl, but since their ampl i tudes were negligible, no converted phases are shown i n the ca lculated travel t imes or ampl i tudes. T h e low seafloor velocity provides an inefficient boundary for the conversion f rom compress ional to shear waves (or vice versa) ( P u r d y and Detr ick, 1986), so it is expected that converted phases have negl igible ampl i tudes. E v e n w i th the large number of mu l t ip le phases that were model led, the ampl i tudes of the secondary arrivals are much too weak. Baggeroer et al. (1986) stated that sec-ondary arrivals w i t h a higher amp l i t ude t han the p r imary arrivals may be obta ined by construct ive interference between peg-leg mult ip les and mul t ip ly - re f racted arr ivals; how-ever, they invest igated much later arrivals than what are shown i n these da ta sections and state that a h igher gradient t han what is observed here is required. A h i gh gradient was not obta ined for Layer A , but the above argument may be va l id i f Layer A has a h igh gradient near O B S 2. T h e amp l i tude of the secondary arrivals is s ite-dependent since the secondar} r arr ivals for O B S 8 do not have as h igh an amp l i t ude as those recorded on O B S 2. It must also be emphas ized that ray theory has been used to obta in these am-pl i tudes. Mo re sophist icated theoret ical procedures should be used to better unders tand the process by wh i ch these h igh ampl i tudes are obta ined. 4.1.3 Line 6 (OBS 2 and OBS 9) L i ne 6 is located ma i n l y on the Pac i f i c P l a t e and extends 15.4 k m southeast f r om O B S 2 to O B S 9 located east of the ridge on the J D F plate! W i t h a to ta l length of ~ 3 0 k m , this is the longest l ine model led. T h e da ta f rom O B S 2 again quick ly deteriorate in signal-to-noise rat io (F igure 4.7a). Noise causes rap id amp l i t ude f luctuat ions even for da ta at shot-receiver distances less than 10.km. Beyond 7 k m , the determinat ion of first Figure 4.7: (a) Observed data for Line 6-OBS 2. Because of their extremely poor quality, data beyond 14 km were only loosely matched in the travel-time modelling, (b) Synthetic section for (a). Comparison of this figure to (a) is difficult because of the poor data quality. T i i i i i i i i i i i i i i i i i i i i i i i i r -1 1 3 5 7 9 11 13 15 17 19 21 2: D I S T A N C E (KM) Figure 4.7: (c) Refracted rays traced through f ina l velocity mode l for L i n e 6 - O B S 2. (d) Compar i son of calculated (lines) and observed (symbols) t rave l t imes. T rave l - t ime picks beyond 14 k m fit poorly because they were only app rox imate l y ma t ched . oo o Chapter 4. INTERPRETATION 81 arrivals becomes d i f f i c u l t — t h e hydrophone channne l was used to determine picks beyond this distance. T h e picks f rom O B S 2 beyond 14 k m were on ly approx imate ly matched; m a n y different sets of picks were chosen and the ones shown are those wh ich best fit the theoret ica l t ravel times ca lcu lated f rom the m o d e l obta ined by more closely match ing the reverse l ine. S imi la r to the da ta collected f rom O B S 5, also located on the J D F P la te , da ta f rom O B S 9 is of excellent qua l i ty (F igure 4.8a). H i gh amp l i t ude energy beh ind the B r a n c h 2 arrivals masks the cont inuat ion of the B r anch 1 arr ivals as a later phase. T h e large bathymetr i c changes cause fluctuations i n the travel t imes and ampl i tudes on b o t h L i n e 6 sections. Mu l t i p l e s are v is ible on bo th d a t a sections, bu t the cont inu i ty and amp l i t ude of these mult ip les is not as strong as on other O B S sections and no travel t imes were determined. F igures 4.7d and 4.8d show the compar i son between the ca lcu lated and observed travel times. T h e match for O B S 9 is excellent, bu t the match for O B S 2 is poor, as expected since arrivals beyond 14 k m were diff icult to pick. Near O B S 9, the veloc ity i n Layer A has been increased to better ma tch the B r a n c h 1 arrivals. Layers A and B are thickest at the ridge crest in order to create the t rave l - t ime delay required for the B ranch 3 arrivals near 7 k m (Figures 4.7c and 4.8c). T h e top three layers have been increased i n velocity and th inned to the northwest to reproduce the h igh apparent veloc ity of the O B S 9 data. T h e velocity i n Layer D also increases to the northwest, but poor contro l f rom the O B S 2 da ta makes this re lat ionship tenuous. T w o sets of travel t imes were chosen near 6 km; at tempts to match these w i t h two different arr ivals were unsuccessful. A g a i n the compar i son of the synthet ic section to the observed section for O B S 2 is d i f f icult (Figures 4.7a and b). E r r a t i c amp l i t ude changes at near offsets are not model led; the ampl i tudes beyond 12 k m are imposs ib le to compare because of the high noise level. Anoma lou s structure beneath the ridge may be the reason for the extremely poor da ta recorded on O B S 2 beyond 12 k m , a l though arrivals f rom O B S 9 that travel th rough the Figure 4.8: (a) Observed data for Line 6-OBS 9. OBS 9 is located at 15.4 km. This is the same section that is shown in Figures 3.2 and 3.8. (b) Synthetic section for (a). oo to Figure 4.8: (c) Refracted rays traced through final velocity model for Line 6-OBS 9. (d) Comparison of calculated (lines) and observed (symbols) travel times. The two groups of arrivals near 6 km could not be matched. oo C h a p t e r 4. INTERPRETATION 84 r idge do not appear to be s imi lar ly affected. T h e amp l i t ude match to the O B S 9 data is shown i n Figures 4.8a and b. If arr ivals f rom Layer C extended to 4 k m , the amp l i tude fit w rould be better, but this is not compat ib le w i t h t ravel - t ime da ta f rom O B S 2. 4.1.4 Line 7 (OBS 9) A l t h o u g h L i ne 7 has two inst ruments , O B S 9 and O B S 5, i t was not reversed. O B S 5 was deployed early and because of s to rm delays, had run out of tape before this l ine was f i na l l y shot. Mos t of this l ine is located on the J D F P la te ; i t extends 11.4 k m northeast f rom O B S 9 to O B S 5. T h e da ta section for O B S 9 is of h igh qua l i ty w i t h clear arrivals for the entire length of the l ine (F igure 4.9a). Noise f rom the sma l l a i rgun para l le l l ing the water-wave arrivals is v is ible s tart ing at 9 k m . M a n y later arr ivals are present, but the mult ip les are not as h igh i n amp l i t ude as on other sections. La te r arrivals for a l l da ta recorded on O B S 9 are of lower amp l i t ude t han later arrivals recorded on O B S 2 and 5. In order to ma tch the rap id apparent velocity increase to the northeast beyond 9 km, a l l the layer velocities were increased and the top three layers decreased i n thickness (F igures 4.9c and d). These two effects are s t i l l not enough to ob ta in calculated travel t imes for Branches 3 and 4 that m a t c h the observed data. E v e n th inner layers and higher velocities cannot be reconciled w i t h interface depths and layer velocities under O B S 5 determined f rom other lines. In add i t i on to the poor match of t ravel t imes beyond 11 k m , the near-offset travel t imes of B r anch 1 have not been reproduced. A l t hough the water-wave arrivals were fit, there st i l l appears to be a stat ic shift in the travel times for B r a n c h 1; however, L i ne 7 does not d irect ly pass over O B S 9 and a relat ively large O B S repos i t ion ing correction was needed (124 m) , bo th of wh ich may cause problems in match ing travel times at distances near the O B S . In the synthet ic section (F igure 4.9b), the amp l i t ude of B ranch 1 beyond 4 k m is lower Chapter 4. INTERPRETATION 85 F igu re 4.9: (a) Observed data for L i ne 7 - O B S 9. T h e large arrow indicates noise f rom the sma l l a i rgun. (b) Synthet ic section for (a). Chapter 4. INTERPRETATION 86 sw LINE7-OBS9 D ISTANCE (KM) NE -5 -3 -1 1 3 5 7 9 11 13 15 17 I I 1 I I 1 1 I I 1 I I I f~l I I I I I I I -5 -3 -1 1 3 5 7 9 11 13 15 17 DISTANCE (KM) F igu re 4.9: (c) Refracted rays traced through f ina l velocity mode l for L i ne 7 -OBS 9. O B S 5 indicates the pos it ion of intersect ing lines, (d) Compar i son of ca lcu lated (lines) and observed (symbols) travel t imes. A n even greater velocity increase is needed to the northeast to ob ta i n a better t rave l - t ime f it, but this is not compat ib le w i t h intersecting lines at O B S 5. C h a p t e r 4. INTERPRETATION 87 t han the observed da ta and terminates early. A n increase i n gradient wou ld strengthen the ampl i tudes but terminate the branch earlier, so the veloc ity in Layer A was not changed. T h e gradient in Layer B was increased in an a t tempt to match the amp l i t ude behav iour of B r a n c h 2, but it was not possible to obta in strong ampl i tudes that g radua l l y weaken. T h e gradient i n Layer D was decreased in order to ob ta i n the lower amp l i t ude arr ivals of B r a n c h 4. 4.1.5 Line 9 (OBS 2) L i n e 9 is a " sp l i t - sp read " profi le over O B S 2 ~ 1 0 k m to the west of and para l le l to the ridge. Since this fine is complete ly on the Pac i f i c P l a t e and approx imate ly a long an i sochron, it is not expected to be h igh ly two-d imens ional . Because of the smaller offset distances recorded on this unreversed l ine, the noise usua l ly recorded on O B S 2 is not as p rob lemat ic (F igure 4.10a). A shoal ing of the ba thymet r y causes the travel t imes beyond -11 k m to be advanced significanthy and the cont inu i ty of the ampl i tudes to deteriorate. H i g h amp l i tude arrivals para l le l l ing the first arr ivals are also evident on this l ine, a feature characterist ic of a l l d a t a recorded on O B S 2. Since this l ine is intersected by other lines i n three locat ions, more contro l on layer depths and velocit ies is avai lable than what is no rma l for unreversed lines. A n ex-cellent ma tch to the travel -t ime da ta (F igure 4.10d) was obta ined us ing the mode l i n F i gu re 4.10c. E v e n the peg-leg mu l t i p l e fits most of the p icked travel times. On l y sma l l changes were made to the s tart ing mode l , ma i n l y to improve the ampl i tude fit. T h e ve loc i ty of Layer A near the O B S was increased to match the B ranch 1 travel times. T h e depth of the Layer A - L a y e r B interface is in agreement w i t h that determined f rom the reversed l ines, but the velocity of Layer A is ~ 1 0 % higher. T h e top three layers are th innest near the O B S , but that character ist ic may be due to travel-t ime constraints be ing avai lable on ly near the O B S . Figure 4.10: (a) Observed data for Line 9-OBS 2. (b) Synthetic section for (a). C O oo F igure 4.10: (c) Refracted rays traced through f ina l velocity model for L i ne 9 - O B S 2. Po int s of intersection w i th L ines 3 and 4 are ind icated, (d) Compar i s on of ca lcu la ted (lines) and observed (symbols) travel times. A peg-leg mul t ip le in Layer A has been traced. Chapter 4, INTERPRETATION 90 T h e synthet ic section (F igure 4.10b) also compares favourably w i t h the observed data. B r a n c h 1 ampl i tudes to the no r th are s l ight ly low in ampl i tude; however, those to the south are too h igh in amp l i tude , especially since they are not v is ib le as later arrivals. Cons t ruc t i ve interference f rom B r a n c h 4 arrivals has caused the amp l i t ude max imums in B r a n c h 3 to occur at the correct distances. 4.1.6 Line 10 (OBS 5) L i n e 10 is the f i na l l ine that w i l l be discussed i n this group of off-ridge lines. S imi la r i n conf igurat ion to L i ne 9, L i ne 10 is ~ 1 0 k m to the east of the r idge on the J D F P l a t e and is centred over O B S 5. T h e section shown i n F igure 4.11a represents the best da ta recorded i n this experiment and was used i n an attempt to ob ta in a one-dimens ional mode l i n the wavefield cont inuat ion a lgor i thm (Section 3.2.1). M a n y later arrivals that m i m i c the shape of the first-arrival t ravel times are v is ible and the locat ion of the first arr ivals can be chosen w i thout diff iculty. T h e lack of a decrease i n ampl i tudes near ± 1 2 k m indicates that B r anch 4 arrivals do not appear on the section w i t h i n the usual offsets and that the B ranch 3 arrivals extend farther than norma l . A low-velocity arr iva l w i t h an apparent velocity of 3.2 km/s is v is ible near - 5 k m at a reduced t ime of 3.25 s; the low velocity wou ld most l i ke ly make it a converted shear wave, but its t ravel t ime is too late. A refracted arr iva l that travels entirely as a shear wave through the crust at a ve loc i ty of Vc = Ip/VZ (cr = 0.25) arrives ~ 2 5 0 ms earlier t han this a r r iva l and has a m u c h lower ampl i tude. Three-d imens iona l effects may be an exp lanat ion for this phase. T h e compar i son of the mode l led travel t imes and those obta ined f rom the data is shown i n F i gure 4 . l i d . Layer A velocities again had to be increased near the O B S to ma tch the B r anch 1 arrivals (F igure 4.11c). W i t h an average thickness of on ly 0.36 km, Layer A is the thinnest of a l l the models; the thickest por t ion of Layer A is to the north Figure 4.11: (a) Observed data for Line 10-OBS 5. Many different multiples were chosen from this section. The low-velocity arrival indicated by the arrow cannot be explained by two-dimensional ray tracing, (b) Synthetic section for (a). All the multiples shown in Figure 4.6 were used to obtain the later arrivals. i i i i i i i i i i i i i 1 1 1 1 1 1 1 1 i 1 1 1 1 — r -15 -13 -11 - g - 7 -5 -3 - 1 1 3 5 7 9 11 DISTANCE (KM) Figure 4.11: (c) Refracted rays traced through final velocity model for Line 10-OBS 5. Points of intersection with Lines 3 and 4 are indicated, (d) Comparison of calculated (lines) and observed (symbols) travel times. All the multiples shown in Figure 4.6 have been modelled for this line. R3 was not obtained to the north because of the bathymetry. Chapter 4. INTERPRETATION 93 under the ba thymet r i c high. T h e h igh apparent velocities to the south were reproduced by t h i nn i n g Layers A and B and increas ing the average velocity w i t h i n Layers B and C. Layer A to the south shou ld be thickened to ma tch the i i i mult ip les (F igure 4.6), bu t increas ing the thickness of Layer A creates diff iculties i n match ing the picks for first arr ivals wh ich have much less t rave l - t ime error. T h e same mult ip les that were traced for L i ne 2 - O B S 5 (F igure 4.6) were also model led on this l ine; the travel -t ime f it for some of these mult ip les is excellent. T h e synthet ic section obta ined f rom a l l the different t rave l - t ime branches is shown i n F i gu re 4.11b. T h e h igh -ampl i tude arrivals beyond ± 9 k m are reproduced by th ickening Layer C so that arrivals refract ing th rough this layer t ravel to farther offsets. To the no r th , the th icken ing of Layer C is not constra ined by t ravel - t ime da ta f rom arrivals refract ing through Layer D. Cons t ruc t i ve interference of the R2, R4, and R5 mult ip les to the no r t h creates relat ively h i gh -amp l i t ude later arrivals; construct ive interference f rom other arrivals not model led m a y create the h igh ampl i tudes required to the south. 4.2 Cross-ridge Lines Th i s group consists of three l ines: L i ne 2, L i ne 3, and L i ne 4. L i ne 2 is the most densely sampled of a l l the lines and was mode l led by W h i t e and Clowes (1988) us ing a travel -t ime invers ion procedure; i t is presented here for amp l i tude model l ing . L ines 3 and 4 were mode l led us ing the same method as the previous models. 4.2.1 Line 2 (OBS 12 and OBS 14) T h e cross-ridge l ine w i th its eight O B S was not model led in the same manner as the other l ines. Us ing travel-t ime da ta f rom a l l the O B S , W h i t e and Clowes (1988) appl ied a tomograph i c invers ion method descr ibed by Wh i te - (1988 ) to obta in the model . Th i s Chapter 4. INTERPRETATION 94 technique i terat ive ly modifies a s tart ing ve loc i ty mode l un t i l t ravel t imes ca lcu lated for the mode l match the observed travel t imes. W h i t e and Clowes (1988) solve bo th for veloc ity and interface perturbat ions. Since the i r method does not use amp l i t ude infor-m a t i o n , the two sections w i t h the furthest separat ion and best data qua l i ty were used to determine i f the mode l obta ined f rom the invers ion repl icated the amp l i t ude data. D a t a f rom O B S 12 and 14 were chosen (F igure 2.2); a l though O B S 5 wou ld p robab ly have had excellent qua l i ty da ta , the ins t rument had r un out of tape and this l ine was not recorded. A i r g u n da ta recorded along L i ne E X - 6 (F igure 2.1) does exist, but is offset f r om L ine 2. W 'h i te and Clowes (1988) used the double-a i rgun L i n e E X - 6 da ta i n their invers ion, but the different ba thymet ry in t roduced a comp l i ca t i on i n the fo rward-mode l l ing procedure, so the next closest inst rument, O B S 14, was chosen. L i n e 2 was shot perpendicu lar to the ridge over a ba thymet r i c h igh, coincident w i t h a mu l t i channe l c rus ta l reflection prof i le that was obta ined earlier i n 1985 (Roh r et ai, 1988; Roh r and Mi lkere i t , 1988). O B S 12 is located on the Pac i f i c P la te , 6.5 k m f rom O B S 2 and ~ 4 k m west of the ridge. O B S 14 is located on the J D F P la te , 16.8 k m f rom O B S 2 and ~ 7 k m east of the ridge. L i ne 2 recorded on O B S 12 is shown in F i gure 4.12a. It is contaminated by h igh-apparent-veloc i ty water-wave arrivals f rom the sma l l a i rgun. L i n e 2 recorded on O B S 14 was contaminated by h igh -ampl i tude high-frequency noise (F igure 4.13). F i l t e r i ng f rom 0-15 H z removed most of this noise, but the degradat ion of the da ta at further offsets is apparent. Mo re B r a n c h 1 first arrivals exist on this section because O B S 14 was shallower than O B S 12. A l t h o u g h mult ip les exist on bo th sections, none were picked or model led. Since the tomograph ic invers ion scheme of W h i t e (1988) has a different mode l pa -rameter i zat ion f rom the model l ing a l go r i thm of Zelt and E l l i s (1988a), i t was diff icult to have the same type of layer boundar ies w i t h the same velocity along the entire boundary. Instead of a lter ing the thickness of a layer as was done i n the forward mode l l ing , Wh i t e ' s Figure 4.12: (a) Observed data for Line 2-OBS 12. OBS 12 is located at 6.5 km. The large arrows indicate coherent noise from the small airgun. (b) Synthetic section for (a). C O C n i i i i i i i i i i i i n i i i i i i i i i i i i i i i r -2 0 2 4 6 8 10 12 14 16 18 20 22 24 DISTANCE (KM) F igu re 4.12: (c) Refracted and reflected arrivals for L ine 2 - O B S 12. T h e velocity mode l is not shown because of its different parameter izat ion. On l y those O B S wh ich ind icate intersect ion points w i t h other lines or wh i ch have been model led i n this thesis are label led; da ta f rom a l l eight were used by W h i t e and Clowes (1988). T h e intersect ion po in t w i t h L i n e D5 is ind icated, (d) Compar i s on of calculated (lines) and observed (symbols) travel t imes. Figure 4.13: (a) Observed data for L i n e 2 - O B S 14. O B S 14 is located at 16.8 k m . (b) Synthetic section for (a). CD w LINE 2-OBS 14 DISTANCE (KM) E -4 -2 0 2 4 6 8 10 12 14 16 18 20 22 24 26 T 1 1 1 1 1 1 1 I 1 I 1 I 1 1 1 1 1 1 1 1 I 1 I I I I I I I -2 0 2 4 6 8 10 12 14 16 18 20 22 24 26 DISTANCE (KM) Figure 4.13: (c) Refracted and reflected arrivals for Line 2-OBS 14. The velocity model is not shown because of its different parameterization. Only those OBS which indicate intersection points with other lines or which have been modelled in this thesis are labelled; data from all eight were used by White and Clowes (1988). The intersection point with Line D5 is indicated, (d) Comparison of calculated (lines) and observed (symbols) travel CD times. 0 0 Chapter 4. INTERPRETATION 99 (1988) invers ion method alters the velocity w i t h i n layers, creat ing va ry ing velocities on the top and bo t t om of a layer. On l y three layers are present in the mode l of W h i t e and Clowes (1988), so an interface was placed at the depth of 4.3 k m where the velocity gradient i n the invers ion mode l decreased. T h e f i na l mode l (F igure 4.14) is shown i n isovelocity contour format because of this different parameter izat ion. T h e travel-t ime fits for bo th O B S 12 and O B S 14 are very good but not as good as those shown by W h i t e and Clowes (1988) because fewer blocks were used i n the mode l shown i n F igure 4.14 t han i n the invers ion scheme. Layer A is much better constra ined in this mode l compared to the other l ines because arr ivals f r om eight O B S were used; however, even w i t h the smaller in s t rument spacing, the surface velocity is on l y d i scont inuous ly constra ined and is not constra ined at the axis of Endeavour Pudge. Th i s layer thickens east of the r idge and has higher velocities over and jus t off the ridge. T h e velocity at the top of Layer B varies f rom 4.6 to 4.9 km/s and the velocity at the b o t t o m of the layer is 5.5 km/s. Th i s b o t t o m veloc i ty is 0.2 km/s faster t han that used in the forward-model l ing s tart ing mode l because W h i t e and Clowes (1988) used a higher gradient i n the invers ion mode l (1.0 s _ 1 instead of 0.75 s _ 1 ) . Th i s higher velocity also causes the thickness of Layer B to be greater t han average (0.78 k m vs. 0.65 km). T h e flat Layer B - L a y e r C interface f rom an intermediate mode l prov ided to the author by D.J. W r h i te (pers. comm., 1988) was only changed s l ight ly near 20 k m . ( W h i t e and Clowes (1988) show a model w i t h the Layer B - L a y e r C interface a smoothed version of the seafloor bathymetry ; this is a later version of the L i ne 2 mode l and more closely resembles the models presented in this thesis f rom two-d imens iona l ray tracing.) T h e region of lowest velocity gradient in Layer B is between 4 and 12 k m , part of wh ich is under the ridge. T h e velocities i n Layer D do not differ greatly f r om the start ing mode l , a l though there is a slight velocity increase to the west of the ridge between 2 and 10 k m . Since this l ine is d i rect ly over the ridge, i f a low-veloc i ty zone ( L V Z ) such as a magma Figure 4.14: Velocity model for Line 2. Layer boundaries are shown as thicker lines; the contour interval is 0.1 km/s. Velocities in Layers A, B, C, and D range from 2.5-2.8, 4.6-5.5, 5.7-6.5, and 6.6-7.3 km/s, respectively. The model has been plotted with a vertical exaggeration of two so that more contours could be shown. There is a slight velocity decrease in Layer B beneath the ridge and a slight velocity increase in Layers C and D to the west of the ridge. o Chapter 4. INTERPRETATION 101 chamber exists, its effects should be apparent as a delay in the travel times or a decrease in ampl i tudes of arr ivals that t ravel through i t . Ne i ther of those characterist ics is v i s ib le on this da ta set, and no L V Z was required to match the t ravel t imes ( W h i t e and Clowes, 1988). To match the amp l i tude data , on ly sl ight changes were made to the mode l velocities because data f rom only 2 O B S were used to check the travel -t ime f it. T h e A R T ampl i tudes ca lcu lated (Figures 4.12a and b and 4.13a and b) compare favourab ly w i t h the observed data; the large number of raypaths used i n the invers ion helped resolve the veloc ity gradients i n the model . T h e amp l i t ude fit to the observed da ta for O B S 14 is better than the fit for O B S 12. Interference f rom the sma l l a i rgun to the east on O B S 12 makes compar i son w i t h the B ranch 4 arrivals diff icult. 4.2.2 Line 3 (OBS 8) L i n e 3 is a cross-ridge l ine ~ 1 0 k m north of L i n e 2, but recorded in- l ine into only one ins t rument, O B S 8 on the Pac i f i c P la te . S im i l a r to other da ta recorded on O B S 8, the observed section is of poor qua l i ty (F igure 4.15a). Trave l - t ime picks beyond 11 k m were obta ined w i t h a id of the hydrophone channel. A r r i va l s to the east travel under the ridge wh ich may account for the poorer s ignal compared to arrivals to the west, a l though this phenomenon was not noted for L i ne 2. T h e m a n y later arrivals present are stronger to the west where the O B S is on the same side of the ridge. Travel - t ime f luctuat ions were model led ma in l y by vary ing the thickness of the upper-most layer (Figures 4.15c and d). Layer A is thickest at t i le r idge and thinnest off ridge and is constrained by the models f rom the ridge-parallel lines as wel l as by the P -wave peg-leg mu l t ip le (Rl)- T h e veloc ity in Layer C between - 7 and - 2 k m was decreased s l ight ly so that refractions f rom Laj 'er C wou l d not overtake the arrivals f rom Layer B too earl}-. T h e slight increase i n the velocity of Layer D away f rom the ridge is poor ly Figure 4.15: (a) Observed data for Line 3-OBS 8. (b) Synthetic section for (a). o t o DISTANCE (KM) 4 Figure 4.15: (c) Refracted rays traced through final velocity model for Line 3-OBS 8. Points of intersection with Lines 9 and 10 are indicated, (d) Comparison of calculated (lines) and observed (symbols) travel times. A peg-leg multiple in Layer A has been traced. i—' o oo Chapter 4. INTERPRETATION 104 constra ined because of the short offset distances. T h e L i ne 3 synthet ic section (F igure 4.15b) does not match the observed da ta as wel l as other lines, bu t the chaotic amp l i t ude character of the real da ta is p robab ly due to effects that cannot be model led w i t h the a lgor i thm used. Ca l cu la ted B r a n c h 3 ampl i tudes beyond 6 k m are too strong, but degradat ion of the da ta due to noise makes compar i son diff icult. 4.2.3 Line 4 (OBS 9) L i n e 4, ~ 1 9 k m to the south of L i n e 3, is a cross-ridge l ine over O B S 9 located "on the J D F P late. T w i c e as many a i rgun shots as no rma l were recorded on this l ine because the spacing used was the same as that for the so-called " doub le " a i rgun lines. Th i s is a h igh signal-to-noise section w i t h clear first arr ivals for the length of the l ine (F i g -ure 4.16a). M a n y h igh ampl i tude, low-apparent-ve loc i ty arrivals are present betAveen ± 3 and ± 6 k m — B r a n c h 1 as a later a r r i va l was diff icult to pick because of these reverbera-tions. There is an arr iva l between - 4 and - 7 k m w i t h an apparent velocity of 2.5 km/s; it is not a mu l t i p l e because it occurs earlier than B r a n c h 1 arrivals. Th i s strange phase cou ld not be mode l led and may be caused by out-of-plane structure. Few changes i n the start ing mode l velocities were needed to match the observed da ta (F igures 4.16c and d) and to reta in consistency w i t h the intersect ion po ints f rom the other lines. S im i l a r to a l l l ines recorded on O B S 9, Layer A was increased in velocity near the O B S ; Layer A also th ins away from the ridge. For the later B r a n c h 1 arrivals to the east, the wrong cycle was chosen and not repicked, creating the travel -t ime offset. A sharp vert ica l offset of 100 m i n the Layer A - L a y e r B interface beneath the O B S was required because B ranch 2 arr ivals to the east arr ive earlier than those to the west. U s i ng a s loping boundary instead of an offset causes d i f f iculty in ob ta in ing ray coverage for deeper layers. B o t h sets of travel t imes near - 9 k m have not been matched, a l though Figure 4.16: (a) Observed data for Line 4-OBS 9. These data were obtained with a shot spacing half that of the other single airgun lines. A low-velocity arrival has been indicated by a large arrow; this phase could not be modelled, (b) Synthetic section for Figure 4.16: (c) Refracted rays traced through final velocity model for Line 4-OBS 9. Points of intersection with Lines 9 and 10 are shown, (d) Comparison of calculated (lines) and observed (symbols) travel times. The travel-time picks for Branch 1 arrivals to the east were picked too late and do not match the calculated arrivals. The two groups of arrivals between -10 and -8 km have not been successfully modelled. Chapter 4. INTERPRETATION 107 the earlier set appears to be correct. Ba thymet r i c changes to the west beyond - 12 k m make i t diff icult to obta in evenly-spaced ray coverage, but the travel t imes have been approx imate ly matched. A l l branches obta ined i n the synthet ic section (F igure 4.16b) compare favourably w i t h the observed d a t a (F igure 4.16a). Ba thymet r i c changes, such as those beyond -12 k m , cause many of the ampl i tude highs and lows. T h e velocity gradient i n Layer C was increased i n order to obta in h igher ampl i tudes for B r anch 3 at distances near - 8 k m , but the ca lcu lated ampl i tudes are s t i l l too weak; veloc ity changes w i t h i n Layer C not model led may cause the h igh observed ampl i tudes. T h e h igh ampl i tudes out to 14 k m are reproduced by refractions th rough Layer C since refractions f r om Layer D wou ld have much lower ampl i tudes and a h igher apparent velocity. 4.3 Along-ridge Line—Line D5 (OBS 9 and OBS 8) L i ne D 5 was shot obl iquely a long the ridge between O B S 9 located s l ight ly to the east of the ridge and O B S 8 located s l ight ly to the west of the ridge and is coincident w i t h L i ne E X - 5 shown i n F igure 2.1. Th i s l ine was to have been done w i t h two 2000 - i n 3 airguns, hence its designation as D5 . A l t h o u g h L i ne D 5 extends further to the nor th and south (F igure 2.1), the po r t i on model led here is 30 k m in length and extends 6 k m south of O B S 9 and 5 k m nor th of O B S 8 w i t h the length of reversal 18.8 k m . Un l i ke other l ines recorded on O B S 9, the qual i ty of da ta for L i ne D5 is poor beyond 9 k m (F igure 4.17a). Near 3 km, there was a p rob lem w i th the a i rgun pressure, so the traces are very close together and there is an overlap of traces recorded. Not a l l the traces were p lo t ted so there is no overlap on the section. Beyond 7 k m there is a very sharp decrease in amp l i t ude concurrent w i t h an increase i n the apparent velocity. A n increase i n the noise level beyond 14 k m causes difficult}- in determin ing the locat ion of the first Chapter 4. INTERPRETATION 108 arrivals. A t shot-receiver offsets greater than 17 k m the ampl i tudes appear to increase in strength and there is a large t rave l - t ime delay; however, the first breaks are actua l ly two cycles ahead of the h igh ampl i tudes between 17 and 24 km. L ine D5 recorded on O B S 8 is shown i n F i gure 4.18a. It is s imi lar to a l l other lines recorded on O B S 8 i n that the signal-to-noise rat io is very low and the da ta qua l i t y deteriorates rap id l y w i t h distance. B r a n c h 1 arrivals are not v is ible as later arrivals due to the h igh amp l i tude energy out to 13 k m and 24 k m beh ind the B r a n c h 2 arr ivals. S imi la r to O B S 9, there is a rap id amp l i t ude decrease at 13 k m w i th a s imultaneous travel -t ime advance. Beyond 6 k m , no coherent branches are v is ib le; again the hydrophone da ta were used to extend the distance to wh ich picks could be made. Because of the poor qua l i ty of the da ta recorded on bo th O B S and the larger error i n the a i rgun shot clock for this l ine (Section 2.3.4), the pick s}'mbols i n Figures 4.17d and 4.18d are ± 2 0 ms i n height instead of ± 1 5 ms. F r om a s imple inspect ion of the data , i t can be seen that these da ta are different f rom those recorded off-ridge and across the ridge. A t rave l - t ime advance such as that beyond 5 k m on O B S 9 requires h igh velocities, but the low ampl i tudes require low gradients. T h e travel - t ime advance is pa r t i a l l y caused by the bathymet r i c h igh between the O B S , but it was found not to be enough. W h e n the s tar t ing mode l was used, the B r a n c h 3 arrivals came i n too late whi le the B ranch 4 arr ivals came i n too early. Dras t i c velocity changes were needed. T h e s tart ing mode l structure near the O B S matched the da ta wel l , so the intersection points at O B S 8 and O B S 9 are consistent w i t h other lines (Figures 4.17c and 4.18c). In the m idd le of the mode l near 8 k m , the cross-ridge l ine, L i ne 2, intersects L i ne D5. Us ing the velocity s t ructure of L i ne 2 for L i ne D5 d id not reproduce the observed travel times. To reta in some consistency w i t h L i ne 2 since the velocit ies of the mode l between the O B S had to be increased, the interface depths f rom the L i ne 2 mode l at the intersection po int w i t h L i ne D5 were used and not altered. To match the travel-t ime advance between 6 and Figure 4.17: (a) Observed data for L i n e D 5 - O B S 9. B r a n c h 3 and 4 arrivals are lower i n ampl i tude compared to other sections recorded on O B S 9. (b) Synthet ic section for (a). o t o Figure 4.17: (c) Refracted rays traced through final velocity model for Line D5-0BS 9. The point of intersection with the cross-ridge line, Line 2 is indicated, (d) Comparison of calculated (lines) and observed (symbols) travel times. The symbols for the observed travel times are ±20 ms in height. Figure 4.18: (c) Refracted rays traced through final velocity model for Line D5-OBS 8. The point of intersection with the cross-ridge line, Line 2, is indicated, (d) Comparison of calculated (lines) and observed (symbols) travel times. The symbols for the observed t—1 travel times are ±20 ms in height. t° Chapter 4. INTERPRETATION 113 16 k m (Figures 4.17d and 4.18d), the velocities i n the top three layers between the O B S were increased. T h i s wTas the on ly method that wou ld speed up Layer C arrivals w i thout adjust ing the interface boundar ies constra ined by L i n e 2. T h e velocities obta ined for Layers A and B are very poor ly constra ined because no arrivals f rom this higher velocity region are model led. A low velocity gradient is needed in Layer C because refractions f rom this layer have weak ampl i tudes. T h e most diff icult part of the mode l l i ng procedure was mode l l i ng the delay i n ar-rivals past 16 k m for O B S 9 and beyond 4 k m for O B S 8. T h i s t ravel -t ime delay was accompl i shed i n two different ways. In the mode l shown i n Figures 4.17c and 4.18c, the higher ve loc i ty regions of Layers A , B, and C are located so that arrivals tu rn ing i n Layer D do not travel th rough regions of increased velocities and are not sped up. T h e veloc ity of Layer D was also decreased to create a travel -t ime delay and an amp l i -tude decrease. A second mode l that fit the da ta inc luded a L V Z . T h e high velocities i n Layer C were extended f rom 1 to 17 k m wh i ch created a s l ight ly better t ravel -t ime and amp l i t ude fit for B r anch 3 arrivals. T h e Layer D arrivals then needed to be delayed so a th in (~0.3 k m th ick) L V Z w i t h a velocity of 5.0 km/s was inserted between Layer C and Layer D extend ing from 1 to 17 k m . A veloc i t j 7 of 5.0 km/s was chosen because that was the velocity used by M c C l a i n et al. (1985) for a low-velocity m a g m a chamber under the E P R . T h e first mode l is preferred over this L V Z model because no L V Z was imaged on the more h i gh l y constrained cross-ridge l ine. A l t h o u g h the L V Z model led is t h i n , it extends obl iquely under the ax ia l h igh and its cross-ridge w i d t h wou ld be ~ 3 k m , a size large enough to be seen on L i ne 2. Ob t a i n i n g two different models for the same da ta set emphasizes the non-uniqueness of refract ion da ta interpretat ions. T h e agreement between the ca lcu lated and observed travel-t imes is adequate (F ig -ures 4.17d and 4.18d). T h e ve loc i ty of Layer A near O B S 9 was again increased to Chapter 4. INTERPRETATION 114 improve the fit to the B ranch 1 arrivals. A group of emergent arrivals was chosen be-tween 2 and 3 k m on L i ne D 5 - 0 B S 9 but has not been matched. T h e calculated travel t imes on O B S 9 are late f rom 9 to 15 k m but increas ing the velocities between 6 and 14 k m causes the ca lculated t ravel t imes on O B S 8 to be too early. A larger t ravel - t ime delay is required to reproduce the B ranch 4 arr ivals on both sections, bu t the error in the B r anch 4 picks, especially for O B S 8, is larger t han the pick size shown so the picked travel times were only approx imate ly matched. T h e general amp l i tude character of bo th sections has been repl icated i n the synthet ic sections (Figures 4.17a and b and 4.18a and b). T h e strength of B r anch 1 and 2 am-pl i tudes and the distance at wh i ch they terminate have been reproduced by increas ing the gradient i n Layer B near bo th O B S . E ven though the veloc ity gradient in Layer C was decreased, the ampl i tudes of B r anch 3 arrivals are s t i l l too h igh. A t tenuat i ve effects beneath the ridge and scattering f rom the rough topography w i t h i n the neovolcanic zone may have decreased the observed ampl i tudes. 4.4 Final Model of Region T h e f ina l models obta ined are shown as contour plots and as an average ve loc i ty-depth funct ion. Figures 4.20, 4.22, and 4.24 are contour plots of the thickness of the top three layers whi le F igures 4.19, 4.21, 4.23, and 4.25 are contour plots of the depths of the four different interfaces, i nc lud ing the bathymetry. Regions wi th R M S velocities 0.05 km/s greater or less t han the R M S velocity of the layers i n the start ing mode l are h igh l ighted on the layer thickness plots for the top three layers and on the Layer C - L a y e r D interface p lot for Layer D. T h e ba thymet r y shown i n F i gure 4.19 has been obta ined f rom the seafloor interfaces used i n the different forward models. Th i s interface was p lot ted w i t h respect to the Chapter 4. INTERPRETATION 115 a i rgun depth, not the sea surface and the depths may not coincide w i t h those obta ined f r o m more accurate methods such as Seabeam. Th i s f igure has been inc luded to show the details of the bathymetry used and to show how the deeper interfaces m i m i c the ba thymet r i c highs and lows. T h e thickness of Layer A (F igure 4.20) varies f rom 0.21 to 0.66 km, w i th the average thickness ~0.40 km. T h e 0.4 k m contour is subpara l le l to the r idge and the bathymet r i c h i gh to the northeast, w i t h the thickest regions (0.50 k m and greater) correlat ing w i t h the ridge. To the east and southeast on the oldest crust i n the s tudy area, Layer A is th innest. A velocity of 2.5-2.7 km/s for Layer A was used for most of the study area but there are regions of higher velocit ies, especial ly along lines para l le l to the ridge and to the southeast. These regions of h igher velocities increase the average velocity of Layer 2 A to 2.6-2.8 km/s. T h e R M S ve loc i ty over the ridge nor th of the intersection point between the cross-ridge and along-ridge l ines is 0.7 km/s faster than the average R M S velocity. It mus t be emphas ized, however, that no arrivals f r om Layer A were obta ined along either l ine at the ridge crest. T h e depth of the Layer A - L a y e r B interface (F igure 4.21) is s imi lar to the bathymetr} ' , as expected since this interface was constra ined i n the s tart ing models to m i m i c the bathymetry . T h e bathymetr i c h igh west of the ridge has been f lattened at this interface and the no r th - s ou th cont inu i ty of the ridge is not as apparent. Layer B has the highest ve loc i ty gradient of the four layers i n the models (~0.85 s _ 1 ) . Its thickness varies by over 0.60 k m , f rom less than 0.50 k m to more than 1.00 k m (F igure 4.22). S im i l a r to Layer A , Layer B also is thickest where the bathymetr i c highs exist and where the crust is } roungest; this connect ion is more tenuous than for Layer A , however, because when mode l l i ng , the interface between Layer B and Layer C was usual ly const ra ined to be natter than the shallower interface. In some regions, especial ly to the east, the velocity w i th i n Layer B is 0.1 km/s faster away from the ridge. T h e R M S Chapter 4. INTERPRETATION 116 129 20 48 10-15 10 129 00 55 128 50 •48 10 Bathymetry 5 + 48 00 + 55 50 4- E n d e a v o u r R i d g e A i r g u n L i n e s + 48 00 + 55 + 50 Al 454 129 20 15 10 10 Kilometres 129 00 15 55 20 • 4 7 4 5 -128 50 Figure 4.19: Bathymetry obtained using data from all modelled lines. Data points used were obtained along the dashed lines. Contours from 2.1 km to 2.5 km; contour interval 0.1 km. Chapter 4. INTERPRETATION 117 129 48 10-20 15 10 129 00 55 128 50 -48 10 T h i c k n e s s o f L a y e r A 48 00 55 --50 ---48 00 --55 Endeavour Ridge Airgun Lines Higher velocity region — Lower velocity region 1 t-50 10 15 Kilometres 47 454 129 20 15 10 129 00 55 20 128 -47 45 50 F igure 4.20: Thickness of Layer A . D a t a points used were obta ined along the dashed lines. Regions a long the lines w i t h an R M S veloc i ty 0.05 km/s faster or slower than 2.60 km/s are h igh l ighted. Contour s f rom 0.3 k m to 0.6 km; contour interva l 0.1 k m . Chapter 4. INTERPRETATION 118 129 48 10-20 15 10 129 00 55 128 50 48 10 Depth to Layer A—Layer B Interface 48 00 55 --50 - Endeavour Ridge Airgun Lines --5 48 00 •55 --50 47 454 129 20 15 10 10 Kilometres 129 00 1 5 55 20 47 45 128 50 F i gu re 4.21: D e p t h to Layer A - L a y e r B interface. D a t a points used were obta ined along the dashed lines. Th i s interface is s imi la r i n structure to the bathymetry. Contour s f rom 2.5 k m to 2.9 km; contour interva l 0.1 km. Chapter 4. INTERPRETATION 119 veloc ity w i t h i n Layer B for the m idd l e cross-ridge l ine is higher because a higher gradient was used for L ine 2 (Wh i t e and Clowes, 1988). Velocit ies for Layer B are higher than average along the entire along-ridge l ine. T h e Layer B - L a y e r C interface has some re lat ionsh ip to the bathymetry , especially to the northeast (F igure 4.23). Th i s interface was constra ined to be flat in the middle, cross-ridge l ine wh ich causes diff icult ies i n contour ing. Laye r C is thicker to the east t h a n to the west; the " h i l l " to the northeast is caused by flattening the lower boundary of this layer (F igure 4.24). Because the Layer C - L a y e r D interface is near l j ' flat, where the over ly ing layers are th innest, Layer C is thickest, and vice-versa; this causes it to be thickest on older crust such as that to the southeast and th innest along the ridge. There is a sl ight veloc ity increase i n this layer for arrivals para l le l to the ridge and awa3 r f rom the ridge, especial ly to the east, and a veloc ity decrease for arr ivals near the ridge. A s ignif icant velocity increase and velocity gradient decrease is present on the along-ridge l ine. T h e Layer C - L a y e r D interface is nearly flat, bu t there is is a sl ight sha l lowing of Layer D away f rom the ridge (F igure 4.25). T h e velocities in Layer D are relat ively h igh compared w i t h those obta ined f rom other refract ion surveys, but the velocities at depth are not wel l -control led; even so, there is a sl ight increase i n velocities of this layer away f rom the ridge, especially to the west. T h e lowest velocities i n Layer D occur near the ridge a n d for regions where Layer D refractions have passed under the ridge. T h e average one-dimensional mode l shown i n F i gure 4.26 was obta ined by ca lcu lat ing the average upper and lower velocity of each layer and the average depth of each interface for a l l ten models. T h e average velocities and depths obta ined are s imi lar to those used in the s tar t ing mode l , as expected. Sl ight differences between the start ing and f ina l model occur w i t h i n Layers A and B. Chapter 4. INTERPRETATION 120 129 48 10 20 15 10 129 00 55 5 --48 00 -55 50 --47 45 128 50 •48 10 Thickness of Layer B ^ 4 - 0 . 6 --48 00 --55 Endeavour Ridge Airgun Lines Higher velocity region 50 1 0 Kilometres 1 5 2 0 129 20 15 10 129 00 55' 128 -47 45 50 F igure 4.22: Th ickness of Layer B. D a t a po ints used were obta ined along the dashed lines. Regions a long the lines w i t h an R M S veloc ity 0.05 km/s faster or slower than 5.05 km/s are h igh l ighted. Contours f rom 0.5 k m to 1.0 km; contour interval 0.1 km. Chapter 4. INTERPRETATION 121 20 15 10 129 00 + 55 128 50 •48 10 D e p t h t o L a y e r B - L a y e r C I n t e r f a c e Endeavour Ridge Airgun Lines 129 20 15 10 10 Kilometres 129 00 15 55 20 48 00 55 + 50 47 45 128 50 F i gu re 4.23: Dep th to Layer B - L a y e r C interface. D a t a points used were obta ined a long the dashed lines. Contours f rom 3.0 km to 3.6 km; contour interva l 0.1 km. Chapter 4. INTERPRETATION 122 20 15 — r -10 - f -129 00 55 128 50 48 10 Thickness of Layer C 48 00 ---5 --48 00 --55 Endeavour Ridge Airgun Lines Higher velocity region -Lower velocity region — -i. 50 i 10 Kilometres 129 20 15 10 h — 129 00 15 55 20 128 -47 45 50 F igu re 4.24: Th ickness of Layer C. D a t a points used were obta ined along the dashed lines. Regions a long the Hues w i t h an R M S veloc ity 0.05 km/s faster or slower than 6.05 km/s are h ighl ighted. Contour s f rom 0.8 k m to 1.3 km; contour interva l 0.1 k m . Chapter 4. INTERPRETATION 123 20 15 10 129 00 55 D e p t h t o L a y e r C - L a y e r D I n t e r f a c e 128 50 •48 10 --48 00 --55 Endeavour Ridge Airgun Lines Higher velocity region -Lower velocity region 1 t ---50 10 15 Kilometres 20 15 10 129 00 55 2 0 47 45 128 50 F igure 4.25: D e p t h to Layer C - L a y e r D interface. D a t a points used were obta ined along the dashed lines. Regions along the lines w i t h an R M S velocity 0.05 km/s faster or slower t han 6.90 km/s are highl ighted. Contours f rom 4.2 k m to 4.4 km; contour interva l 0.1 k m . Chapter 4. INTERPRETATION 124 Figure 4.26: Average one-dimensional model of final models. Only slight differences exist between the starting and final model. Chapter 5 DISCUSSION AND CONCLUSIONS T h e layered structure of young oceanic crust -obta ined by the mode l l i ng procedure corre-sponds to the class ical mode l of the oceanic crust discussed i n Section 1.1.1 (F igure 1.1). Layer 2, d iv ided into 2A , 2B, and 2C , has been determined, w i t h the four th layer (D ) the top of Layer 3 (3A) . F r o m an analysis of a l l the lines s tud ied, the fo l lowing questions w i l l be addressed: 1. W h a t are the possible causes of the low veloc ity i n the shallowest layer? 2. W h a t causes the layering and associated velocity discont inuit ies? 3. W h y are the crustal velocities faster on the along-ridge l ine? 5.1 Comparison of One-dimensional Model with Other Refraction Surveys Compar i s on of the average mode l obta ined i n this survey w i t h those obta ined f rom other refract ion surveys i n the region (F igure 1.4 and Tab le 5.1) is not stra ightforward because the other surveys have sparser da ta sets and have been interpreted us ing different tech-niques. T h e survey of Davis et al. (1976) at the Sovanco Trans fo rm Faul t was interpreted assuming isovelocity layers; the mode l for the l ine recorded along the J u a n de Fuca R idge ( J D F R ) is shown i n Table 5.1. T h e average velocity and thickness of Layer 2 determined in the study of Dav i s et al. (1976) agree w i t h those of this s tudy (5.1 km/s and 2.0 km) . A u and Clowes (1982) determined only two layers for Layer 2 f rom a r-p invers ion of t rave l times and model l ing w i t h one-d imens ional W K B J synthet ic seismograms of data 125 Chapter 5. DISCUSSION AND CONCLUSIONS 126 S E I S R I D G D a v i s et al. A u &c C l o w e s M c C l a i n &: L e w i s L a y e r ( 1 9 8 8 ) ( 1 9 7 6 ) ( 1 9 8 2 ) ( 1 9 8 2 ) h (km) (km/s) h ( km) (km/s) h (km) (km/s) h (km) vP (km/s) 2 A 0.40 2.56 2.76 0.9 4.3 4.8 1.0 4.7 2B 0.67 4.80 5.37 2.18 5.2 1.4-2.2 5.0 6.4 5.8 2C 0.94 5.81 6.34 " 1.0 6.5 6.7 3 (A ) 2.74 6.50 7.30 3.68 6.1 var iab le 6.5 3.8 6.9 6.95 h is the thickness of the layer. VP is the P - w a v e velocity. Tab le 5.1: Compar i s on of thicknesses and velocities of oceanic crust obta ined f rom re-f ract ion studies near the J u a n de Fuca Ridge. W h e n two velocities are given, they are the velocities at the top and the b o t t o m of the layer, respectively. T h e thickness of Layer 3 for A u and Clowes (1982) is extremely var iable and so no number is given. Chapter 5. DISCUSSION AND CONCLUSIONS 127 obta ined near the Noo t ka Fau l t Zone. T h e velocities and thicknesses of Layer 2 A are greater than those for this study, but the characterist ics of Layer 2 B are s imi lar to Lay -ers 2 B and 2C combined (1.6 k m and 5.7 km/s). A u and Clowes (1982) also determined veloc ity discontinuit ies or t rans i t ion zones between Layers 2 A and 2B and 2 B and 3. T h e M c C l a i n and Lewis (1982) mode l for the J D F R south of the C o b b Fracture Zone was determined by travel -t ime mode l l i ng only and differs f rom the S E I S R I D G mode l i n hav-ing h igh velocities on ly 1 k m below the seafloor. Layers i n the i r mode l are not label led; those shown in Tab le 5.1 have been assumed. T h e basic crusta l mode l of this s tudy does not differ greatly f rom other results near the Endeavour R idge. T h e M c C l a i n and Lewis (1982) mode l is anomalous, but it also does not agree w i t h general crusta l models such as that determined by H o u t z and E w i n g (1976; Table 1.1). 5.2 Layer 2 5.2.1 Layer 2A One of the more interest ing results of this exper iment is the detection of a t h i n upper-most crusta l layer w i t h a low velocity. T h e surface velocity of 2.5 km/s is one of the lowest velocities ever reported for the oceanic crust. Th i s layer cannot be model led w i th an extremely h igh vert ica l gradient as has been done i n other surveys (e.g., Spud ich and Orcu t t , 1980b; E w i n g and Pu rdy , 1982) because arrivals f rom B r a n c h 1 are sometimes v i s ib le as secondary arrivals. If this layer were model led w i th a h igh gradient, a t r ip l i ca -t ion of the travel -t ime curve wou ld occur w i t h i n the h igh -ampl i tude water-wave arrivals and no refracted or reflected arrivals wou ld be present beh ind the first arrivals. Inter-preters of other refraction surveys obta in a gradient for this layer by ext rapo lat ion since the direct water wave usual ly masks any arrivals f r om this uppermost layer; however, Chapter 5. DISCUSSION AND CONCLUSIONS 128 these surveys inc lude port ions of Layer 2B i n their calculat ions wh ich greatly increase the velocity gradients obta ined (3.5-4.4 s _ 1 ; e.g., W h i t m a r s h , 1978; E w i n g and Purdy , 1982; B r a t t and Pu rdy , 1984). If Layer 2 A determined from this survey was model led as a veloc ity increase f rom 2.5 km/s at the surface to 4.8 km/s at 0.45 k m depth, the gradient wou ld be 5.1 s _ 1 ; i f the tota l thicknesses of 2 A and 2B determined here are combined, w i t h the velocity at the surface 2.5 km/s and the veloc ity at 1.0 k m . 5.3 km/s, the gradi-ent i n this layer wou ld be 2.5 s _ 1 . Th i s range of poss ible gradients (2.5-5.1 s _ 1 ) is s imi lar to the range of h igh gradients obta ined i n other refract ion surveys. A refract ion survey on the M i d - A t l a n t i c R idge ( M A R ) obta ined a low gradient in the uppermost oceanic crust ( P u r d y and Detr ick, 1986). U s i ng explosive sources and ocean b o t t o m hydrophone ( O B H ) receivers, P u r d y and Det r i ck (1986) obta ined a 450 m-thick low-velocity layer (2.55 km/s at the top of the layer and 2.65 km/s at the bot tom) near one O B H f rom observat ion of a low-velocity phase that was v i s ib le out to ~ 7 k m and f rom peg-leg mu l -tiples f r om the interface below this layer. These results are very s imi lar to those of this study. M c C l a i n et al. (1985) also determined a t h i n (<400 m) low-velocity-gradient layer f rom explos ion da ta over the Ea s t Pac i f i c Rise ( E P R ) at 13°N; the seafloor velocities of 2.9-3.6 km/s that they determined are higher t han those obta ined i n this study. F r om E S P ( E xpand i n g Spread Prof i le) da ta i n the same region of the E P R , Ha rd i ng et al. (1988a; 1988b) also determined a Layer 2 A but on ly ~100 -200 m th ick w i t h velocities i n the range 2.35-2.6 km/s. M o r e accurate surveys us ing b o t h bo t t om sources and bo t t om receivers have, however, also determined h igh gradients w i t h i n Layer 2A. P u r d y (1987) obta ined a seafloor velocity of 2.1 km/s and a velocity gradient of 4 s - 1 w i t h i n the med ian val ley of the M A R ; Jacobsen and Lewis (1988) determined a veloc ity of 2.7 km/s and a gradient of 4.6 s _ 1 for 0.4 M a o ld crust near the B l an co Fracture Zone (F igure 1.4). F r om sonic logging i n D S D P Hole 504B, Sal i sbury et al. (1985) obta ined h igh velocities (4 to 5 km/s) and high Chapter 5. DISCUSSION AND CONCLUSIONS 129 gradients (4.4 s _ 1 ) w i t h i n the top 200 m of the oceanic crust below the sediments, but Hole 504B is on older crust (~6.2 M a ) wh i ch is sealed w i t h a th ick layer of siliceous sediments (Ander son et al, 1982). In a l l these cases, however, the thickness of Layer 2 A is on ly 200 m, hal f tha t determined here. To create the refract ion interface, a h igh gradient most l ike ly exists at the base of Layer 2 A near Endeavour Ridge, but it was not determined because of the frequencies of the data. T h e b o t t o m source -bot tom receiver surveys may be better able to determine this h i gh gradient layer, especial ly i f i t is th inner that that model led i n th is survey. H igher gradients may exist i n this layer because secondary arrivals are not observed on a l l s e c t i on s—Laye r 2 A is l ike ly more heterogeneous than is suggested by the forward mode l l i ng : i t was sampled i n very few locat ions and seismic refract ion waves have an inherent low resolut ion. Large hor i zonta l velocity gradients over distances of 1-3 k m as found by Stephen (1988) at Ho le 504B may also exist. T h e low veloc i ty obta ined is l i ke l y the result of h igh poros i ty caused by vert ica l fis-sures, subhor i zonta l cracks (Ander son and Newmark , 1985), and voids wh ich are a l l f i l led w i t h seawater, lower ing the average veloc ity of the basalt. T h e impor tance of poros i ty i n determin ing seismic velocities has been noted by other authors (e.g., W h i t m a r s h , 1978; E w i n g and Pu rdy , 1982; Spud ich and Orcut t , 1980a) and studied by Moos and Zoback (1983). Depend ing on the shape and or ientat ion of cracks, the composite velocity of a Layer 2 A cons ist ing of basalt and seawater could range anywhere between the Hash in -Sh t r i kman (HS) upper and lower bounds of 5.52 km/s ( < 5 % poros i ty) and 2.56 km/s ( 5 0 % poros ity; F i gu re 5.1). These velocities d im in i sh as the pores get natter; the veloci-ties for spher ical pores are shown i n F i gure 5.1 wh i le the HS lower bound is for extremely th in discs. O the r theories show large-scale cracks are more efficient in lowering crustal velocities than the mode l of penny-shaped cracks and spherical pores used to obta in the curves in F i gure 5.1 (Spudich and Orcu t t , 1980a; 1980b). Poros i ty measurements f rom Hole 504B are 1 2 - 1 4 % for Layer 2 A (top 100 to 150 m; Becker, 1985) w i t h velocities Chapter 5. DISCUSSION AND CONCLUSIONS 130 ] I 1 — : 1 i i 0 10 20 30 4 0 50 POROSITY (%) Figure 5.1: P-wave velocity of a fractured rock as a function of porosity. The initial unfractured rock is taken to be Hyndman and Drury's (1976) average Mid-Atlant ic Ridge basalt. A t any porosity, composite velocity can range between the Hashin-Shtrikman (HS) upper and lower bounds, depending on pore shape and orientation. The dashed line shows velocity as a function of porosity if all pores are spherical. (Figure and caption from Spudich and Orcutt. 1980b.) Chapter 5. DISCUSSION AND CONCLUSIONS 131 from 3.3 to 5.7 km/s (Moos et al., 1986); the lowest velocities i n 2 A are i n p i l low basalts. These sonic-log velocities are higher than those obta ined f rom this refraction survey, but Ho le 504B is on older crust (~6.2 M a ) wh ich is sealed w i t h sediments that accelerate cementat ion of fractures and cracks. Seismic velocities are also dependent upon conf in ing pressure and pore fluid pressure since pressure closes cracks and voids and decreases the poros i ty of the basalt. T h e low gradient i n Layer 2 A can be exp la ined s imply by increas-ing differential pressure (the difference between conf in ing pressure and pore pressure) w i t h depth. Ve loc i t y var iat ions w i t h i n Laye r 2 A may be caused by di f fer ing amounts of differential pressure, changes i n the geometry of fissures, cracks and voids, and by differing amounts of a l terat ion minerals that fill the cracks and voids. For example, the region about O B S 9 has a general ly h igh velocity; since it is near the flanking por t ion of the ridge wh ich K a p p e l and R y a n (1986) found to be less fissured, velocities in this region should be higher. T h e highest velocities obta ined for Layer 2 A were over the flanks of the ridge, wh ich have few fissures (Crane et al, 1985), especial ly to the no r th of the cross-ridge l ine where the velocities are highest in La}'er 2A. T h e narrow zone of faults and Assur ing, the neovolcanic zone, should have lower velocities, bu t none were found. Th i s may be because the neovolcanic zone is on ly 1 k m wide and the travel -t ime da ta along the ridge were poor. H i g h velocities at the ridge were not determined f rom the cross-ridge l ine, but the region of increased velocities determined from the along-ridge l ine is narrow perpendicu lar to the r idge and to the nor th of the cross-ridge l ine. T h e m o d e l of fissured ax ia l graben and unfissured ridge flanks was used by T i vey and Johnson (1987) to exp la in changes i n magnet ic field strength, stronger magnet ic signatures be ing measured over unfissured crust. M c C l a i n et al. (1985) also found upper crusta l velocities on the E P R rise axis at 13°N to be higher than those off-ridge. Hyd r o t he rma l vents w i t h h igh water temperatures ind icated large, i so lated fissures at the rise axis rather than extensive cracking. Chapter 5. DISCUSSION AND CONCLUSIONS 132 For l ines para l le l to the r idge (Lines D5 , 9, and 10), the velocity of Layer 2 A is 0.2 to 0.7 km/s faster than the ve loc i ty for arrivals perpend icu lar to and obl ique to the ridge. T h e p redominant l y r idge-paral le l fissures formed by tectonic extension evident i n seafloor mapp i n g da ta and v i s ib le from submersibles (e.g., T i v e y and Johnson, 1987) cause this anisotropy. An i so t ropy has previous ly been detected i n Layer 2 (e.g., W h i t e and W h i t m a r s h , 1984) but this survey shows that i t appears to be confined to the up-permost part of Layer 2, i n agreement w i t h Stephen (1986) who, through an obl ique seismic exper iment centred on D S D P Hole 504B, found the an isotropy to be confined to the uppermost 500 m. Mo re accurate studies of P -wave anisotropy such as those done by Stephen (1985; 1986) determined velocities wh ich ranged from 4.0 to 5.0 km/s (i.e., an i sot ropy of 1 km/s) wh i le those of Shearer and Orcu t t (1986) found velocity anisotropy of 0.2 km/s w i t h i n the top 1 to 1.5 k m . B o t h of these studies concluded that para l le l cracks w i t h i n the upper crust are the most l ikely cause of the anisotropy. F r om the lines in terpreted i n this survey, i t is not possible to determine the type of anisotropy, but the upper crust is most l ikely transversely i sotropic w i t h a hor i zonta l s ymmetry axis as found by S tephen (1985; 1986) for the crust near Hole 504B. It must be emphas ized, however, that the higher velocities para l le l to the ridge are not weD constra ined, and may s imply be caused by latera l heterogeneities. Observat ions of polar ized converted shear waves wou l d be required to determine i f these higher velocities are caused by anisotropj r or la tera l heterogeneities. F r ac tu r i ng also causes the at tenuat ion of ampl i tudes for arrivals refracting i n Layer 2A. A low Q of 25 (Jacobson and Lewis , 1988) was used i n the synthet ic seismogram mod -el l ing to account for the scatter ing expected in Layer 2A. In the mode l l i ng program (Zelt and E l l i s , 1988a), this Q is assumed to be anelastic ( intr ins ic) at tenuat ion, even though the observed da ta inc lude a combinat ion of anelastic and apparent attenuat ion ( s t ruc tu ra l effects due to scatter ing, in t rabed mult ip les , etc.). Th i s combinat ion is called Chapter 5. DISCUSSION AND CONCLUSIONS 133 effective at tenuat ion (Jacobson et al, 1984). W i t h o u t this as sumpt ion of a low Q value, synthet ic seismograms for Layer 2 A arrivals ( B ranch 1) wou ld have been too h igh i n am-p l i t ude at a l l distances; using a low Q attenuates the B r anch 1 arrivals as later arrivals so that they more closely resemble the observed data. H i gh attenuat ion of seismic waves has also been observed in dr i l l -hole studies. D u r i n g well- logging i n Ho le 504B, seismic body waves were l i i g l i l y attenuated i n the h i gh l y porous, extensively f ractured uppermost 100 m (Ander son et al, 1982). A cont inuous interface at a depth comparab le to that obta ined f rom this refraction survey has on ly imaged at two sites: beneath the E P R (Herron, 1982) and beneath the Endeavour R idge (Rohr and M i lkere i t , 1988). Herron (1982) descr ibed a reflector obta ined beneath the E P R at 9 ° N as the base of the porous p i l l ow layer; the interva l ve loc i ty (2.8 km/s) and thickness (420 m) are very s imi lar to those obta ined i n this refract ion survey. R o h r and Mi lkere i t (1988) obta ined a shal low in termi t tent reflector on the reflection l ine coinc ident w i t h the midd le cross-ridge reflection l ine (L ine 2). Is the reflector they obta ined the same interface that creates the veloc ity d i scont inu i ty i n the refract ion data? R o h r and M i l kere i t (1988) obta ined interva l velocities for the layer above the interface of 3 .0 -3 .5±0.5 km/s and thicknesses of 0 .6-1.0±0.10 k m ; the refraction interpretat ion for the cross-ridge l ine obtained R M S velocities of 2.6±0.2 km/s and thicknesses of 0.26-0 .66±0.05 k m ( W h i t e and Clowes, 1988). T h e reflection and refraction velocities thus m a t c h w i t h i n their errors, a l though the reflection velocities are generally higher. T h e interface depths match w i th i n their errors i n places, but do not i n most places. If the refract ion velocity instead of the reflection velocities is used w i t h the reflection travel t imes, the depth of the reflector decreases by ->-25% to 0.45-0.75zbO.05 km, close to the depth of the refract ion interface a l though s t i l l s l ight ly too deep (F igure 5.2). T h e reflec-t i on interface is much too shallow to be the second refraction interface. T h e presence of w LINE 2 - 4 - 2 0 2 I I I I I I (NJ LL) Q in ' CO J L DISTANCE (KM) 8 10 12 14 _J I I I I I I 16 18 20 22 _ l 1 I I I l l 24 E 26 Comparison of Reflectors and Refraction Interfaces 2A 2B 2C Ridge T T -shallow reflector xiii axial reflector intermediate reflector Q CJ C o O § Q O O o C o Figure 5.2: Compar i son of reflectors and refraction interfaces for L i n e 2, the cross-ridge l ine. T h e reflectors f rom Roh r and Mi lkere i t (1988) have been converted to depth us ing the refraction velocities. T h e triangles ind icate the locat ions of the eight O B S f rom wh ich da ta were used in the tomographic travel-t ime invers ion. ( F i gu re after W h i t e and Clowes, 1988.) Chapter 5. DISCUSSION AND CONCLUSIONS 135 peg-leg mult ip les f r om the first refract ion interface supports the choice of the uppermost interface. T h e different frequencies of the two types of da ta and error i n determin ing the ref lect ion travel t imes could cause the discrepancies in the interface depths. Tranverse an isotropy could cause the differences i n velocities since refracted arrivals have nearly ho r i zonta l paths and reflected arrivals have nearly vert ica l paths. R o h r and M i l kere i t (1988) proposed two different models to exp la in the shallow reflec-tor. T h e first is s imi la r to that given by Her ron (1982) who suggested that the reflector is the base of the porous p i l low layer. U s ing poros i ty i n fo rmat ion f rom D S D P holes 395A ( H y n d m a n and Sal isbury, 1984) and 504B (Sa l i sbury et ai, 1985), Roh r and Mi lkere i t (1988) stated that the t rans i t ion f rom h igh porosity, low veloc ity basalts to low porosity, 6.0 km/s basalts cou ld reflect seismic energy. M a p p i n g of oph io l i te sequences (Casej ' et al, 1981) shows that this basalt-diabase dike contact may be sharp enough to reflect seismic energy. T h e second possible cause of the reflector is a metamorph i c front caus-ing a decrease i n poros i ty (Roh r and M i l ke re i t , 1988). H i gh temperature minera l i zat ion occurs over 80 m i n the top of the p i l low to dike t rans i t ion zone i n Hole 504B ( A l t et al., 1986) and cou ld create a reflection. T h e refract ion da ta do not support the first mode l for the interface. T h e second interface at a depth of ~1 k m is more l ike ly the contact between the p i l low basalt layer and the sheeted dikes; the velocity of 5.7 km/s obta ined by Her ron (1982) for the 1.42 k m -th ick layer beneath his interface is comparab le to the average velocity (5.7 km/s) of the 1.6 km- th i ck combined Layers 2B and 2C i n this refract ion model . Roh r and Mi lkere i t (1988) use poros i ty da ta obta ined by logging down Hole 504B (Becker et al., 1982; Becker, 1985) to support their exp lanat ion of an interface between 2B and 2C (F igure 5.3), but a poros i ty change near 400 m in F igure 5.3 (225 m below the sediments) could also be an interface. Becker (1985) stated that the boundar ies of Becker et al. (1982) were a rb i t ra r i l y assigned to the midpo int s of the two zones of steep gradients in F igure 5.3 but Chapter 5. DISCUSSION AND CONCLUSIONS 136 Apparent bulk porosity (%) 0 2 4 6 8 10 12 14 Log, 0 bulk permeability (m 2) F igure 5.3: Va r i a t i on of apparent bu l k poros ity and measured permeabi l i ty w i t h depth in Ho le 504B. Ver t i ca l bars are bu lk permeab i l i t y values averaged over the ind icated vert ica l intervals. T h e approx imate boundar ies between Layers 2A, 2B , and 2C are arb i t ra r i l y based only on the gradients i n the apparent bu lk porosit ies. (F igure adapted f rom R o h r and M i lkere i t , 1988 after Becker et ai, 1982.) Chapter 5. DISCUSSION AND CONCLUSIONS 137 the analysis of Sa l i sbury et al. (1985) suggests that a better def in i t ion of Layers 2A , 2B, and 2C could have been obta ined f rom the large-scale resistivities and bu lk porosities by choosing the boundar ies between these layers so that the two zones of gradients i n bu l k porosities were i nc luded entirely w i t h i n the respective over ly ing layers. Th i s wou ld increase the thickness of Layer 2 A i n F i gure 5.3 to ~ 3 0 0 m so that it more closely agrees w i t h the thickness of the interface obta ined i n this study. Compress iona l velocities logged i n Ho le 504B also support an interface i n Ho le 504B at a depth close to the depth of the refracted interface (Sa l i sbury et ai, 1985). T h e above arguments thus show that the first mode l of R o h r and M i l kere i t (1988) does not agree w i t h the refract ion data , but that the second mode l , a poros ity change w i t h i n the p i l low basalts, is va l id . A s descr ibed above, the extremely low veloc ity of the uppermost layer can on ly be caused by h igh porosity. T h e depth to the first interface is most l i ke ly the extent of pervasive f ractur ing, or as found i n Ho le 504B, the depth at wh i ch poros ity and permeab i l i t y drop because of an increase i n the degree to wh ich low temperature a l terat ion minerals such as clays and zeolites f i l l fractures and voids (Zoback and Anderson, 1982). A l i tho log ica l change f rom most ly p i l low basalts, t h i n sheet flows and breccias to th icker sheet flows cou ld also a id i n the poros ity decrease at the interface by decreasing the size and number of vesicles. S im i l a r to the da ta anahyzed by Herron (1982), a shallow reflector is observed on a ref lect ion l ine wh i ch extends f rom Endeavour R idge east to the cont inenta l marg in ( A . J . Ca lve r t , pers. comm., 1988). On l y a por t ion of this l ine has been processed, so it is not certa in that this reflector is the cont inuat ion of the shal low reflector seen at the ridge crest by Roh r and M i l ke re i t (1988). Interval velocities between the sed iment -Layer 2 A interface and the reflector on ~ 9 M a - o l d crust are higher (3 .8-4.0±0.3 km/s) than those obta ined over zero-age crust and the interva l times are less (0.15-0.20 s) so that the thickness of this layer (0.57-0.80x0.05 km) is s l ightly th inner than the layer to the west Chapter 5. DISCUSSION AND CONCLUSIONS 138 near the riclge. A n increase i n in terva l velocity f rom crust at the ridge to older crust is also observed in Herron ' s (1982) da ta . Interpretat ion of the refract ion da ta reveals a t h i nn i ng of the uppermost layer away f rom the ridge (F igure 4.20). Th i s t h i nn i ng could instead have been par t ia l l y model led by an increase i n velocity of the uppermost layer. T h e reflection da ta (Rohr and Mi lkere i t , 1988) show th i nn i ng of the uppermost layer to the east on the J D F P l a te but ind icate a sl ight th icken ing of the layer to the west (F igure 5.2). T h e refraction d a t a also reveal that Laye r 2 A is thickest under ba thymet r i c highs, an observat ion i n agreement w i t h the hypothesis that this layer is del ineated by a zone of h i gh porosity. M a n y of these ba thymet r i c highs are volcanic i n or ig in (Francheteau and B a l l a r d , 1983) and formed of more porous p i l low basalts l ike the ax ia l f lanks of the ridge. F ractur ing wou ld occur as these vo lcan ic edifices are rafted away f rom the ax ia l r idge, further increasing the poros ity of ba thymet r i c highs. T h e detect ion of Layer 2 A velocities increasing w i t h age is wel l -documented. Hou t z and E w i n g (1976) and Hou t z (1976) described an increase i n Layer 2 A velocity f rom 3.3 km/s at the ridge crest to 5.2 km/s at crusta l ages of ~ 1 5 M a at wh ich t ime Layer 2 A becomes ind i s t ingu i shab le f rom the under ly ing Layer 2B. T h e reflection da ta document an increase f rom 3.0-3.5±0.5 km/s at the Endeavour R idge crest (Rohr and Mi lkere i t , 1988) to 3 .8-4.0±0.3 km/s at 9 M a ( A . J . Ca lver t , pers. commun., 1988), a somewhat smal ler increase t han that of H o u t z and E w i n g (1976). To exp la in a reflection bo th beneath the ridge crest and i n the oceanic crust up to the age of 9 M a w i t h an associated increase in velocit}', the fo l lowing model is suggested. Pervas ive f ractur ing at the ridge crest and large numbers of void-f i l led basalt flows cause the extremely low velocities observed, w i th the depth of the fai lure of rocks by fractur-i ng control led by therma l gradients (Lister, 1974). A w a y f rom the ridge, hydro therma l Chapter 5. DISCUSSION AND CONCLUSIONS 139 c i rcu la t ion and conduct ive cool ing cause h igh temperature and low temperature meta -morph i sm, respectively, creat ing a l terat ion products that p l ug the fractures and pore spaces, changing the mineralogy of Layer 2 A a n d replacing low-veloc i ty seawater, thereby increas ing the veloc ity of Layer 2A. Since the permeab i l i t y of Layer 2 A wou ld also de-crease, especial ly w i t h the th ick tu rb id i te cover on the J D F P l a te , a porosit} ' as we l l as a mineralog} ? contrast wou ld s t i l l exist between Layers 2 A and 2B , wh ich wou l d be detectable as a reflector. T h e depos i t ion of sediments wou ld accelerate the fo rmat ion of metamorph i c minerals by sealing the crust and increas ing its temperature. For such a mode l , the amp l i t ude of seismic energy reflected f rom the interface i n older crust shou ld be weaker than that reflected f rom the interface i n young crust, but this has not been determined. T h e behav iour of Layer 2 A on the Pac i f i c P l a t e should be different f rom that on the J D F P l a t e because of the lack of sediment cover to the west. R o h r and M i l kere i t (1988) pred icted that cracks in the upper J D F crust should f i l l faster t han those i n the upper Pac i f i c crust. If this is true, the velocity of Layer 2 A to the west should be lower and have a higher poros i ty contrast w i t h a stronger reflector amp l i tude t h a n crust of s imi lar age on the J D F P la te . T h e refraction mode l l i ng does not show lower velocities for Layer 2 A on the Pac i f i c P l a te , but constraints on the thickness and veloc ity of this layer are poor so that any changes that have occurred i n jus t 0.40 M a wou ld be diff icult to detect. 5.2.2 Layer 2B A decrease i n poros i ty and an increase i n conf in ing pressure wou ld most l ike ly cause the increase i n ve loc i ty in Layer 2B f rom 4.8 km/s at the top of the la3'er to 5.3 km/s at the b o t t o m of the ~0 .65 k m thick layer. However, i n Ho le 504B, Zoback and Ander son (1982) found an increase i n fractures and vo id density w i t h depth i n Laye r 2 B — t h e increase i n the sonic-log average P - w a v e velocity to 4.8 ± 0.8 km/s (Newmark et al, 1985a; 1985b) Chapter 5. DISCUSSION AND CONCLUSIONS 140 was a t t r ibuted to an increase i n clays and zeolites filling fractures and voids. A higher p ropor t ion of massive sheet flows and an increase i n the propor t ion of sheeted dikes at depth as found by A d a m s o n (1985) i n Hole 504B wou ld also decrease the poros i ty and increase the veloc ity gradient i n this layer. A m p l i t u d e fluctuations of arrivals refract ing th rough this layer that are not matched by topograph ic focussing and defocussing i m p l y changing velocity gradients, poss ib ly caused by la tera l var iat ions i n the rat io of p i l l ow basalts to sheet flow basalts, by changes i n the amount of a l terat ion products depos ited by hyd ro the rma l c i rcu lat ion, or by differential pressure changes (Chr i stensen, 1984b). A s described i n Section 5.2.1, the interface at ~1 k m depth p robab ly corresponds to the volcanic extrus ives-sheeted dike boundary. T h e dramat ic change i n l i tho logy at 1055 m below the sea floor in Ho le 504B supports a minera log ica l change f rom extrusives to massive units and dikes ( Adamson , 1985) as wel l as the drop i n poros i ty f rom 7 - 1 0 % w i t h i n Layer 2 B to < 3 % in Layer 2C (F igure 5.3) determined f rom res ist iv i ty logging (Becker, 1985). Th i s interface is not at a constant depth and s imi lar to the interface above, it is not flat (F igure 4.23). Studies f r om the B a y of Islands ophio l i te have deter-m ined that the interface between Layer 2B and 2 C is not p lanar but is wavy i n three dimensions w i t h a m a x i m u m relief of several hundred metres and a range i n wavelength f rom a few hundred metres to a lmost a k i lometre (Casey et al., 1981), s imi lar to the Layer 2 B - 2 C interface determined i n this refract ion survey. A reflector interpreted to be f rom the base of Layer 2B has been obta ined i n the A t l a n t i c by Musgrove and A u s t i n (1983). 5.2.3 Layer 2C T h e velocities of the th i rd layer (5.8-6.3 km/s) , Layer 2C, are s imi lar to those of un -fractured basalt (~6.0 km/s). T h e slight increase i n velocity away f rom the ridge (F i g -ure 4.24) could poss ib ly be caused by the closing of any fractures created at the ridge Chapter 5. DISCUSSION AND CONCLUSIONS 141 and a. drop in tens ional stresses and strains (Brocher and Ten B r i n k , 1987). For example, Layer 2C i n Hole 504B was found to conta in subhor izonta l fractures wh ich were l ikely formed at the ridge axis by hor i zonta l extension (Newmark et ai, 1985a; 1985b). C los ing of fractures l ike these would increase the ve loc i ty of Layer 2C away f rom the ridge. Re-gions of lowered velocities may conta in more of these fractures or may s imply be caused by heterogeneities in poros i ty or differential pressure w i t h i n Layer 2C. D S D P sonic log-ging i n Hole 504B obta ined velocities of 5.6 ± 0.3 km/s (Newmark et al., 1985a; 1985b) for Layer 2C; these velocities are low because the complete layer has not been dr i l led. T h e b o t t o m of this layer is the depth l im i t of the sheeted dikes and the beg inn ing of the massive gabbro formed w i t h i n a m a g m a chamber. T h e roof of a m a g m a chamber is hypothes ized to be formed of gabbro crystals freezing f rom the top of the chamber. Th i s change i n crys ta l l in i ty (i.e., crystals i n gabbro are coarser than those i n basa l t ) could cause an increase in velocity gradient wh ich wou ld better repl icate some of the h igh amp l i t ude B r anch 3 arrivals that refract at the base of Layer 2C (e.g., F i gure 4.11b). 5.2.4 G e n e r a l C h a r a c t e r i s t i c s o f L a y e r 2 T h e layer ing determined for Layer 2 is pa r t i a l l y a n art i fact of the mode l l ing procedure and the resolut ion of the seismic waves used, but layer ing does exist i n this region as shown by the presence of reflectors (F igure 5.2) on the cross-ridge seismic reflection l ine (Rohr et ai, 1988; Roh r and Mi lkere i t . 1988). Reflections f rom w i t h i n Layer 2 have also been obta ined over the E P R (Herron et ai, 1978) and i n the A t l a n t i c (Musgrove and A u s t i n , 1988; M c C a r t h y et al, 1988). T h e reflection da ta of Herron et al. (1978) show three reflectors. T h e first is that a lready discussed (Herron, 1982) wh i ch is p robab ly the Layer 2 A - L a y e r 2B interface and the th i rd is the top of a possible ax ia l magma chamber (Herron et ai, 1980; Hale et ai, 1982), the Layer 2 C - L a y e r 3 interface; however, Chapter 5. DISCUSSION AND CONCLUSIONS 142 the second reflection is not discussed in the later studies and cou ld be the Layer 2 B -Layer 2C interface interpreted i n this da ta set. W h i t e (1979) and Spud i ch and Orcu t t (1980b) determined interna l s t ructure w i t h i n Layer 2 f r om reflections w i t h i n this layer, and Ma lacek and Clowes (1978) used the amp l i t ude behav iour of P-waves at close range to infer i n te rna l s t ructure w i t h i n Laye r 2. Clowes and K n i z e (1979) processed a mu l t i channe l reflection profi le in which layer ing can be seen w i t h i n the igneous crust of the J D F P l a te near the base of the cont inenta l slope; they interpreted reflections f r om the base of a 700-m-thick Layer 2A , f rom w i t h i n Layer 2B , and f rom the top of Layer 3. T h e shallowest interface closely mimics the ba thymet r y wh i le the deeper interfaces follow the ba thymet ry less closely. Th i s may be par t i a l l y caused by the decrease i n resolut ion of the seismic waves w i t h depth and increas ing velocity, but P u r d y (1982b), W h i t e and P u r d y (1983), W h i t e and W h i t m a r s h (1984), and W h i t m a r s h et al., (1986) also found that a l though the topography of i sovelocity layers w i t h i n Layer 2 decreases w i t h increas ing depth, they s t i l l reflect a subdued version of the seafloor relief. B r a t t and P u r d y (1984) determined that s t ruc tu ra l boundar ies w i t h i n shal low crust undu la te vert ica l ly by hundreds of metres over distances of tens of k i lometres; these results are s imi lar to the results of the refract ion da ta presented i n Chap te r 4 (Figures 4.21 and 4.23). Since these fluctuations i n the interfaces occur at the ridge crest, they must manifest var iab i l i t y i n the or ig ina l c rusta l accret ion process (B ra t t and Pu rdy , 1984). H igher velocities than n o r m a l i n Layer 2 have been obta ined along the ridge crest. Since this region of higher velocities was not required for the cross-ridge data , this zone of h igher velocities must be less than 2 k m wide. Such a narrow zone of higher veloci-ties impl ies that oceanic crust r ap id l y evolves to the observed off-ridge structure w i th i n 0.03 M a by processes of f ractur ing and hyd ro the rma l c i rcu lat ion ( A l t et ai, 1986). As exp la ined for Layer 2A, this veloc ity increase occurs under the less-fissured ax ia l flanks and may suggest a lack of pervasive crack ing throughout Layer 2 at the ridge crest. Chapter 5. DISCUSSION AND CONCLUSIONS 143 Tecton ic extension wh ich occurs as the l i thosphere s lowly moves away f rom the zone of accret ion wou ld cause the decrease i n velocity off-ridge. T h e ampl i tudes of refracted arr ivals f rom this zone of elevated velocities are extremely weak compared to those f rom r idge-paral le l l ines 10 k m off the ridge. T h e decrease i n veloc ity gradient required to reproduce these ampl i tudes could also be an effect of fewer fractures and lower porosity. T h e rough ridge-crest topograph} ' wou ld also decrease the ampl i tudes by scatter ing, but this effect was not model led. K o n g et al. (1988) also found higher velocities at an along-axis h igh at 26°N on the M A R ; they suggested this h igh-veloc i ty vo lume is a zone of hot, re lat ively uncracked mater i a l and may represent the most recent locus of magmat i c in t rus ion . T h i s exp lanat ion is on ly pa r t i a l l y va l id for the Endeavour R idge because K o n g et al. (1988) determined steep veloc ity gradients w i t h i n the crust i n contrast to the low veloc ity gradients determined f rom the S E I S R I D G data. T h e th i nn i ng of Layer 2 off-ridge has also been determined i n other surveys ( M c C l a i n et al., 1985; Rosendah l et al., 1976). A t the ridge crest, velocities are relat ively h igh, and then decrease because of f ractur ing. Beyond the ax i a l f lanks, Layer 2 th ins and increases i n velocity, b o t h effects caused by f i l l i ng of cracks by a l terat ion products. N o majo r a symmetry i n velocities or layer thicknesses exist on bo th sides of Endeavour R idge, a l though velocities i n Layers 2B and 2C are s l ight ly faster on the J D F P late. Th i s may s imp ly be caused by la tera l heterogeneities or, as discussed for Layer 2A , these velocit ies may be faster because of more rap id i n f i l l i ng of cracks caused by more sediment cover on the J D F P la te . 5.3 Layer 3 T h e interface between Layer 2C and 3 is a change i n vert ica l velocity gradient, a result also found i n other studies (Spudich and Orcut t , 1980b). Th i s change in velocity gradient Chapter 5. DISCUSSION AND CONCLUSIONS 144 m a y be the boundary between sheeted dike basalts and p lated and cumulate gabbros or a metamorph ic boundary de l ineat ing the base of water penetrat ion. Casey et al. (1981) believed this interface is caused by a complex interre lat ionsh ip between igneous and metamorph ic processes and that the resolut ion of a seismic exper iment does not differentiate between these processes. Reflections f rom this interface have been observed by Ha le et al. (1982) and M c C a r t h y et al. (1988). T h e velocity increase of Layer 3 away f r om the ridge is poor l y constra ined due to the larger errors i n da ta at further offsets, but this increase and the ve loc i ty decrease for arr ivals that have travel led under the ridge (F igure 4.25) may ind icate a region of lower ve loc i ty or lower gradient under the ridge at Layer 3 depths. H igher temperatures below the ax ia l ridge w i t h i n Layer 3 cou ld cause the veloc ity decrease: a l a te ra l temperature contrast of 400°C at depths of 3-5 k m below the seafloor wou ld account for a hor izonta l va r i a t i on in velocity of about 0.25 km/s (Toomey et al., 1988; Chr i s tensen. 1979), a ve loc i ty decrease comparab le to that used i n the refract ion mode l l i ng (7.0 km/s instead of 7.3 km/s). H i gh temperature hyd ro the rma l discharges ind icate that c rus ta l temperatures at the ridge crest are higher than those off-ridge. P u r d j ' and Detr ick (1986) determined a region of lowered Layer 3 velocities centred below a major along-axis h igh on the M A R . T h e y interpreted this region to be the remnants of the most recent phase of inject ion that has temporar i l y left beh ind it a region of elevated temperatures and pervasive cracking, resu lt ing in lowered velocities. T h e y suggested that a l terat ion products r ap id l y seal these cracks so that hydro therma l c i r cu la t ion is restr icted to Layer 2. reduc ing the Layer 3 porosit ies and p roduc ing t yp i ca l Layer 3 velocities off-ridge. Th i s mechan i sm could also exp l a i n the low velocity determined under Endeavour Ridge. Due to the short offsets of the data model led, no interface has been determined w i t h i n this layer comparable to Hor i zon R, suggested by M u t t e r and N A T Stud} ' G r oup (1985) to be the top of the layered cumulate sequence. Chapter.5. DISCUSSION AND CONCLUSIONS 145 5.4 Axial Reflector A short (<1 k m long) reflection at a two-way travel t ime of 1.0 s below the seafloor and d i rect ly below the ax ia l h igh is v is ible i n the reflection da ta of Roh r et al. (1988) and is shown i n F i gure 5.2. U s ing the refract ion velocities obta ined f rom the cross-ridge mode l l i ng , this reflector occurs at a depth of 1.8 k m below the seafloor or ~4 .0 k m below the sea surface. Th i s depth is shallower t han the Layer 2 C - L a y e r 3 interface but corresponds to the average depth of the L V Z obta ined i n the alternate mode l for the along-r idge l ine. R o h r et al. (1988) suggested two causes of this ax ia l reflector: a velocity contrast at the top of an ax ia l m a g m a chamber or the rma l effects i n conjunct ion w i t h the dep th of fau l t ing and hyd ro the rma l c i rcu lat ion. W h i t e and Clowes (1988) inverted the refract ion da ta us ing a tomograph ic technique and d id not ob ta in any large veloc ity anomalies at the depth of the ax ia l reflector; a m a g m a chamber w i t h a roof on ly 1 -km wide, however, cannot be resolved w i t h an inver-sion of refract ion data. T h e resolv ing power of the cross-ridge refraction da ta is such that pockets of melt up to 1 k m i n d iameter wou ld not be imaged; on ly a few rays wou ld refract through the areas of melt whi le man) ' wou ld refract a round i t . N o b road low-veloc i ty zone associated w i t h "hot rock" su r round ing a m a g m a chamber was observed (e.g., Ha rd ing et al., 1988b), further l im i t i ng the size of a hypo the t i ca l m a g m a chamber ( W h i t e and Clowes, 1988). M c C l a i n and Lewi s (1982) also fai led to find a magma chamber south of the C o b b Fracture Zone. B o t h of these studies suggest that crusta l accretion is episodic at low intermediate-spreading rates and that hyd ro the rma l c i rcu lat ion is r ap id and may extend to the base of the crust (L i ster, 1972). A n y m a g m a chamber present wou ld have to be l ike that proposed by Stakes et al. (1984) for the M A R — s e i s m i c a l l y undetectable and i n a wan ing phase. Submers ib le da ta also prov ide evidence that the m a g m a chamber beneath the Endeavour R idge is absent or in a quiescent phase: l ight ly-sedimented pi l low Chapter 5. DISCUSSION AND CONCLUSIONS 146 lavas and large v i s ib le sulfide deposits suggest a lack of recent erupt ions (Davis et ai, 1984), the elongate summi t depression has formed dur ing a per iod of tectonic extension w i t hou t extrusive volcanic ac t i v i t y ( K a p p e l and Ryan , 1986; T i v e y and Johnson, 1987) and the preponderance of p i l low lavas over more fluid sheet flows i n the neovolcanic zone is evidence against a recently replenished m a g m a chamber (Francheteau and Ba l l a r d , 1983; Johnson and Delaney, 1984). In many surveys (e.g., O r cu t t et al, 1975; 1976; 1984; Lewis and Garmany, 1982), fan shot da ta f rom profiles para l le l to the ridge crest t yp i ca l l y prov ide the strongest evidence for a m a g m a chamber. A separate s tudy w i l l invest igate fan shot da ta f rom this survey for signs of a magma chamber. A s an alternate exp lanat ion for the ax ia l reflector, Roh r et al. (1988) suggested that f au l t i ng w i th i n the neovolcanic zone has prov ided pathways for hyd ro the rma l c i rcu lat ion to cool the rock w i t h i n Layer 2 so that a h igh the rma l gradient and concomitant change i n ve loc i ty exist between Layer 2 and Layer 3, the latter sharp enough to reflect seismic, energy. N o evidence for a large change i n velocity at the ridge crest at the depth of the ax i a l reflector has been obta ined f rom mode l l i ng the cross-ridge l ine or the along-ridge l ine, bu t a velocity d i scont inu i ty on ly 1 k m i n la tera l extent at a depth below 2 k m cannot be resolved by the refract ion data. R o h r et al.'s (1988) exp lanat ion is supported by submers ib le and S e a M A R C da ta wh ich show large numbers of fractures w i t h i n the neovolcanic zone and few on the ax ia l flanks of the ridge. Th i s la tera l heterogeneity i n Assur ing wou ld create a narrow condu i t of fractures at depth and cause the hydro therma l c i r cu la t ion to be s imi la r l y conf ined. A n o t h e r poss ible exp lanat ion for the ax ia l reflector is petrologic. T h e t rans i t ion to crys ta l l ine cumulate gabbros, a l though model led as a velocity gradient, may be abrupt enough to cause a reflection at frequencies of 10 Hz. The lack of a reflector off-ridge i n this region cou ld be caused by d i s rupt ion of the Layer 2-La.yer 3 boundary th rough Assur ing or a decrease in velocity contrast caused b j ' metamorph ic a l terat ion of the rocks Chapter 5. DISCUSSION AND CONCLUSIONS 147 i n the t rans i t ion zone. Un l i k e the refraction in terpretat ion of W h i t e and Clowes (1988), no large asymmetr ic veloc ity anomal ies have been determined w i t h i n Layer 3. N o evidence for the d ipp ing intermediate reflector of Roh r et al. (1988) has been found i n the refract ion model l ing, but the depth of this reflector is below that of the depth of ray coverage on the cross-r idge l ine (F igure 5.2). Interpretat ion of the explos ion data and the double-a i rgun da ta at farther offsets w i l l determine i f this reflector has an analogue i n the refraction data. 5.5 Conclusions T h e shal low crusta l structure of the Endeavour R idge has been determined f rom ray-trace fo rward mode l l i ng of fifteen in- l ine O B S - a i r g u n refraction profiles para l le l , perpendicular, and ob l ique to the ridge. T h e general ized velocity structure of the crust consists of four layers (F igure 4.26): • Laye r 2A: ~0.40 k m thick; 2.6 km/s at top of layer, 2.8 km/s at b o t t o m • Layer 2B: ~0.65 k m thick; 4.8 km/s at top of layer, 5.4 km/s at bo t t om • Layer 2C : ~0.95 k m thick; 5.8 km/s at top of layer, 6.3 km/s at bo t t om • Layer 3 (A ) : ~2.65 k m thick; 6.5 km/s at top of layer, 7.3 km/s at bot tom. T h e layer ing is caused by large poros i ty contrasts due to the d i s t r ibut ion of fractures and vesicles and by changes in petrology. L a t e r a l heterogeneities on the scale of a. few kilometres (Figures 4.20, 4.22, 4.24, and 4.25) are super imposed on this basic velocity structure. Mos t of them are probab ly caused b} r l a tera l changes in porosity. Layer 2A thins and increases i n velocity by 0.1 km/s away f rom the ridge; ridge-parallel cracks create veloc ity anisotropy of ~ 10 -25% , w i th the Chapter 5. DISCUSSION AND CONCLUSIONS 148 faster direction parallel to the ridge. The velocities within Layers 2B and 2C also increase by 0.1 km/s away from the axial high. Increased Layer 2 velocities at the ridge crest reveal evidence of high lateral velocity contrasts in very young crust. Careful analysis of amplitude fluctuations will better determine changes in velocity gradients and their relationship to crustal evolution. Further work in this region should investigate the low-velocity uppermost layer. Lat-eral changes within Layer 2A need to be quantified and possibly related to fissure density. The bottom source and bottom receiver configuration of Purdy (1987) and Koelsch et al. (1986) would be an excellent technique to investigate this layer. The deeper crustal structure along the ridge crest should also be studied; a submersible would be needed to safely place the instruments in the rough topography of the neovolcanic zone. The lack of evidence for a large magma chamber from this seismic refraction stud}- is somewhat surprising. Although the bathymetric high and high temperature hot springs are evidence for a magma chamber, the lack of recent sheet flows in the neovolcanic zone and the presence of an elongate summit depression indicate that the magma chamber is waning. This refraction survey corroborates the evidence from the reflection data that any magma chamber is smaller than 1 km in width. Obtaining denser ray coverage beneath the ridge by increasing the shot spacing and number of OBS would not be suc-cessful in delineating a small magma chamber because of the inherent resolution of seismic wavelengths, although stacking of data from more closely-spaced shots would allow de-termination of travel times in regions where the present data set had noise problems. Careful analysis of the travel times and amplitudes of the fan shot data may resolve any additional velocity anomalies and the presence or absence of a low velocity zone beneath Endeavour Ridge. References A b r a m s , L.J., R.S. Detr ick and P.J. Fox, 1988. Morpho logy and crusta l s t ructure of the K a n e Fracture Zone, J. Geophys. Res., 93, 3195-3210. A d a m s o n , A . C , 1985. Basement l i thostrat igraphy, Deep Sea D r i l l i n g P ro jec t Ho le 504B, Init. Repts. Deep Sea Drilling Project, 83, 121-127. A l t . J . C , J . Honnorez, C. Laverne and R. E m m e r m a n , 1986. Hyd r o t he rma l a l terat ion of a 1 k m section through the upper oceanic crust, Deep Sea D r i l l i n g P ro ject Ho le 504B: Mineralogy, chemistry, and evo lut ion of seawater-basalt interact ions, J. Geophys. Res., 91, 10309-10335. Ander son , R.N., J . Honnorez, K. Becker, A . C A d a m s o n , J .C . A l t , R. E m m e r m a n n , P.D. K e m p t o n , H. K i no sh i t a , C. Laverne, M . J . M o t t l and R.L. Newmark , 1982. D S D P Ho le 504B, the first reference section over 1 k m through Layer 2 of the oceanic crust. Nature, 300, 589-594. Ander son , R.N., and R.L. Newmark , 1985. Use of geophysical logs for quant i ta t i ve determinat ion of f ractur ing, a l terat ion, and l i thos t ra t i g raphy i n the upper oceanic crust, Deep Sea D r i l l i n g P ro ject , Holes 504B and 556, Init. Repts. Deep Sea Drilling Project, 83, 443-478. A u , D., and R .M. Clowes, 1982. C ru s t a l s t ructure f rom an O B S survey of the Noo tka fau l t zone off western Canada , Geophys. J. R. Astr. Soc, 68, 27-48. A u , D., and R .M. Clowes, 1984. Shear-wave velocity s t ructure of the oceanic l i thosphere f r om ocean b o t t o m seismometer studies, Geophys. J. R. Astr. Soc, 77, 105-123. Baggeroer, A.B. , G.L. Duckwo r th , K . J . E l lefson and E.K. Scheer, 1986. T h e relat ive am-pl i tudes of p r ima ry and mu l t i p l e signals refracted i n the ocean crust, Ocean Seismo-Acoustics: Low Frequency Underwater Acoustics, NATO Conference Series IV, Ma-rine Sciences, 16, T . A k a l and J . M . Berkson, eds., P l e n u m Press, New York, 565-578. Becker, K., 1985. Large-scale e lectr ica l res ist iv ity and bu lk poros ity of the oceanic crust, Deep Sea D r i l l i n g Project Ho le 504B, Co s t a Fuca rift, Init. Repts. Deep Sea Drilling Project, 83, 419-427. 149 References 150 Becker, K., R.P. V o n Herzen, T . J . G . Francis, R.N. Ander son , J . Honnorez, A . C . A d a m -son, J .C . A l t , R. E m m e r m a n n , P.D. K e m p t o n , H. K i no sh i t a , C. Laverne, M . J . M o t t l and R.L. Newmark , 1982. In situ e lectr ical res ist iv i ty and bu lk poros i ty of the oceanic crust Co s t a R i c a R i f t , Nature, 300, 594-598. B r a t t , S.R., and G . M . Purdy , 1984. S t ructure and var iab i l i ty of oceanic crust on the flanks of the Eas t Pac i f ic R i se between 11° and 13°N, J. Geophys. Res., 89, 6111— 6125. B r a t t , S.R., and S.C. So lomon, 1984. Compres s iona l and shear wave s t ructure of the Eas t Pac i f i c R i se at 11° 20 'N : Const ra int s f rom three-component ocean b o t t o m seismometer da ta , J. Geophys. Res., 89, 6095-6110. Brocher , T . M . , and U.S. Ten B r i n k , 1987. Var ia t ions i n oceanic layer 2 elastic velocities near Hawa i i and their correlat ion to l i thospher ic flexure, J. Geophys. Res., 92, 2647-2661. B r y a n , W.B. , and J . G . Moore. 1977. Compo s i t i o na l var iat ions of young basalts i n the M i d - A t l a n t i c R idge rift valley near 36° 49 ' N , Geol. Soc. Am. Bull., 88, 556-570. Bu rne t t , M.S., J .A . O rcu t t and J.S. M c C l a i n , 1985. Further refract ion evidence for a crusta l m a g m a chamber, AGU Trans. EOS, 66, 1091. C a n n , J.R., 1974. A mode l for oceanic crusta l s t ructure developed, Geophys. J. R. Astr. Soc, 39, 169-187. Casey, J .F. , J . F . Dewey, P.J. Fox, J .A . Ka r s on and E. Rosencrantz, 1981. Heterogeneous nature of oceanic crust and upper mant le: A perspective f rom the Bay of Islands oph io l i te complex, The Sea, The Oceanic Lithosphere, 7, C. E m i l i a n i , ed.; A W i l e y -Interscience Pub l i c a t i on , J o h n W i l e y and Sons, New York, 305-338. Cerveny, V., I. Mo lo tkov , and I. Psenc ik , 1977. Ray Method in Seismology, Char les Un iver s i t y Press, Prague, Czechos lovakia, 214 pp. Chr i s tensen, N.L, 1979. Compress iona l wave velocities in rocks at h igh temperatures and pressures, c r i t i ca l thermal gradients, and crusta l low-velocity zones, J. Geophys. Res., 84, 6489-6857. Chr i s tensen, N.L, 1984a. T h e magn i tude, s ymmetry and or ig in of upper mant le anisotropy based on fabr ic analyses of u l t ramaf ic tectonites, Geophys. J. R. Astr. Soc, 76, 8 9 -111. Chr i s tensen, N.I., 1984b. Pore pressure and oceanic crustal seismic structure, Geophys. References 151 J. R. Astr. Soc, 79, 411-423. Chr i s tensen, N.I., and M . H . Sal isbury, 1985. Seismic velocities, densities, and porosities of layer 2B and layer 2C basalts f rom Hole 504B, Init. Repts. Deep Sea Drilling Project, 83, 367-370. Chr i s tensen, N.I., and J .D. Smewing, 1981. Geo logy and seismic st ructure of the northern section of the O m a n ophio l i te, J. Geophys. Res., 86, 2545-2555. C l a y t o n , R.W., and G.A. M c M e c h a n , 1981. Inversion of refraction da ta by wavefield cont inuat ion , Geophys., 46, 860-868. Clowes, R.M., and S. Kn i ze , 1979. C ru s t a l s t ructure from a mar ine seismic survey off the west coast of Canada , Can. J. Earth Sci., 16, 1265-1280. Crane, K., F. A i k m a n III, R. Embley , S. H a m m o n d , A . Malahoff , and J . L u p t o n , 1985. T h e d i s t r ibut ion of geothermal fields on the J u a n de Fuca R idge, J. Geophys. Res., 90, 727-744. Cud rak , C.F., D.J. W h i t e , and R .M. Clowes, 1987. A 3-D seismic refract ion survey of crusta l structure: Endeavour Segment, J u a n de Fuca R idge, AGU Trans. EOS, 68, 1371. Dav i s , E.E., and R.G. Cur r ie , 1985. Juan de Fuca Atlas: SeaMARC II Acoustic Image Mosaic, Northern Juan de Fuca Ridge, Open F i l e Repor t 85-2, E a r t h Phys ics B ranch , Energy, M ines and Resources, Sidney, B r i t i s h C o l u m b i a . Dav i s , E.E., and J .L . Ka r s ten , 1986. O n the cause of the asymmetr ic d i s t r ibut ion of seamounts about the J u a n de Fuca ridge: ridge-crest m ig ra t ion over a heterogeneous asthenosphere, Earth Planet. Sci. Lett., 79, 385-396. Dav i s , E.E., C.R.B. L ister, and B.T.R. Lewis, 1976. Seismic structure of the J u a n de F u c a ridge: ocean bo t t om seismometer results f rom the med ian valley, J. Geophys. Res., 81, 3541-3555. Dav i s , E.E., C.R.B. L i s ter, U.S. Wade and R.D. H y n d m a n , 1980. Deta i led heat flow measurements over the J u a n de Fuca ridge system, J. Geophys. Res., 85, 299-310. Dav i s , E.E., R.G. Cur r ie , B.S. Sawyer and D .M. Hussong, 1984a. Juan de Fuca Ridge Atlas: SeaMARC II Acoustic Imagery, Open F i l e 84-17, E a r t h Phys ics B r anch . Dav i s , E.E., R.G. Cur r ie , R.P. R i dd i hough , W . B . F . R y a n , K. Kastens, D.M. Hussong, S.R. H a m m o n d and A. Malahoff , 1984b. Geo log ica l mapp ing of the northern J u a n de References 152 Fuca and Exp lo re r Pudges us ing Sea M A R C I, Sea M A R C II and Seabeam acoustic imag ing, AGU Trans. EOS, 65, 1110. Dehler, S.A., 1986. A Seismic Refraction Study of the Queen Charlotte Fault Zone, M.Sc. thesis, Un iv . of B r i t i s h C o l u m b i a , Vancouver, 112 pp. Delaney, J.R., R.E. M c D u f f and J . E . L u p t o n , 1984. Hyd r o t he rma l f lu id temperature of 400°C on Endeavour Segment, Nor the rn J u a n de Fuca R idge, AGU Trans. EOS, 65, 973. Delaney, J.R., J . L . Ka r s ten and S.R. H a m m o n d , 1986. Petro logy and tectonics of the J u a n de Fuca R idge, A Symposium on the Juan de Fuca Ridge System, Programme and Abstracts, Geolog ical A s soc ia t ion of Canada , Pac i f i c Sect ion, Pac i f i c Geoscience Centre, 23-25. Det r i ck , R.S., and G . M . Purdy, 1980. T h e crusta l s t ructure of the K a n e fracture zone f rom seismic refract ion studies, J. Geophys. Res., 85, 3759-3778 Detr ick, R.S., P. B u h l , E. Vera, J . Mu t te r , J . O rcu t t , J . Mad sen and T. Brocher , 1987. Mu l t i - channe l seismic imag ing of a crusta l m a g m a chamber a long the Eas t Pac i f i c Rise, Nature, 326, 35-41. E a r t h Phys ics B r a n c h , 1984. Juan de Fuca Ridge Atlas: SeaMARC II acoustic imagery, O p e n F i l e Repo r t 84-1, Energy, Mines and Resources, O t t awa , Canada . E a t o n , R.M., D.E. Wel ls, and N. Stuifbergen, 1976. Satel l i te nav igat ion i n hydrography, International Hydrographic Review, 53, 99-116. E w i n g , J .L , and G . M . Pu rd j ' , 1982. Uppe r crusta l ve loc i ty s t ructure i n the R O S E area of the East Pac i f i c Rise, J. Geophys. Res., 87, 8397-8402. Fox, P.J., and J .B . Stroup, 1981. T h e p lu ton ic foundat ion of the oceanic crust, The Sea, The Oceanic Lithosphere, 7, C. E m i l i a n i , ed., A Wi lej r - Intersc ience Pub l i c a t i on , J o h n W i l e y and Sons, New York, 119-218. Francheteau, J . , and R.D. B a l l a r d , 1983. T h e Eas t Pac i f i c R ise near 21°N, 13°N and 20°S: inferences for along-strike var iab i l i t y of ax ia l processes of the M i d - O c e a n Ridge, Earth Planet. Sci. Lett, 64, 93-116. F r a n k l i n , J .M . , and E.S. K a p p e l , 1986. Compos i t i ona l characterist ics and vo lcanic sett ing of massive sulf ide deposits on J u a n de Fuca and Exp lo re r Ridges, A Symposium on the Juan de Fuca Ridge System, Programme and Abstracts, Geolog ica l A s soc ia t ion of Canada . Pac i f i c Section, Pac i f i c Geoscience Centre, 35-38. References 153 Get t ru s t , J.F., K. Fu rukawa and W . B . Kempner , 1982. Va r i a t i on i n young oceanic crust and upper mant le structure, J. Geophys. Res., 87, 8435-8445. G M A Compu te r Technology, L t d . , 1986. OBS Time Code Software, Program Documen-tation, Dept. of Geophys ics and Ast ronomy, Un iver s i ty of B r i t i s h Co l umb i a , 18 pp. G ran t , S.T., 1973. R h o - r h o L o r a n - C combined w i t h satell ite nav igat ion for offshore surveys, International Hydrographic Review, 50, 35-54. G ran t , S.T., 1980. B I O N A V — T h e Bedford Ins t i tute of Oceanography Integrated Nav iga-t ion System, Proceedings, U.S. National Ocean Survey 7th Annual Conference, Wash-ington, 19 pp. Hale, L.D., C . J . M o r t o n , and N.H. Sleep, 1982. Re interpretat ion of seismic reflection d a t a over the Eas t Pac i f i c R ise, J. Geophys. Res., 87, 7707-7717. Ha rd i ng , A . J . , J .A . O rcu t t and M . E . Kappu s , 1988a. Seismic st ructure of the East Pac i f i c R i se at 13°N, AGU Trans. EOS, 69, 304. Hard ing , A . J . , J .A . Orcut t , M . E . Kappu s , E.E. Vera , J .C . Mu t te r , P. B u h l , R.S. Detr ick, T . M . Brocher, 1988b. T h e s t ructure of young oceanic crust at 13° N on the E P R f rom E x p a n d i n g Spread Prof i les, J. Geophys. Res., ( submitted). Hefner, D.E., and D.L. Bar ret t , 1979. O B S development at Bedford Inst i tute of Oceanog-raphy, Mar. Geophys. Res., 4, 227-245. Her ron, T . J . , 1982. L a v a flow l a y e r — E a s t Pac i f i c Rise, Geophys. Res. Lett., 9, 17-20. Her ron, T . J . , W . J . L u d w i g , P.L. Stoffa, T . K . K a n , and P. B u h l , 1978. Structure of the East Pac i f i c R i se crest f r om mu l t i channe l seismic reflection data , J. Geophys. Res., 83, 798-804. Her ron, T . J . , P.L. Stoffa, and P. B u h l , 1980. M a g m a chamber and mant le r e f l e c t i on s— Eas t Pac i f i c Rise, Geophys. Res. Lett., 7, 989-992. Hey, R.N., and D.S. W i l s o n , 1982. P ropaga t i ng rift exp lanat ion for the tectonic evolut ion of the N E P a c i f i c — T h e pseudomovie, Earth Planet. Sci.'Lett., 58, 147-188. He}-, R.N., F.K. Duennebier and W . J . Mo r gan , 1980. P ropagat ing rifts on mid-ocean ridges, /. Geophys. Res., 85, 3547-3658. Houtz , R.E., 1976. Seismic properties of layer 2 A in the Pac i f ic , J. Geophys. Res., 81, 6321-6331. References 154 Hou tz , R., and J . Ew i ng , 1976. U p p e r crusta l structure as a funct ion of p late age, J. Geophys. Res., 81, 2490-2498. H y n d m a n , R.D., and M J . Drury , 1976. T h e phys ica l properties of oceanic basement rocks f rom deep dr i l l i ng on the M i d - A t l a n t i c R idge, J. Geophys. Res., 81, 4042-4060. H y n d m a n , R.D., and M . H . Sa l i sbmy, 1984. T h e phys ica l nature of young oceanic crust on the M i d - A t l a n t i c R idge, Deep Sea D r i l l i n g Pro ject Ho le 395A, Init. Repts. Deep Sea Drilling Project, 78B, 839-848. Jacobson, R.S., and B.T.R. Lewis, 1988. A t tenua t i on results f rom the uppermost young oceanic crust, AGU Trans. EOS, 69, 406. Jacobson, R.S., G . G . Shor, Jr. , and M . Bee, 1984. A compar i son of velocity and attenua-t ion between the N icobar and Benga l deep sea fans, J. Geophys. Res., 89, 6181-6196. Jacobson, R.S., L.D. Bibee, R.W. Emb ley and S.R. H a m m o n d , 1987. A microse i smic i ty survey of A x i a l Seamount, J u a n de Fuca R idge, Bull. Seism. Soc. Am., 77, 160-172. Johnson , H.P., J . Ka r s ten , J.R. Delaney, E.E. Dav i s , R.G. Cu r r i e and R.L. Chase, 1983. A detai led s tudy of the C o b b offset of the J u a n de Fuca Ridge: E vo l u t i on of a propagat ing r i f t , J. Geophys. Res., 88, 2297-2315. Johnson , H.P., M .A . Tivey, M . Botros and J . E . Par i so, 1986. Magnet i c and tectonic s t ructure of the J u a n de Fuca region, A Symposium on the Juan de Fuca Ridge System, Programme and Abstracts, Geo log ica l As soc ia t ion of Canada , Pac i f i c Sect ion, Pac i f i c Geoscience Centre, 16-17. K a p p e l , E.S., and W.B . F . R y a n , 1986. Vo l can i c episodic i ty and a non-steady state rift va l le j ' a long northeast Pac i f i c spreading centers: Ev idence f rom Sea M A R C I, J. Geophys. Res., 91, 13925-13940. Ka r s ten , J.L., S.R. H a m m o n d , E.E. Dav i s and R.G. Cur r ie , 1986. Deta i led geomorphol -ogy and neotectonics of the Endeavour segment; J u a n de Fuca Ridge: New results f rom Seabeam swath mapp ing , Geol. Soc. Am. Bull., 97, 213-221. Koe l sch , D.E., W . E . W i t z e l l , Sr., J . E . B r oda , F.B. W o o d i n g and G . M . Pu rdy , 1986. A deep towed explosive source for seismic exper iments on the ocean floor, Mar. Geophys. Res., 8, 345-361. K o n g , L.S.L., S.C. Solomon and G . M . Pu rdy , 1988. Mechan i ca l and s t ructura l hetero-geneity a long the M i d - A t l a n t i c R idge at 26 N: Ev idence for recent magmat ic in t rus ion, AGU Trans. EOS, 69, 304. References 155 K u z n i r . N.J., 1980. T h e r m a l evolut ion of the oceanic crust, its dependence on spreading rate and effect on crusta l structure, Geophys. J. R. Astr. Soc, 61, 167-181. Lewis , B.T.R., and J .D . Garmany , 1982. Const ra int s on the structure of the East Pac i f i c R i se f rom seismic data, J. Geophys. Res., 87, 8417-8425. Lewis , B.T.R., and W . E . Syndsman, 1977. Ev idence for a low veloc ity layer at the base of the oceanic crust, Nature, 266, 340-344. L ister, C.R.B., 1972. O n the t he rma l balance of a mid-ocean ridge, Geophys. J. R. Astr. Soc, 26, 515-535. L ister, C.R.B., 1974. O n the penetrat ion of water into hot rock, Geophys. J. R. Astr. Soc, 39, 465-509. L i s ter, C.R.B., 1982. " A c t i v e " and "pass ive" hyd ro the rma l systems i n the oceanic crust: P red i c ted phys ica l condit ions, The Dynamic Environment of the Ocean Floor, K A . Fann i ng and F.T. M a n h e i m , eds., Lex ing ton Books, D.C. Heath and Company , Lex-ing ton , Mass., 441-459. L i s ter, C.R.B., 1983. O n the in te rmi t tency and crys ta l l i zat ion mechanisms of sub-seafloor m a g m a chambers, Geophys. J. R. Astr. Soc, 73, 351-365. Lonsda le, P., 1983. Over l app ing rift zones at the 5.5°S offset of the East Pac i f i c Rise, J. Geophys. Res., 88, 9393-9406. Lonsda le, P., 1985. Non- t rans fo rm offsets of the Pac i f i c -Cocos plate boundary and their trace on the rise f lank, Geol. Soc. Am. Bull., 96, 313-327. L u d w i g , W . J . , J . E . Nafe and C L . Drake, 1970. Seismic refract ion, The Sea, 4 ( 1 ) , A . E . M a x w e l l , ed., Wi ley- Intersc ience New York, 53-84. M a c d o n a l d , K . C , 1982. M id -ocean ridges: F ine-scale tectonic, volcanic and hydrother-m a l processes w i t h i n the p late boundary zone, Ann. Rev. Earth Planet. Sci., 10, 155-190. M a c d o n a l d , K . C , 1983. A geophysical compar i son between fast and slow spreading centers: Const ra ints on m a g m a chamber format ion and hydro the rma l act iv i ty, Hy-drothermal Processes at Seafloor Spreading Centers, NATO Conference Series IV, Marine Sciences, 12, P.A Rona , K. Bo s t r om, L. Laub ie r and K.L. Sm i th , Jr. , eds., P l e n u m Press, New York, 27-51. References 156 M a c d o n a l d , K.C., 1986. T h e crest of the M i d - A t l a n t i c R idge: Mode l s for crusta l gen-erat ion processes and tectonics, The Geology of North America, The Western North Atlantic Region, M, P.R. Vogt and B.E. Tu lcho lke, eds., Geolog ica l Society of Amer i c a , Bou lder , Co., 51-68. M a c d o n a l d , K.C., and P.J. Fox, 1983. Ove r l app ing spreading centers: N e w accretion geometry on the East Pac i f ic Rise, Nature, 302, 55-57. M a c d o n a l d , K.C., and B.P. Luyendyk , 1985. Invest igation of fau l t ing and abyssal h i l l f o rmat ion on the flanks of the East Pac i f i c R i se (21°N) us ing A L V I N , Mar. Geophys. Res., 7, 515-535. M a c d o n a l d , K., J . - C . Sempere and P.J. Fox, 1984. East Pac i f i c R i se f rom Siqueiros to Orozco fracture zones: A l o n g strike cont inu i ty of ax ia l neovolcanic zone and structure and evo lut ion of over lapping spreading centers, Geophys. Res., 89, 6049-6069. Malecek, S.J., and R.M. Clowes, 1978. C r u s t a l s t ructure near Exp lo re r R idge f rom a mar ine deep seismic sounding survey, J. Geophys. Res., 83, 5899-5912. M c C a r t h y , J . , J . C . Mut te r , J . L . Mo r t on , N.H. Sleep and G.A. T h o m p s o n , 1988. Re l i c m a g m a chamber structures preserved w i t h i n the Mesozo ic N o r t h A t l a n t i c crust?, Geol. Soc. Am. Bull, 100, 1423-1436. M c C l a i n , J.S., and B.T.R. Lewis, 1982. Geophys i ca l evidence for the absence of a crusta l m a g m a chamber under the nor thern J u a n de Fuca Ridge: A contrast w i t h R O S E results, J. Geophys. Res., 87, 8477-8489. M c C l a i n , J.S., J . A . Orcut t and M . Bu rnet t , 1985. T h e East Pac i f i c Rise i n cross-section: A seismic mode l , J. Geophys. Res., 90, 8627-8639. M c M a n u s , D.A., M .L . Holmes, B. Carson, and S.M. Ba r r , 1972. La te Quaternary tecton-ics, nor thern end of J u a n de F u c a R idge (northeast Pac i f ic ) , Mar. Geol, 12,. 141-164. Moores, E.M.. J.S. M c C l a i n and R.J. Varga, 1986. Detachment fau l t ing at oceanic spreading centers: Ev idence f r om Troodos and oceanic crust. Geol. Soc. Am. Abstracts, Annual Meeting Cordilleran Section, 160. Moos , D., and M .D . Zoback, 1983. In s i tu studies of velocity i n fractured crysta l l ine rock, J. Geophys. Res., 88, 2345-2358. Moos , D., D. Go ldbe rg , M .A . Hoba r t and R.N. Ander son , 1986. E la s t i c wave velocities i n Layer 2A f rom fu l l waveform sonic logs at Ho le 504B, Init. Repts. Deep Sea Drilling Project, 92, 563-570. References 157 M o r t o n , J.L., and N.H. Sleep, 1985a. A mid-ocean ridge the rma l model: Const ra ints on the vo lume of ax ia l hydro therma l heat f lux, J. Geophys. Res., 90, 11345-11353. M o r t o n , J.L., and N.H. Sleep, 1985b. Seismic reflections f r om a L a u Ba s i n m a g m a chamber, Geology and Offshore Resources of Pacific Island Arcs-Tonga Region, D.W. Schol l and T . L . Val l ier, eds., C i r c um-Pac i f i c C o u n c i l for Energy and M i ne r a l Resources E a r t h Science Series, 2, Hous ton, Texas, 441-453. M o r t o n , J.L., N.H. Sleep, W .R . No rmark and D.H. Tompk in s , 1987. St ructure of the southern J u a n de Fuca R idge f rom seismic reflection records, J. Geophys. Res., 92, 11315-11326. Musgrove, L.A., and J .A . A u s t i n , Jr. , 1983. Intrabasement st ructure i n the southern A n g o l a Ba s i n , Geology, 11, 169-173. M u t t e r , J . C , and No r th A t l a n t i c Transect ( N A T ) Study G r o u p , 1985. Mu l t i c hanne l seismic images of the oceanic crust 's i n te rna l structure: Ev idence for a m a g m a chamber beneath the Mesozo ic M i d - A t l a n t i c R idge, Geology, 13, 629-632. N A T S tudy G r o u p , 1985. N o r t h A t l a n t i c Transect: A wide-aperture, two-ship m u l t i -channel seismic invest igat ion of the oceanic crust, J. Geophys. Res., 90, 10321-10341. Newmark , R.L., R.N. Ander son , D. Moos and M .D . Zoback, 1985a. Sonic and u l t rason ic logging of Hole 504B and its imp l i cat ions for the structure, porosity, and stress regime of the upper 1 k m of the oceanic crust, Init. Repts. Deep Sea Drilling Project, 83, 479-510. Newmark , R.L., R.N. Anderson, D. Moos and M .D . Zoback, 1985b. Structure, poros i ty and stress regime of the upper oceanic crust: Sonic and ultrasonic logging of D S D P Ho le 504B, Tectonophysics, 118, 1-42. N i sbet , E.G., and C.M.R. Fowler, 1978. T h e M i d - A t l a n t i c R idge at 37° and 45°N: some geophys ical and petro log ica l constraints, Geophys. J. R. Astr. Soc, 54, 631-660. O rcu t t , J .A., 1987. Structure of the E a r t h : Oceanic crust and uppermost mant le, Rev. Geop., 25, 1177-1196. O rcu t t , J . , B. Kennet t , L. D o r m a n , and W . Prothero , 1975. A low velocity zone under-l y i n g a fast spreading rise crest, Nature, 256, 475-476. O r cu t t , J .A., B .L .N. Kennet t , and L .M. D o r m a n , 1976. S t ructure of the East Pac i f i c Rise f rom an ocean bo t t om seismometer survey, Geophys. J. R. Astr. Soc, 45, 305-320. References 158 Orcu t t , J.A., J.S. M c C l a i n and M . Bu rnet t , 1984. E vo l u t i o n of the ocean crust: results f r om recent seismic experiments, Ophiolites and Oceanic Lithosphere, Geol. Soc. Spe-cial Publication No. 13, I.G. Gass, S.J. L i p p a r d and A . W . Shelton, eds., Geo l . S o c , B l ackwe l l Scientif ic Pub l i ca t i ons , Ox fo rd , 7-16. Pa l l i s ter , J.S., and C A . Hopson, 1981. S ama i l oph io l i te p lu ton i c suite: F i e ld relations, phase var iat ion, c rypt ic var iat ion and layer ing, and a mode l of a spreading ridge m a g m a chamber, J. Geophys. Res.. 86, 2593-2644. P i ke , C . J . , 1986. A Seismic Refraction Study of the Hecate Sub-basin, British Columbia, M.Sc. thesis, Un i v . of B r i t i s h C o l u m b i a , 133 pp. Poehls , K.A., 1974. Seismic refraction on the M i d - A t l a n t i c R idge at 37°N, J. Geophys. Res., 79, 3370-3373. Pu rdy , G.M. , 1982a. T h e correction for the t ravel t ime effects of seafloor topography i n the interpretat ion of mar ine seismic data, J. Geophys. Res., 87, 8389-8396. Pu rdy , G.M. , 1982b. T h e var iab i l i ty i n seismic s t ructure of Layer 2 near the East Pac i f i c R i se at 12°N, J. Geophys. Res., 87, 8403-8416. Pu rdy , G.M. , 1987. New observations of the shal low seismic structure of young oceanic crust, J. Geophys. Res., 92, 9351-9362. Pu rdy , G.M. , and R.S. Detr ick, 1986. C r u s t a l s t ructure of the M i d - A t l a n t i c R idge at 23°N f rom seismic refract ion studies, J. Geophys. Res., 91, 3739-3762. Pu rdy , G.M. , and J . E w i n g , 1986. Seismic s t ructure of the ocean crust, The Geology of North A merica, The Western North Atlantic Region, M , P.R. Vogt and B.E. Tulcholke, eds., Geo l . Soc. A m . , Bou lder , Co., 313-330. Raff, A.D., and R.G. Ma son , 1961. Magnet i c survey off the west coast of N o r t h Ame r i c a , 4 0°N to 52°N la t i tude, Geol. Soc. Am. Bull, 82, 1267-1270. Re id . I., J .A . O r cu t t and W . A . Prothero, 1977. Seismic evidence for a zone of par t i a l me l t i ng under ly ing the East Pac i f i c Rise at 21°N, Geol. Soc. Am. Bull., 88, 678-682. R i dd i hough , R., 1984. Recent movements of the J u a n de Fuca plate system, J. Geophys. Res., 89, 6980-6994. Rohr . K .M.M. . and B. Mi lkere i t , 1988. A shal low crusta l reflector across the J u a n de Fuca ridge, J. Geophys. Res., submit ted. Rohr . K .M.M. . B. M i l kere i t and C . J . Yo r a th , 1988. As } r mmetr ic deep crusta l structure References 159 across the J u a n de Fuca R idge, Geology, 1 6 , 533-537. Rosendah l , B.R., R .W. Ra i t t , L .M . D o r m a n , L.D. Bibee, D .M. Hussong, and G .H . Sut-t on , 1976. E vo l u t i o n of oceanic crust, 1, A phys ica l mode l of the East Pac i f i c R i se crest f rom seismic refraction da ta , J. Geophys. Res., 8 1 , 5294-5304. R y a n , W.B.F. , 1986. Features and properties of the Endeavour and Exp lo re r Ridges i n common w i t h the M i d - A t l a n t i c R idge and East Pac i f i c Rise, A Symposium on the Juan de Fuca Ridge System, Programme and Abstracts, Geo log ica l Assoc iat ion of Canada , Pac i f i c Sect ion, Pac i f i c Geoscience Centre, 21-22. Sal isbury, M .H. , and N.I. Chr i s tensen, 1978. T h e seismic veloc ity structure of a traverse th rough the B a y of Islands ophio l i te complex, Newfound land , an exposure of oceanic crust and upper mant le, J. Geophys. Res., 8 3 , 805-817. Sal isbury, M .H. , N.I. Chr i stensen, K. Becker and D. Moos, 1985. T h e velocity s t ructure of layer 2 at Deep Sea Dr i l l i ng P ro ject Site 504 f rom logging and laboratory exper iments, Init. Repts. Deep Sea Drilling Project, 8 3 , 529-539. Sempere, J . - C , and K.C. M a c d o n a l d , 1987. Ma r i ne tectonics: Processes at mid-ocean ridges, Rev. Geophys., 25, 1313-1347. Shearer, P., and J . Orcut t , 1985. Anisotrop} ' i n the oceanic l i t ho sphe re—theo r y and observations f r om the Ngendei seismic refract ion exper iment i n the south-west Pac i f ic , Geophys. J. R. Astr. Soc, 8 0 , 493-526. Shearer, P.M., and J .A . Orcut t , 1986. Compress iona l and shear wave an i sotropy in the oceanic l i t ho sphe re—the Ngende i seismic refract ion exper iment, Geophys. J. R. Astr. Soc, 8 7 , 967-1003. Sleep, N.H., 1975. Fo rmat ion of oceanic crust: Some the rma l constraints, J. Geophys. Res., 8 0 , 4037-4042. Sleep, N.H., 1978. T h e r m a l s t ructure and kinematics of mid-ocean ridge ax i s - some impl i cat ions to basalt ic vo lcan i sm, Geophys. Res. Lett., 5, 426-428. Sleep, N.H., 1983. Hyd ro the rma l convect ion at ridge axes, Hydrothermal Processes at Seafloor Spreading Centers, NATO Conference Series, IV, Marine Sciences, 1 2 . P.A. Rona , K. Bo s t r om, L. Laub ie r and K .L . Sm i th , eds.. P l e n u m Press, New York , 71-82. Sleep, N.H., and B.R. Rosendahl , 1979. Topography and tectonics of mid-ocean ridge axes, J. Geophys. Res., 8 4 , 6831-6839. References 160 Smewing, J.D., 1981. M i x i n g characterist ics and compos i t iona l differences i n mant le-derived melts beneath spreading axes: Ev idence f rom cyc l ica l ly layered rocks i n the ophio l i te of north O m a n , J. Geophys. Res., 8 6 , 2645-2659. Spud ich , P., and J . Orcut t , 1980a. A new look at the seismic veloc ity s t ructure of the oceanic crust, Rev. Geop. Space Phys., 1 8 , 627-645. Spud ich , P., and J . Orcut t , 1980b. Petro logy and porosity of an oceanic crusta l site: Results f rom waveform mode l ing of seismic refract ion data, J. Geophys. Res., 8 5 , 1409-1433. Stakes, D.S., J . W . Shervais and C A . Hopson, 1984. T h e volcan ic- tecton ic cycle of the F A M O U S and A M A R valleys, M i d - A t l a n t i c R idge (36° 47'N): E v i dence f rom basalt glass and phenocryst compos i t i ona l var iat ions for a steady state m a g m a chamber beneath the val ley midsect ions, A M A R 3, J. Geophys. Res., 8 9 , 6995-7028. Stephen, R.A., 1985. Seismic anisotropy i n the upper oceanic crust, J. Geophys. Res., 90,11383 -11396 . Stephen, R.A., 1986. Seismic anisotropy i n the upper crust at D S D P Site 504B, Ocean Seismo-Acoustics: Low Frequency Underwater Acoustics, N A T O Conference Series IV , Ma r i ne Sciences, 1 6 , T . A k a l and J . M . Berkson, eds., P l e n u m Press, New York , 599-607. Stephen, R.A., 1988. La te ra l heterogeneity i n the upper oceanic crust at Deep Sea D r i l l i n g Project Site 504, J. Geophys. Res., 9 3 , 6571-6584. Stoffa, P.L., P. B u h l , T . J . Her ron, T . K . K a n , and W . J . L udw i g , 1980. M a n t l e reflections beneath the crestal zone of the East Pac i f i c R i se f rom mul t i - channe l seismic d a t a , Mar. Geol, 3 5 , 83-97. Taba ta , S., and J . Peart . 1985. Canadian Data Report of Hydrography and Ocean Sci-ences, 3 8 , 411-414. Tive5', M.K.. and J.R. Delaney, 1986. G r o w t h of large sulfide structures on the Endeavour Segment of the J u a n de Fuca R idge, Earth Planet. Sci. Lett., 77, 303—317. T ivey, M.A., and H.P. Johnson , 1987. T h e centra l anomaly magnet ic h igh: Impl icat ions for ocean crust construct ion and evo lut ion, J. Geophys. Res., 9 2 , 12685-12694. Toomey, D.R., S.C. Solomon and G . M . Pu rdy , 1988. M icroearthquakes beneath median val ley of M i d - A t l a n t i c R idge near 23°N: Tomography and tectonics, /. Geophys. Res., 9 3 , 9093-9112. References 161 W h i t e , D.J., 1988. Two-d imens iona l seismic refract ion tomography, Geophys. J., sub-mi t ted . W r hite, D.J., and R.M. Clowes, 1988. Shal low crusta l structure beneath the J u a n de Fuca R idge f rom seismic tomography, Geophys. J., submi t ted . W h i t e , R.S., 1979. Oceanic upper crusta l s t ructure f rom var iable angle seismic reflection-refraction profiles, Geophys. J. R. Astr. Soc, 57, 683-726. W h i t e , R.S., 1984. A t l a n t i c oceanic crust: seismic st ructure of a s low-spreading ridge, Ophiolites and Oceanic Lithosphere, Geol . Soc. Spec ia l P u b l i c a t i o n 13, I.G. Gass, S.J. L i p p a r d and A . W . Shelton, eds., Geo l . S o c , B l a c kwe l l Scientif ic Pub l i ca t i on s , Ox fo rd , 101-111. W h i t e , R.S., R.S. Detr ick, M . C . S i nha and M . H . Cormier , 1984. Anoma lou s seismic crusta l s tructure of oceanic fracture zones, Geophys. J. R. Astr. Soc, 79, 779-798. W h i t e , R.S., and G . M . Purdy , 1983. C ru s t a l veloc ity s t ructure on the flanks of the M i d - A t l a n t i c R idge at 24°N, Geophys. J. R. Astr. Soc, 75, 361-385. W r hite. R.S., and R.B. W h i t m a r s h , 1984. A n invest igat ion of seismic an isotropy due to cracks i n the upper oceanic crust at 45°N, M i d - A t l a n t i c R idge, Geophys. J. R. Astr. Soc, 79, 439-467. WTi i tmarsh, R.B., 1978. Seismic refract ion studies of the upper igneous crust i n the N o r t h A t l a n t i c and poros i ty estimates for Layer 2, Earth Planet. Sci. Lett., 37, 451-464. WTi i tmarsh, R.B., J .A . Orcut t , T . H . J o rdan , R.G. A d a i r and P .M. Shearer, 1986. Ve loc i ty bounds on the seismic structure of Mesozo ic crust and upper mant le f rom downhole observations at Deep Sea D r i l l i n g Project Ho le 595B, Init. Repts. Deep Sea Drilling Project, 91, 437-444. W i l s o n , D.S., R.N. Hey and C. N i sh imura , 1984. P ropaga t i on as a mechan i sm of reori-entat ion of the J u a n de Fuca R idge, J. Geophys. Res., 89, 9215-9227. Yo r a th , C .J . , R .M. Clowes, R.D. M a c D o n a l d , C. Spencer, E.E. Dav i s , R.D. H y n d m a n , K. Rohr , J . F . Sweeney, R.G. Cur r ie , J . F . Ha lpenny and D.A. Seeman, 1988. Marine Multichannel Seismic Reflection, Gravity and Magnetic Profiles—Vancouver Island Continental Margin and Juan de Fuca Ridge, O p e n F i l e Repor t 1661, Geo log ica l Sur-vey of Canada . Zelt, C.A., and R.M. E l l i s , 1988a. P r a c t i c a l and efficient ray t rac ing in two-d imens iona l med i a for rap id travelt ime and amp l i tude fo rward mode l l ing , Can. J. Expl. Geophys., References 162 24, 16-31. Zelt, C.A., and R .M. E l l i s , 1988b. Seismic st ructure of the crust and upper mant le i n the Peace R i ve r A r c h Region, Canada , J. Geophys. Res., submi t ted . Zoback, M.D., and R.N. Anderson, 1982. U l t ra son ic borehole televiewer invest igat ion of oceanic crusta l layer 2A, Co s t a R i c a R i f t , Nature, 295, 375-379. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0052628/manifest

Comment

Related Items