Citation
Geochemistry of Eastern Pacific Morb : implications for morb petrogenesis and the nature of crustal accretion within the neovolcanic zone of two recently active ridge segments

Material Information

Title:
Geochemistry of Eastern Pacific Morb : implications for morb petrogenesis and the nature of crustal accretion within the neovolcanic zone of two recently active ridge segments
Creator:
Smith, Matthew C
Publication Date:
Language:
English
Physical Description:
vii, 164 leaves : ill. ; 29 cm.

Subjects

Subjects / Keywords:
Basalt ( jstor )
Crystallization ( jstor )
Lava ( jstor )
Mid ocean ridges ( jstor )
Ocean floor ( jstor )
Ridges ( jstor )
Seamounts ( jstor )
Submersibles ( jstor )
Volcanology ( jstor )
Waxes ( jstor )
Dissertations, Academic -- Geological Sciences -- UF ( lcsh )
Geological Sciences thesis, Ph. D ( lcsh )
City of Gainesville ( local )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1999.
Bibliography:
Includes bibliographical references (leaves 153-163).
General Note:
Printout.
General Note:
Vita.
Statement of Responsibility:
by Matthew C. Smith.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. This item may be protected by copyright but is made available here under a claim of fair use (17 U.S.C. §107) for non-profit research and educational purposes. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
021577808 ( ALEPH )
43707752 ( OCLC )

Downloads

This item has the following downloads:


Full Text











GEOCHEMISTRY OF EASTERN PACIFIC MORB:
IMPLICATIONS FOR MORB PETROGENESIS AND THE NATURE OF CRUSTAL ACCRETION WITHIN THE NEOVOLCANIC ZONE OF TWO RECENTLY ACTIVE RIDGE SEGMENTS










By

MATTHEW C. SMITH


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


1999















ACKNOWLEDGMENTS


There are many that have contributed to this research and deserve

acknowledgement. I would like to thank Dr. Michael Perfit for helping to put me in a position to take advantage of numerous opportunities to participate on the fascinating oceanographic cruises that have led to this research. Without his contacts, guidance, mentorship, and friendship, none of this would have been possible. Drs. Robert Embley and Daniel Fornari served as chief scientists on many of the cruises to the Juan de Fuca and East Pacific Rise respectively. Their example of how to be an effective sea going scientist made a lasting impression on me, and this research benefited from many discussions of the structure and volcanic morphology of eastern Pacific spreading centers. The crews of the ALVIN, Atlantis II, ROPOS and Discoverer deserve thanks for their willingness to pursue the excellence we required, even in the wee hours of the morning. Dr. William Chadwick graciously prepared the sample location maps in chapter 2 and provided AMS -60 data to aid in interpretation of regional geologic associations. H. Paul Johnson generously donated splits of basalt samples collected during three cruises to the CoAxial Segment. Dr. Ann Heatherington assisted greatly in the isotopic analyses and I thank her. Drs. Tim O'Hearn and W. I. Ridley analyzed data from the Smithsonian and USGS respectively, and Dr. Ian Jonasson provided ICP data from the GSC at no cost to this project. I owe them all a debt of gratitude. Dr. Karen Von Damm, Dr. John Lupton and Dr. Marvin Lilley graciously provided unpublished data for vent fluids discussed in chapter 3. D. John Chadwick also graciously provided his unpublished data for basalts









from Axial Seamount south rift zone used in chapter 2 discussions. Lastly, I would like to thank Robin Milam for her assistance in the formatting and preparation of this manuscript, and my family for their continual love and support.















TABLE OF CONTENTS
pAge


A CKN OW LED G M ENTS ............................................................................................... ii

A B S T R A C T ....................................................................................................................... vi

CHAPTERS

1. GEN ERA L INTRO D U CTION ..................................................................................... 1

1.1 The C oA xial Segm ent ........................................................................................ 7
1.2 The East Pacific Rise between 9 and 10�N ........................................................ 8

2. GEOCHEMICAL CHARACTERISTICS AND PETROGENETIC
RELATIONSHIPS OF RECENTLY ERUPTED MORB FROM THE
CENTRAL JUAN DE FUCA RIDGE: SEGMENT SCALE,
INTERFLOW AND INTRAFLOW VARIATIONS ....................................... 11

2.1. Introduction ................................................................................................... . . 11
2.1.1 Field Investigations and Previous Studies ........................................ 11
2.1.2 R egional G eology ............................................................................... 12
2.2. Data Collection and Sample Distribution ........................................................ 16
2.2.1 Alvin and ROV Based Sample Acquisition ..................................... 16
2.2.2 Surface Based Sample Acquisition ................................................... 16
2.2.3 Sidescan Sonar and Near Bottom Photographic Data Acquisition ....... 20
2.3. A nalytical M ethods ....................................................................................... 20
2.3.1 Sam ple Preparation ............................................................. . ..... 20
2.3.2 M ajor Elem ent Analyses ................................................................... 22
2.3.3 Trace Elem ent Analyses ................................................................... 26
2.3.4. Isotopic A nalyses ............................................................................ 33
2.4. Regional CoAxial Segment Geochemical Characteristics .............................. 33
2.4.1 M ajor Elem ent Trends ...................................................................... 33
2.4.2 Isotopic and Trace Element Signals ................................................. 40
2.5. Discrimination of Discrete Flow Units ............................................................ 46
2.5.1 Visual and Bathymetric Flow Discrimination ................................... 46
2.5.2 Geochemical Flow Discrimination .................................................... 48
2.6. Distinction of Regional Petrochemical Provinces .......................................... 61
2.6.1 Axial Seam ount Province ................................................................. 62
2.6.2 The Western Fault Block Ridge ........................................................ 70
2.6.3 Sm all N RZ Volcanoes ...................................................................... 71









2.6.4 The CoAxial Segment ........................................................................ 73
2.7. D iscussion ..................................................................................................... . . 74
2.7.1 Recent CoAxial Segment Volcanism ............................................... 74
2.7.2 Intra-flow Geochemical Variability ................................................. 78
2.7.3 Regional Petrochemical Characteristics ............................................ 81
2.7.4 Regional Mantle Heterogeneity and the Origins of CoAxial
Segment Geochemical Depletion ...................................................... 85
2.8. C onclusions ..................................................................................................... 87

3. SUBMARINE INVESTIGATIONS OF A THIRD-ORDER OSC A 9'37' N:
ESTABLISHING A CAUSE AND EFFECT RELATIONSHIP BETWEEN OSC PROPAGATION AND MAGMATIC ACTIVITY .......................................... 89

3.1. Introduction ................................................................................................... . . 89
3.2. Previous Studies and Regional Geology ....................................................... 91
3.2.1 Previous Studies .............................................................................. 91
3.2.2 General Second-Order Scale Observations ....................................... 99
3.2.3 Previous Geologic Observations of the EPR Axis Between 90
49.9'N and 9' 37.1'N ......................................................................... 102
3.3. Data Acquisition and Alvin Observations ......................................................... 104
3.3.1 In Situ Geologic Observations ............................................................. 104
3.3.2 Hydrothermal and Biologic Activity ................................................... 109
3.4. B asalt G eochem istry .......................................................................................... 111
3.4.1 Basalt Sample Recovery and Analysis Procedures ......................... Il
3.4.2 Second-Order Segment Scale Geochemical Observations .................. 118
3.4.3 Fourth-Order Segment Scale Geochemical Observations of
Segm ent D ........................................................................................... 130
3 .5 . D iscu ssion ......................................................................................................... 13 1
3.5.1 Southward Aging of Volcanic Activity ............................................... 131
3.5.2 The 9' 37' OSC as a Magmatic and Hydrothermal Divide ................. 133
3.5.3 Evidence for a Magmatic Perturbation at 9' 37'. ................................ 136
3.5.4 Association of E-Type MORB ............................................................ 138
3.6 . C onclu sions ....................................................................................................... 144

4. C O N C L U SIO N S ........................................................................................................ 147

APPENDIX FREQUENTLY USED ACRONYMS AND DESCRIPTIONS OF
SAM PLIN G TECHN IQ U ES .................................................................................... 151

R E F E R E N C E S ................................................................................................................ 153

BIO G RA PH ICA L SKETCH ........................................................................................... 164









Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the Requirements for the Doctor of Philosophy GEOCHEMISTRY OF EASTERN PACIFIC MORB:
IMPLICATIONS FOR MORB PETROGENESIS AND THE NATURE OF CRUSTAL ACCRETION WITHIN THE NEOVOLCANIC ZONE OF TWO RECENTLY ACTIVE RIDGE SEGMENTS

By

Matthew C. Smith

December 1999

Chairman: Michael R. Perfit

Major Department: Geological Sciences

A detailed investigation of two recently volcanically active mid-ocean ridge (MOR) segments in the N.E. Pacific Ocean has resulted in a better understanding of crustal accretion within the neovolcanic zones of medium- to fast-spreading ridges. The use of submersible and other near-bottom systems, in conjunction with rock sampling by wax core, provides excellent spatial control, resulting in a database of precisely located in situ geologic observations and samples. The combination of dense high-resolution sampling, temporal control gained by recent eruptive activity, and multi-year monitoring efforts makes these investigations unique in MOR research.

The CoAxial Segment of the Juan de Fuca Ridge (JdFR) is a medium-spreading rate ridge that has experienced at least three volcanic eruptions between 1981 and 1993. Evidence of amagmatic extension and chemical heterogeneity between different lava flows indicates that the CoAxial Segment has behaved as a magma limited system, despite the occurrence of recent eruptive activity. The current CoAxial Segment neovolcanic zone is greater than 1 km wide, and spatial focussing of eruptive activity is









poor relative to the more magmatically robust Cleft Segment of the JdFR and 9-10 N segment of the East Pacific Rise (EPR). A broader survey of the central JdFR reveals that, on a regional scale, geochemically different melt regimes can be associated with distinct structural, morphologic and volcanic provinces.

Detailed study of ridge-axis morphology, structure, and chemistry of the faster spreading, magmatically robust, 9- 10' N segment of the EPR shows a different relationship between ridge-axis structure and magmatic activity than that observed at the CoAxial Segment. Data from a third-order overlapping spreading center (OSC) located at 90 37' N suggests that there is a close association between magmatic activity and axial discontinuities, and further suggest that this OSC is a magmatic and hydrothermal divide between adjoining third-order ridge segments. Temporal constraints correlate a recent magmatic event to southward propagation of the eastern OSC limb. These data establish a direct causal relationship between magmatic activity and evolution of ridge-axis discontinuities, and confirm that segmentation of melt delivery to the ridge-axis may affect ridge-axis structure along the fast spreading EPR.















CHAPTER 1
GENERAL INTRODUCTION

The mid-ocean ridge (MOR) system is a globe-encircling volcanic mountain chain more than 60,000 km long. It extends throughout all of the world's major ocean basins, rising some 1000-3000 m above the abyssal sea floor. Mid-ocean ridges are the site of oceanic crustal accretion and lithospheric genesis and, as such, are important to the study of plate tectonics and Earth's geological processes. While an understanding of magmatic processes occurring at MORs is an essential part of understanding plate tectonics, the significance of magmatic activity at ocean ridges extends far beyond geological interests. Magmatic activity along MORs provides the source of heat that drives extensive hydrothermal circulation within the oceanic crust which, in turn, exerts a strong influence on ocean chemistry. Not only does this volcanically driven hydrothermal system help to control seawater chemistry, but it also provides the energy that supports a rich and diverse abyssal ecosystem that utilizes chemosynthetic reduction as the basis of its food web.

Surprisingly, relatively little of this immense volcanic system has been studied in detail. Difficulties in investigating such a remote and extreme environment limited direct examination until the 1960's when advances in technology allowed man to travel to and freely access the mid-ocean ridges, which typically range from 2000-4000 meters below sea level. Early studies sampled the MOR system with coarse spatial resolution, providing insights on some of the first-order differences between different ridges. The









advent of swath mapping systems that allow for rapid underway bathymetric mapping of large areas of seafloor greatly advanced the understanding of MOR morphology and structure.

Discontinuities and offsets in the strike of the ridge-axis separate the MOR system into segments of varying scale, depending on the magnitude of the axial discontinuity. Macdonald et al. (1988) established four different orders of ridge axis discontinuity classified on the basis of the discontinuity's spatial dimensions. In this categorization, which is today generally recognized as the accepted nomenclature, first-order spreading segments are tectonically defined and bounded by large transform offsets or propagating rifts whose offsets are such that plate boundaries behave rigidly, juxtaposing crust of greater than approximately 0.5-1 Myr age difference across the offset boundary. Offsets are typically greater than 50 km and partition the ridge-axis at intervals of 100-1000 km. Second-order segments have a length scale of 50-300 km, and they are bounded by smaller, non-rigid discontinuities such as large overlapping spreading centers (OSC) with an offset greater than 3-5 km or small offset (less than 20 km) non-rigid transform faults. Third-order segmentation divides the ridge-axis into characteristic lengths of 30-100 km, and these segments are bounded by smaller OSCs (0.5 to 3-5 km). These smaller OSCs are often associated with small (10's of meters) increases in axial depths. Lastly, fourthorder segments have typical length scales of 10-50 km and are not commonly associated with increased axial depth near the segment ends. These fourth-order discontinuities are defined by small non-overlapping offsets of the ridge axis and/or small (1 '-5) bends in the ridge-axis referred to as deviations in axial linearity (DEVALS) [Langmuir et al., 1986]. Although first-order segments and their associated axial discontinuities are









thought to be persistent on the order of millions of years, smaller ridge offsets may not be as temporally stable [Macdonald et al., 1988b].

As investigations of the MOR system were undertaken at finer and finer spatial scales, it became apparent that the mid-ocean ridge environment was the product of complex interactions between magmatic, tectonic, hydrologic, and biologic processes. Only by compiling detailed investigations of processes occurring at different spreading segments would a more comprehensive understanding of this environment be achieved. To date, only a small percentage of the mid-ocean ridge system has been subjected to detailed multi-disciplinary investigation at the first to fourth-order segment scale.

Oceanic ridges exhibit a range of plate divergence or "spreading rates". Full

divergence rates vary by over an order of magnitude, and to a first-approximation ridge morphology, structure, and magmatic activity are strongly influenced by this spreading rate [Perfit and Chadwick, 1998]. Typically, slow-spreading ridges (full plate-divergence rate on the order of 1-4 cm/year), exemplified by ridges in the Atlantic and Indian oceans, rise steeply from the abyss relative to faster-spreading MORs. They are generally rifted by an axial graben that can be 10's of kilometers across and greater than 1500 meters in relief. Slow-spreading ridge terrains are generally more greatly affected by extensional tectonics and faulting and, as such, tend to be much rougher than at faster-spreading ridges. Volcanic morphology of the axis floor, which is the site of active crustal accretion, tends to be dominated by discrete constructional features that may coalesce to form mounds or ridges. Lava morphology is often pillowed, a characteristic associated with relatively slow lava effusive rates. The zone of crustal accretion, or "neovolcanic zone," is thought to be relatively wide and unfocussed (2-12 km wide), with relatively









less control on the location of magma emplacement in the crust relative to faster spreading regimes.

Contrary to the above, fast spreading ridge systems (-8-16 cm/yr full-rate),

exemplified by the East Pacific Rise, exhibit very different characteristics. Characteristic morphology of fast spreading ridges is that of a broad swell or "rise" in the abyssal seafloor. Fast spreading ridges generally lack a large fault bounded axial valley, and when an axial graben is present, it is of a much smaller scale than those present at slow spreading ridges. Overall ridge structural fabric is less rough, and melt focussing appears to be greater, with a majority of volcanic activity being restricted to a much narrower zone of accretion (100-200 m wide). Magmatic activity appears to be much less episodic than in slower-spreading environments, and overall ridge fabric appears to be more greatly dominated by magmatic processes rather than tectonic processes. Many sections of fast-spreading ridge can be associated with a seismically detectable axial magma chamber in the shallow crust, while these features are largely absent from seismic imaging of slow-spreading ridge-axes. Medium spreading-rate ridges (-4-8 cm/yr fullrate), exemplified by the Juan de Fuca Ridge in the northeast Pacific, exhibit characteristics intermediate between the two end-member axial morphologies (neovolcanic zone 200-2000 m wide).

While recognition of the first-order characteristics has allowed for great

advancing in the understanding of mid-ocean ridges, details of the process of oceanic crustal accretion has been hampered by a lack of any high resolution temporal control to the origins of observable phenomenon. The first documented recent eruption was not until the 1990s when anomalies between repeat bathymetric surveys were able to show the









existence of a seafloor volcanic constructional terrain that was not present 6 years prior (time of eruption constrained to being between 1981 and 1987) [Chadwick et al., 1991, Embley et al., 1991]. Since then, fewer than 10 additional eruptions have been confirmed and investigated, several of which are reported on in this dissertation. Numerous questions fundamental to understanding MOR crustal accretion and the linkages between geologic activity and important biologic and oceanographic processes still exist. Outstanding questions, among the as-of-yet unsatisfactorily addressed geologic issues include (but are not limited to): (1) How frequently and with what spatial consistency do volcanic eruptions occur on a given segment of MOR? (2) What is the dependence of eruptive frequency on spreading rate and melt supply? (3) How does spreading rate and melt supply affect melt focussing and the physical configuration of the neovolcanic zone?

(4) How does magmatic activity affect ridge axis structure and segmentation or viceversa? (5) It has been observed that different segments of ridge spreading at similar overall rates can have very different axial morphologies. Is this observation best explained by a temporal cyclicity to the intensity of magmatic and tectonic activity at discrete spreading segments, or do discrete segments have inherent morphologies resulting from differences in volcanic activity that are constant over longer time scales?

Answers to these and other questions that relate to temporal variability and evolution of magmatic, hydrothermal and biological systems can only be answered through the integration of detailed, long-term studies and active monitoring of many different portions of the global MOR system. While this dissertation does not provide holistic answers to any of the above questions, it does present the results of detailed









geologic and petrologic studies of two recently active ridge segments, each occurring in different spreading-rate and apparent melt-supply regimes.

Documentation of recent eruptive activity along two second-order segments of the eastern Pacific mid-ocean spreading system provides a rare opportunity to examine the geometry and petrogenetic associations of magmatic events that form the basis for crustal accretion along medium to fast spreading divergent plate boundaries. The CoAxial Segment of the medium spreading-rate Juan de Fuca Ridge and the 90- 100 N Segment of the fast-spreading East Pacific Rise were the sites of eruptive activity in 1993 and 199192 respectively [Fox et al., 1995; Haymon et al., 1993]. The Juan de Fuca Ridge located approximately 500 kilometers west of the Washington/Oregon/British Columbia coast. The 9-10' N latitude segment of the East Pacific Rise is located west of the Central american coast, between the Clipperton and Siqueros fracture zones. Detailed maps of these two regions are provided in chapters 2 and 3 respectively.

Not only are these two second-order spreading segments [Macdonald et al., 1988] different in terms of spreading rate, but they also differ greatly in axial morphology, and possibly current melt supply rates from the upper mantle. The 9�- 100 N segment of the East Pacific Rise has an axial morphology typical of faster spreading ridge segments dominated by magmatic activity. Contrary to this, the CoAxial segment has an axial morphology intermediate in nature between the fast- and slow-spreading end-members, and is much more indicative that amagmatic extension plays a prominent role in the development of axial morphology. Additionally, these two recently active ridge segments provide a good contrast to the Cleft Segment, another comparably studied portion of the Juan de Fuca Ridge that although spreading at a similar rate to the CoAxial









segment, displays many of the volcanic features observed on the EPR [Smith et al., 1994]. The two investigations presented in this dissertation, and the results that arose from them, are briefly summarized below in "Abstract form" for the convenience of the reader, providing a bit more detail than is presented in the overall dissertation abstract.

1.1. The CoAxial Segment


Recent investigations along the central portion of the Juan de Fuca Ridge have identified at least three different eruptive units that were emplaced along the CoAxial Segment between 1981 and 1993 [Chadwick et al., 1995]. Major element, trace element and isotopic compositions of mid-ocean ridge basalt (MORB) samples recovered by submersible, ROV and wax core are examined in order to critically assess the spatial and temporal chemical variability on regional and local scales. The data allow us to quantify the chemical variances within and between each of the mapped flow units, identify and define genetic relationships between the flows, and to document the relationship of these recent eruptives to older basalts recovered along the CoAxial and other nearby segments of the Juan de Fuca Ridge.

Geochemical data show that each of the three recent eruptives are distinct from one another, supporting the hypothesis that at least three separate volcanic units have been emplaced along the CoAxial segment since 1981. Further, when compared to fresh looking eruptives sampled from within Axial Seamount caldera and along its associated north and south rift zones, recent CoAxial eruptives have isotopic and elemental characteristics distinct from lavas associated with Axial Seamount volcanism. This suggests that magmas parental to lavas sampled from each of these regions have unique origins, and that it is highly unlikely that Axial Seamount was the source of the 1993









CoAxial eruption, as has been suggested. CoAxial segment lavas are geochemically very depleted and show no geochemical evidence for influence of Axial Seamount hotspot. The integration of these geochemical data with observational, photographic and sidescan sonar data defines several different tectono-magmatic provinces in the region between the Vance and Cobb segments of the Juan de Fuca Ridge. The presence of these different provinces suggests quite distinct sub-axial melting regimes exist along this medium rate spreading ridge.

1.2. The East Pacific Rise between 9' and 10'N


Alvin based investigations of a small ridge-axis discontinuity, located at the crest of the East Pacific Rise crest between -9' 35-37'N, provide strong evidence that the feature is actually a small overlapping spreading center (OSC) that divides two magmatic and hydrothermal systems. Two overlapping axial summit collapse troughs (ASCT) morphologically express this third-order ridge-axis discontinuity. The eastern ASCT that extends from the north has an approximately 0.4 km right-lateral offset from the western ASCT that extends from the south and overlaps it by at least 1.2 km. Changes in volcanogenic morphology and increased hydrothermal and biologic activity observed between surveys of the area in 1989 and 1991 suggest that a magmatic event affected this region during this time-period. These changes combined with geochemical data from basalt samples and hydrothermal fluids recovered during three Alvin dives suggest that the eastern limb of the OSC is actively propagating to the south. Observational data that support this hypothesis include: (1) the occurrence in 1991 of extensive areas of bacterial bloom and diffuse venting of 390C, low-chlorinity hydrothermal fluids along the southern extension of the eastern ASCT; (2) regions of local collapse of drained out young-









looking lava flows, the presence of lava pillars, and hydrothermal fluids venting directly out of glassy, young-looking lava flows along the southward extension of the trend of the eastern ASCT; and (3) the presence of several extinct sulfide edifices within the western ASCT, and venting of high chlorinity hydrothermal fluids at temperatures of greater than 113'C in 1991 and 1994 within that region.

MORB geochemistry lends further support to the hypothesis that southward

propagation of the eastern OSC limb at 90 37' N is related to recent magmatic activity in the 90 37-52' N area. Basalt samples with the highest MgO (most primitive, highest temperature) are typically the freshest-looking lavas and tend to be spatially associated with the ASCT of the eastern OSC limb, and its extension to the south. More fractionated (lower temperature) basalt samples are found along the western limb ASCT and in the zone of overlap between the two summit troughs. The more evolved MORB western ASCT MORB are consistently older in appearance than the fresher, younglooking high-MgO lavas to the east. Recently enhanced magmatic activity of the OSC's eastern limb is apparent by MgO content of the most recent eastern ASCT lavas which have Mg#'s that are anomalously high relative to the trend observed in the regional MORB geochemical data. These lavas have major element chemistry similar to the 1991 eruptive at 9' 46-52' N and may be related to the same phase of magmatic activity. Additionally, the chemistry of hydrothermal fluids recovered both within the region of the OSC and along segments of the EPR crest proximal to it, indicates that the OSC marks a hydrographic divide. Hydrothermal fluids generally have higher chlorinity than seawater south of the OSC, and a lower chlorinity than seawater north of it (including fluids from the eastern OSC limb). Investigation of hydrothermal fluids between 9017-









54'N have shown that low-chlorinity fluids are associated with either eruptive/intrusive events, or occur in areas with very shallow and robust heat sources. With time, the fluids often increase in chlorinity. It is only at the 9�37'N OSC and further south that extensive extinct sulfides and fluids with chlorinities -1.5 times seawater have been sampled, suggesting that hydrothermal systems north of the OSC have been more recently perturbed by a magmatic event than those on the failing rift and further south. It is proposed that the OSC's southward propagation is locally driven, resulting from a combination of renewed and enhanced magmatic activity north of the propagator, and a waning magmatic system south of it. Local occurrence of E-MORB on the Pacific plate near the OSC may reflect lower degrees of melting in the region of the retreating OSC limb.














CHAPTER 2
GEOCHEMICAL CHARACTERISTICS AND PETROGENETIC RELATIONSHIPS OF RECENTLY ERUPTED MORB FROM THE CENTRAL JUAN DE FUCA RIDGE:
SEGMENT SCALE, INTERFLOW, AND INTRAFLOW VARIATIONS


2.1. Introduction


2.1.1 Field Investigations and Previous Studies

The 1993 detection of a seismic swarm thought to be related to dike intrusion

along what is now known as the CoAxial Segment of the Juan de Fuca Ridge [Fox et al., 1995] triggered an extensive series of investigative cruises to this portion of the midocean ridge system. A total of three rapid response cruises undertaken between July and October of 1993 (the seismic event lasted from June 26 until July 10, hydrocasting began on July 3 and the first direct sea floor observations began on July 9) investigated the remotely detected seismic events. Investigators participating on these cruises sampled hydrothermal event plumes and visually confirmed that a basalt flow had indeed erupted onto the ocean floor in the location where the most intense seismic swarms had been located [Embley et al., 1995; Baker et al., 1995]. These cruises concentrated on mapping and sampling lavas, biota, and hydrothermal waters associated with the new flow and along the path of proposed dike emplacement. Subsequently, at least 7 cruises to this region have expanded the sampling and mapping efforts to include other portions of the CoAxial Segment as well as Axial Seamount and its associated north and south rift zones.









Prior to the aforementioned events, very little was known from this section of the Juan de Fuca Ridge (with the exception of Axial caldera). The CoAxial Segment had not been recognized as a distinct second-order MOR segment [Macdonald and Fox, 1983; Embley et al., 1995], and geochemical data from only four relatively poorly located dredges had been reported in the literature (Eaby et al., 1984; Rhodes et al., 1990).

2.1.2 Regional Geology

The geology and morphology of the CoAxial Segment and its surrounding terrain is discussed in detail by Embley et al. [in press], and, as such, will only be discussed here in sufficient detail to provide basic background information. This paper is, in many ways, complimentary to and an extension of a recent paper by Embley et al. [in press].

The Juan de Fuca Ridge (JdFR) is a first-order, medium spreading rate ( -6 cm/yr full rate) ridge segment located approximately 500 km off of the Oregon/Washington/ British Columbia coast (Figure 2-1). It extends some 490 km along strike (-020') between the Blanco and Sovanco Fracture Zones which intersect the ridge-axis at

-44'30' and 48'45' N latitude respectively. The JdFR has been subdivided by various authors into at least 6 different second-order ridge segments ranging from 50-150 km in length. More detailed discussions of the JdFR and its second-order subdivisions have been published [Delaney et al., 1981; Johnson and Holmes, 1989; Kappel and Ryan, 1986; Kappel and Normark, 1987; Hey, 1982]. Axial Seamount, a large ridge-axis centered volcano that is the latest expression of the Cobb melting anomaly [Desonie and Duncan, 1990], is located in the middle portion of the JdFR at -45' 55'N latitude and 130' W longitude (Figure 2-1). This large edifice rises more than 1000 m from the surrounding abyssal terrain and has produced a large summit caldera and two major rift





























Figure 2-1: Bathymetric map of the CoAxial Seamount region taken after Figure 4 of Embley et al., in press. True north is towards the top of the page. Regional NE Pacific map is shown in inset (upper left). Shading depicts interpretation of volcanic terrain based on sidescan sonar and bathymetric data (Embley et al., in press). Basalt sample locations are shown as crosses (wax core samples), filled circles (ROV/Submersible samples) and filled squares (starting/ending points of dredges). Solid black outline depicts the interpreted aerial extent of Axial Seamount north rift zone volcanism (see text). Dashed lines mark major rift zones of Axial Seamount and the CoAxial Segments. Rectangular outlines mark the Floc and Flow Sites which are shown in Figures 2-2a (southern outline) and 2-2b (northern outline) respectively. Regional shading is based on interpretation of sidescan and bathymetric data and is keyed to the legend (lower right).































Aberdeen V.


Brown Bear Smt.


Floc Site Source Site


't 1


Cobb


460 50'N





460


460 30'N


460 20'N


460
1 ON


Rogue V.



2' 42t


20'W









zones extending north (to -46'15' N) and south (to at least 450 30' N) from the main edifice, which is some 20 km in diameter [Johnson and Holmes, 1989]. Though this edifice is the youngest volcano of the Cobb-Eickelberg chain, a greater than 300 km long seamount chain, Axial Seamount lavas do not display geochemical enrichments typically associated with hot spot volcanism [Rhodes et al., 1990]. Together, Axial Seamount and its north and south rift zones comprise the Axial Seamount Province.

The southern end of the Coaxial Segment exists in an overlapping relationship

with the Axial Seamount's north rift zone (NRZ) to its west, overlapping it by at least 25 km. The southern terminus of the CoAxial segment is at Helium Basin (-460 N and 129055' W) which lies at the base of the steep, faulted NE flank of Axial Seamount [Lupton, 1990]. The CoAxial Segment extends north-northeast along a strike of -022' to

-46' 40' where it overlaps some 10 km with the Cobb segment to its east [Embley et al., in press]. The ridge is shallowest at its southern end, and deepens northward along the strike of the ridge by approximately 425 m to a depth of -2550 m at its northern terminus. The CoAxial Segment's axial valley is quite wide compared with other JdFR second-order segments (up to 2 km), as is the apparent current zone of active crustal accretion, which will be discussed further below. It is intriguing that the only other segment on the JdFR with as wide of an axial valley is the Vance Segment which lies immediately south and with Axial Volcano's south rift zone (SRZ) [Embley et al., in press].

Based upon our sampling regimen and the criteria presented in Embley et al. [in press], we have designated three different sites along the CoAxial segment as the primary focus of our investigations. These sites listed from south to north are the "Source Site",









the "Floc Site" and the "Flow Site" (Figure 2-1). Each of these sites is described in detail in Embley et al. [in press], and the reader is referred to that paper for a detailed discussion of their geology and morphology.

2.2. Data Collection and Sample Distribution


2.2.1 Alvin and ROV Based Sample Acquisition

A total of 136 site-specific in situ basalt samples have been recovered from the

CoAxial segment since the first rapid response cruises in late 1993 (Figures 2-1 and 2-2). These samples were collected utilizing a variety of submersible (104 samples) and remotely operated vehicle (ROV) platforms (32 samples) employed during 7 different oceanographic expeditions conducted between 1993 and 1996.

The majority of ROV and submersible sampling concentrated around the three primary study sites (described above and in Embley et al. [in press]), with an emphasis placed on sampling lavas of different apparent age and flow morphology. In addition, special attention was paid to recovering a denser and well distributed suite of samples for areas where discrete eruptive units could be identified and mapped on the basis of visual observations and depth differences between repeat bathymetric surveys [Chadwick et al., 1995]. Relative ages and flow contact relationships were deduced from stratigraphic relationships, flow morphology, amount of sediment cover and presence or absence of sessile organisms.

2.2.2 Surface Based Sample Acquisition

The suite of basalt samples recovered by submersible and ROV is supplemented by 57 samples recovered by wax corer (out of 66 attempts), two dredge hauls and one TV Grab (Figures 2-1 and 2-2). Surface based sampling was generally performed at night or






























Figure 2-2: Expanded view of the Floc (A) and Flow (B) sites depicted by rectangle outlines in Figure 2-1. Figures are after figures 11 and 8 of Embley et al., in press. Basalt sample locations, mapped boundaries of recent flows discussed in text, major fissures and low temperature hydrothermal venting locations are superimposed on images generated by deep tow AMS 60 sidescan sonar (Embley et al., in press). Flow boundaries were determined in consort with visual and photographic mapping, repeat bathymetic surveying and sidescan data. Wax core samples are shown as crosses, and submersible/ROV collected samples as circles. Fissures marked by bold white lines (22A) depict the current zone of hydrothermal activity, and interpreted ridge-axis, which extends northward passing directly underneath the 1993 Flow (2-2B).










460 20'N


46' 19'N 460 18'N 46 17'N 460 16'N
1290 45'W


129� 44'W 1290 43'W


F
Jr .. ..... U3"j"


129�42'W





















46' 32'N 460 31'N Graben -


129' 35'W


129' 34'W








during submersible/ROV down time. Wax cores primarily sampled the Axial Seamount NRZ (15) and regions along the CoAxial segment outside of the major sites of submersible investigation. Dredges sampled Rogue Volcano and a topographic high bounding the east side of the CoAxial Segment. The video guided grab sample recovered basalt from a fresh flow outcropping along Axial Seamount NRZ. This flow was also sampled by wax core.

2.2.3 Sidescan Sonar and Near Bottom Photographic Data Acquisition

Photographic data gathered by submersible, ROV and near bottom camera tow, as well as sidescan imagery acquired using SeaMARC II and AMS-60 sidescan sonar systems, have been incorporated in our regional geologic interpretations. Submersible, ROV and camera tow data were generally located using long-baseline transponder navigation that provides positional accuracy to within approximately 20 meters. AMS-60 sidescan data have a spatial resolution of 1 m. Details of the distribution of both the photographic data and the sidescan sonar imaging used to place samples in a geologic context are presented in Embley et al. [in press].

2.3. Analytical Methods


2.3.1 Sample Preparation

All basalt samples were described and catalogued at sea. Geologic context and the location for each sample were carefully documented, and glassy rinds sub-sampled when present. Shoreside, a representative suite of samples were sectioned for petrographic analyses. Natural basaltic whole rock glasses (quenched glass + crystals) were cleaned and prepared for analysis by x-ray fluorescence. Glasses were rinsed in 2X distilled water and the cleanest portion (generally 5-50 grams for ALVIN/ROV recovered









samples) of the glass was hand picked under binocular microscope for further preparation. Picked glass subsamples were subsequently coarsely crushed and sonicated first in a solution comprised of equal parts 2.5N HCL and 30% H20- (15-30 min) and then in 2X distilled water. Samples were then dried and picked free of any visible surficial alteration or precipitated coating using a binocular microscope. Glass splits to be analyzed by electron microprobe, thermal ionization mass spectrometry (TIMS) or laser ablation ICP-MS (LA ICP-MS) were then further hand picked to assure that only the most crystal-free and pristine glass was analyzed. Glass splits to be analyzed by XRF or solution ICP-MS were ground using either a tungsten-carbide disk mill, a SPEX agate ball mill, or some combination of the two. Use of the tungsten-carbide disk mill was kept to a minimum, but it is likely that in at least some of the samples the elements Ta and W may be to some degree compromised by tungsten-carbide contamination [Thompson and Bankston, 1970]. Samples (< 100 mg) analyzed by TIMS were subjected to a second leaching (-10 minutes) using warmed distilled 6N HCL and digested in a solution of distilled HF and ultra pure HNO3. Dissolved glasses were then chlorinated using distilled 6N HCL and Sr and Nd separated and collected using standard ion chromatographic techniques.

2.3.2 Major Element Analyses

The major element abundances of 153 natural basaltic glasses are reported here and a representative subset of the data are presented in Tables 2-1 and 2-2. The majority of these analyses were measured using a JEOL 8800 electron microprobe (EMP) at the USGS in Denver. Multiple standards were frequently measured during the course of each analytical run, including the mid-ocean ridge basalt (MORB) glass standards JdF-D2 and









Table 2-1. Electron microprobe Analyses of Representative Natural Glasses Recovered by Submersible/ROV Sample ID Latitude N. Longitude W. Loc. # SiO2 A1203 TiO2 FeO MnO MgO CaO Na20 K20 P205 Total


221-0445 460 30.743' 1290 35.147' 2670-4 460 31.553' 1290 34.739' 2672-2 460 31.353' 1290 34.864' 2788-4R 460 31.598' 1290 34.721' 2792-1R-2 460 31.778' 1290 34.678' 2794-2R -460 32.56' -1290 33.974' 2983-R2 460 31.636' 1290 34.775' 2983-R6 460 31.624' 1290 34.749' 2993-R4 460 31.024' 1290 35.040' JA187-7 460 31.696' 1290 34.665' 2672-6 460 31.266' 129� 34.353' 2792-4R 460 31.788' 1290 34.321' 2792-5R 460 31.772' 1290 34.247' 2792-7R 460 31.534' 1290 34.270' 2985-RI 46' 31.424' 129034.161' 2985-R3 460 31.442' 1290 34.232' JA187-4 460 31.475' 1290 35.014' 2948-8R 460 17.439' 129043.425' 2948-9R 460 17.742' 1290 43.276' 2948-11R 460 17.998' 1290 43.177' 2995-R2 460 18.610' 129� 42.940' JA188-la 460 18.632' 1290 42.944' JA188-3 460 18.898' 1290 42.761' JA188-5 460 18.898' 1290 42.761' 2791-3R 460 18.416' 1290 42,525' 2791-5R 460 18.522' 1290 42.461' 2793-5R 460 18.241' 1290 42.642' 2946-1R 460 18.850' 1290 42.440' 2946-2R 460 18.819' 1290 42.228' 2947-1 R-2 460 18.549' 1290 42.446' 2948-IOR 460 17.850' 1290 43.199' 2948-1R 460 17.035' 1290 42.737' 2948-2R 460 17.092' 1290 43.087'


50.6 50.4 50.5
50.5 50.5 50.6
50.5 50.5 50.3 50.5 49.4 49.6 49.6 49.3 49.4 49.4 49.6 50.5 50.3
50.4 50.4 50.4 50.6
50.6 50.2
50.0 50.2 50.1
50.2 50.2 50.5 50.6 50.6


13.8 13.5 13.6 13.6 13.5 13.5 13.6
13.5 13.6 13.6 14.3 14.4 14.4 14.8 14.6 14.6 14.7 14.3 14.4 14.4 14.3 14.3 14.0 14.1 14.4 14.4 14.5 14.3 14.5 14.6 14.4 14.5 14.2


1.61 1.69 1.67 1.66 1.70 1.65 1.68 1.70 1.67 1.68
2.08 2.02 2.03 1.93 1.97 1.97 1.95
1.45 1.40 1.40 1.55
1.48 1.50 1.52
1.43 1.46 1.29
1.45 1.47 1.41 1.35 1.28 1.30


12.5 0.23 6.90 12.9 0.24 6.69 12.7 0.24 6.77 12.7 0.22 6.73 12.9 0,23 6.65 12.7 0.23 6.67 12.8 0.24 6.72 12.9 0.24 6.70 12.6 0.25 6.79 12.8 0.23 6.73 12.7 0.22 6.94 12.3 0.20 6.95 12.4 0.23 6.83 12.3 0.22 7.20 12.4 0.23 7.11 12.3 0.22 7.06 12.1 0.22 7.06 11.1 0.21 7.27 10.9 0.21 7.50 11.1 0.20 7.39 11.3 0.18 7.21 11.4 0.21 7.17 11.5 0.22 7.01 11.6 0.20 7.03 10.9 0.16 7.94 10.8 0.15 7.85 10.9 0.20 7.93 10.7 0.21 7.83 10.8 0.18 7.69 10.6 0.24 7.91 10.5 0.24 7.76 10.2 0.16 7.73 10.7 0.18 7.85


11.4 2.65 0.12 0.14 99.9 11.4 2.66 0.12 0.14 99.6 11.4 2.67 0.12 0.14 99.7 11.3 2.65 0.13 0.15 99.6 11.3 2.71 0.12 0.17 99.6 11.3 2.65 0.12 0.14 99.6 11.3 2.63 0.13 0.15 99.8 11.3 2.69 0.13 0.15 99,8 11.5 2.59 0.11 0.14 99.5 11.2 2.65 0.12 0.15 99.7 10.9 2.71 0.15 0.20 99.5 10.8 2.74 0.16 0.21 99.4 10,8 2.77 0.17 0.19 99.4 10.9 2.77 0.15 0.19 99.7 10.8 2.75 0.16 0.20 99.5 10.9 2.76 0.16 0.20 99.6 10.9 2.76 0.17 0.19 99.6 11.9 2.62 0.15 0.16 99.6 12.0 2.6 0.14 0.13 99.6 11.9 2.68 0.15 0.16 99.8 11.7 2.58 0.14 0.14 99.5 11.8 2.74 0.16 0.15 99.7 11.8 2.69 0.16 0.16 99.7 11.7 2.69 0.16 0,17 99.7 12.3 2.48 0.08 0.1 100.0 12.2 2.56 0.08 0.11 99.7 12.4 2.48 0.08 0. 1 100.1 12.4 2.48 0.08 0.111 99,7 12.3 2.47 0.09 .1 1 99.9 12.4 2.47 0.08 0.12 100.0 12.3 2.44 0.09 0.13 99.7 12.6 2.08 0.10 0.12 99.3 12.2 2.41 0.08 0.14 99.6


Sample Location Key: 1-1993 Flow Site Lava; 2-1980s Flow Site Lava. 3-1980s Floc Site Lava; 4-Floc Site; 5-Flow Site; 6-Source Site; 7-Other CoAxialSegment; 8- Western Fault Block
Ridge, 9-Axial Seamount North Rift Zone; 9*-Recent Flow in NRZ;O-Small Volcanoes in NRZ; II -Rogue Volcano.










Table 2-1 continued.
Sample ID Latitude N. Longitude W. Loc. # Si02 A1203 TiO2 FeO MnO MgO CaO Na20 K20 P205 Total


2948-6R 460 17.185' 129' 43.092' 4 2948-7R 460 17.232' 1290 43.507' 4 2949-1R 46' 17.102' 129043.121' 4 2949-4R 460 17,218' 1290 43.053' 4 2995-R4 460 20.599' 1290 33.525' 4 221-0551 460 30.807' 1290 35.194' 5 2788-6R 460 31.591' 1290 34.831' 5
2792-1 R- 1 460 3 1.778' 129' 34.677' 5 2792-2R 460 31.955' 1290 34.416' 5 2792-8R 460 31.354' 129' 34.418' 5 2794-1R 46' 32.024' 1290 34.412' 5 2983-RI 460 31.71 ' 129' 34.921' 5 2983-R5 460 31.624' 1290 34.749' 5 2989-RI 46� 31.951' 129 34.440' 5 2989-R5 460 31.772' 129' 34.009' 5 2989-R6 460 32.096' 129' 34,395' 5 2990-RI 460 28.994' 1290 36.409' 5 2990-R2 460 29.071' 1290 36.279' 5 2990-R3 460 28.984' 129� 36.281' 5 2991-R3 460 32,280' 1290 34.322' 5 2991-R4 460 32.280' 1290 34.322' 5 2993-R2 460 30.980' 1290 34.838' 5 2993-R5 460 31.068' 129' 35.185' 5 2787-1R 460 09.404' 1290 48,524' 6 2790-1R 460 09.137' 1290 48.615' 6 2790-4R 460 09.336' 1290 48.589' 6 2945-5R 460 09,342' 1290 48.588' 6 1993 Flow Site Unit Average(n=24) �2 Std. Dev.
1980s Flow Site Unit Average(n= 10) �2 Std. Dev.
1980s Floc Site Unit Average(n=12) �2 Std. Dev.


49.4 50.4 50.3 50.5 50.3 48.4 49.4 48.9 49.8 50.0 50.0 48.9 48.1 49.0 48.9 50.1 49.7 50.3 50.0 49.5 49.3 50.4 48.3 50.3 50.4 50.4 50.5
50.5 0.2 49.4 0.4
50.5 0.2


14.7 1.33 14.3 1.29 14.3 1.31 14.4 1.31 14.5 1.30 17.3 1.04 15.5 1.19 16.2 1.15 14.4 1.98 13.7 1.76 13.6 1.63 15.5 1.52 17.1 1.04 16.0 1.19 14.1 2.15 13.7 1.66
14.3 1.48 14.8 1.18 14.3 1.53 15.6 1.18 15.9 1.19 13.7 1.66 17.2 1.06 13.8 1.63 14.1 1.44 14.0 1.49 14.2 1.48 13.6 1.68 0.2 0.05 14.5 2.00 0.3 0.12 14.2 1.45 0.3 0.21


10.7 0.27 8.29 12.7 2.23 0.12 0.12 99.8 10.8 0.19 7.76 12.2 2.48 0.08 0.13 99.7 10.7 0.23 7.89 12.3 2.43 0.07 0.09 99.6 10.6 0.21 7.94 12.0 2.42 0.07 0.10 99.6 10.6 0.20 7.86 12.3 2.37 0.11 0.12 99.7 9.0 0.17 9,29 12.1 2.43 0,04 0.06 99.8 9.8 0.24 8.25 12.5 2.62 0.06 0.08 99.6 9.6 0.18 8.63 12.2 2.6 0.06 0.08 99.7 12.2 0.23 6.87 10,9 2.77 0.17 0.20 99.6 12.2 0.24 7.01 11.8 2.82 0.09 0.14 99.7 12.3 0.28 6.99 11.5 2.59 0.09 0.13 99.1 10.7 0.22 7.68 11.4 2.83 0.11 0.15 99.0 8.8 0.20 9.56 12.1 2.34 0.03 0.07 99.3 9.6 0.20 8.62 12.2 2.55 0.04 0.07 99.5 12.9 0.22 6.70 10.9 2.75 0.15 0.21 99.0 12.1 0.23 6.88 11.5 2.66 0.08 0.13 99.1 10.8 0.21 7.70 12.4 2.61 0.07 0.12 99.5 9.9 0.22 8.06 12.7 2.52 0.05 0.10 99.8 10.8 0.24 7.49 12.5 2.61 0.07 0.12 99.5 9.7 0.22 8.39 12.6 2.52 0.05 0.07 99.8 9.7 0.19 8.53 12.3 2.53 0.04 0.08 99.8 12.5 0.24 6.81 11.4 2.63 0.11 0.15 99.7 9.1 0.15 9.39 12.0 2.41 0.03 0.06 99.7 12.1 0.23 6.97 11.7 2.59 0.12 0.15 99.6 11.4 0.19 7.53 11.9 2.51 0.11 0.13 99.7 11.4 0.18 7,41 11.8 2.54 0.11 0.13 99.5 11.1 0.23 7.38 12.0 2.51 0.13 0.15 99.6 12.8 0.24 6.73 11.3 2.67 0.13 0.15 0.2 0.02 0.11 0.1 0.06 0.01 0.02 12.4 0.23 7.03 10.9 2.75 0.16 0.20 0.4 0.02 0.22 0.1 0.04 0.01 0.02 11.2 0.20 7.28 11.9 2.63 0.15 0.15 0.6 0.03 0.55 0.5 0. 19 0.03 0.05


Sample Location Key: 1-1993 Flow Site Lava; 2-1980s Flow Site Lava; 3-1980s Floc Site Lava; 4-Floc Site; 5-Flow Site; 6-Source Site; 7-Other CoAxialSegment; 8- Western Fault Block
Ridge, 9-Axial Seamount North Rift Zone; 9*-Recent Flow in NRZ;0-Small Volcanoes in NRZ; II -Rogue Volcano.






Table 2-2. Electron microprobe Analyses of Natural Glasses Rcovered by Rock Core, Dredge or Video Grab Sample ID Latitude N. Longitude W. Loc. # SiO2 A1203 TiO2 FeO MnO MgO CaO Na20 K20 P205 Total


94RC 19 94RC20 95 RC43 95 RC42a 95 RC42b 95 RC41 95 RC40 95 RC39 95 RC38 95 RCI3 95 RCI4 95 RCI6 95 RCI5 94 RCI9 94 RC18 94 RCI7 94 RCI5 94 RC 16 94 RC20 94 RC21 95 RC27 95 RC28 95 RC25 95 RC24 95 RC23 94 RCIa 94 RC Ib 95 RC22 94 RC2 94 RC3 94 RC4 95 RC21


460 16,358' 460 18.089' 460 10.790' 460 11.279' 460 11.279' 460 11.707' 460 12.089' 460 12.595' 46" 12.989' 460 14.308' 460 15.238' 460 15.431' 460 15.950' 460 16.358' 460 16.441' 46� 16.760' 460 17.631' 460 17.674' 460 18.090' 460 18.277' 460 19.160' 460 19.225' 460 21.726' 460 22.030' 460 22.624' 460 22.664' 460 22.664' 460 23.117' 460 23.762' 460 24.360' 460 24.750' 460 24.900'


129" 44.052' 1290 43.127' 1290 47.280' 1290 46.107' 1290 46.107' 1290 46.650' 129" 46.318' 129" 46.987' 1290 45.825' 1290 44.231F 1290 43.684' 1290 43.540' 129" 43.163' 1290 44.052' 1290 43.703' 1290 44.197' 129" 43.236' 129" 43.824' 129" 43.127' 129" 43.261' 129" 41.653' 129" 41.519'
129" 40.386' 129" 41.090'
129" 40.098' 129" 40.009'
129" 40.009' 1290 40.858' 1290 39.321' 129" 38.957' 129" 38.500' 129" 40.374'


50.5 14.5 50.6 14.2 50.0 14.6 49.6 15.4 50.2 13.8 50.2 13.9 50.4 13.8 50.4 13.9 50.4 13.8 50.9 14.3 50.7 14.0 50.4 14.6 50.5 14.7 50.5 14.5 50.5 14.4 50.5 14.1 50.6 14.1 50.7 14.0 50.6 14.2 50.0 14.6 49.8 14.6 49.8 14.7 50.3 14.1 50.6 14.3 50.4 14.4 49.9 14.6 50.1 14.4 49.5 15.4 50.0 14.5 50.1 14.3 50.4 14.4 50.1 14.3


aStandard data based on replicate analyses run over a several month period, each being the average of 5 or more spot analyses.


1.16 10.3 0.19 1.46 11.3 0.22 1.45 10.6 0.22 1.23 10.1 0.15 1.86 12.4 0.20 1.85 12.2 0.21 1.69 12.5 0.19 1.67 12.1 0.21 1.79 12.3 0.23 1.30 10.4 0.20 1.37 11.1 0.21 1.23 10.3 0.21 1.32 10.5 0.20 1.16 10.3 0.19 1.32 10.4 0.21 1.47 11.3 0.22 1.49 11.5 0.22 1.47 11.4 0.22 1.46 11.3 0.22 1.24 10.6 0.21 1.59 11.1 0.18 1.58 11.1 0.20 1.74 11.9 0.22 1.65 11.2 0.22 1.34 10.6 0.22 1.32 10.7 0.21 1.41 10.9 0,20 1.19 10.6 0.20 1.37 10.8 0.21 1.43 10.9 0.22 1.50 11.0 0.20 1.68 11.8 0.22


8.0 12.6 2.37 0.11 0.09 99.7 7.1 11.8 2.69 0.16 0.15 99.6 7.6 12.2 2.85 0.10 0.12 99.8 8.2 12.2 2.55 0.08 0.12 99.7 6.5 11.2 3.21 0.18 0.17 99.7 6.5 11.2 3.08 0.18 0.17 99.5 6.5 11.3 2.94 0.14 0.15 99.7 6.8 11.4 2.86 0.15 0.15 99.6 6.7 11.3 2.93 0.16 0.18 99.7 7.8 12.3 2.41 0.09 0.10 99.7 7.5 11.9 2.54 0.09 0.13 99.6 8.2 12.4 2.40 0.09 0.14 99.9 7.8 12.2 2.42 0.14 0.13 99.8 8.0 12.6 2.37 0.11 0.09 99.7 7.8 12.3 2.48 0.11 0.13 99.6 7.4 11.9 2.63 0.11 0.13 99.8 7.0 11.7 2.72 0.15 0.15 99.7 7.3 11.9 2.62 0.12 0.14 99.8 7.1 11.8 2.69 0.16 0.15 99.6 8.0 12.6 2.44 0.10 0.11 99.8 7.5 11.7 2.78 0.13 0.14 99.5 7.4 11.7 2.83 0.14 0.14 99.6 7.0 11.6 2.84 0.16 0.14 99.8 7.1 11.7 2.79 0.20 0.17 99.8 7.9 12.1 2.55 0.09 0.13 99.8 8.0 12.2 2.61 0.09 0.13 99.8 7.7 12.3 2.59 0.10 0.12 99.8 8.4 12.1 2.33 0.11 0.08 99.9 7.8 12.3 2.60 0.09 0.13 99.8 7.7 12.3 2.63 0.09 0.13 99.8 7.4 11.9 2.81 0.11 0.14 99.7 7.1 11.4 2.90 0.17 0.18 99.8






Table 2-2 continued.


Sample ID Latitude N. 94 RC5 46� 25.253' 94 RC6 460 25.505' 94 RC23 460 26.011F 94 RC7 460 26.018' 94 RC8 460 26.367' 94 RC9 460 27.002' 94 RCI0 46027.663 94 RCI I 46� 27.997' 94 RC12 460 29.115' 94 RC14 460 34.201' 95 RC32 460 19.51 95 RC34 460 22.418 95 RC35 460 20.464 94 RCI3 460 34.359' 94 RC22 460 18.474' 94 RC25 460 02.231' 94 RC26 460 08.137' 94 RC28 46� 09.719' 94 RC27 46� 09.646' 94 RC29 460 10.384' 94 RC30 460 10.260' 94 RC34a 460 15.855' 99GTVA- I a 460 09.720' 94 RC32 460 11.566' 94 RC33 460 13.174' 62DR2-2 -460 23' 62DR2-4a -460 23' 62DR2-6a -460 23'


JdF-D2 2392-9 JdF-D2 2392-9


average measured value average measured value 2-Sigma Std. Dev. (relative %) 2-Sigma Std. Dev. (relative %) Error used in plots (relative %)


# Si02 A1203 TiO2 FeO MnO


Longitude W. Loc. 1290 38.404' 7 1290 38.724' 7 1290 38.341' 7 1290 38.345' 7
1290 38.001' 7 1290 37.483' 7 1290 37.136' 7 1290 36.706' 7 1290 36.053' 7 1290 32.590' 7 1290 46.255' 8 129� 43.991' 8 1290 45.991' 8 1290 35.227' 8 1290 46.687' 8 1300 00.712' 9* 1290 56.754' 9 1290 55.265' 9* 1290 56.253' 9
1290 55.394' 9 1290 54.716' 9 1290 49.456' 9 1290 55.230' 9* 1290 53.461' 10 1290 51.162' 10
- 1290 28' 11
- 1290 28' 11
- 1290 28' 11


50.1 14.3
50.2 14.5 49.6 14.8 50.0 14.6 50.5 14.2 49.8 14.7 50.2 14.4 49.6 15.4 50.2 15.0 49.6 15.6 50.3 14.3
49.9 14.9 48.1 17.1 47.5 17.1
47.4 17.4 49.9 14.3 49.5 14.4 49.9 13.9
49.7 14.2 49.8 14.2 49.5 14.5 49.5 14.4 49.8 14.1 48.2 17.1
48.1 17.3 50.1 15.2
50.0 15.0 50.1 14.8 50.8 13.8
49.9 15.6 0.4 0.6 0.6 1.1 0.5 1.5


1.40 11.0 0.19 1.41 10.6 0.20 1.42 10.6 0.20
1.39 10.6 0.20 1.51 11.4 0.21 1.58 11.2 0.21 1.50 10.9 0.22
1.10 9.7 0.18 1.10 9.7 0.17 1.20 9.9 0.19
1.48 11.0 0.22 1.39 10.0 0.21 1.20 9.5 0.19 1.06 9.6 0.18
0.97 9.6 0.19 1.54 10.9 0,22 1.55 11.3 0.21
1.69 11.7 0.22 1.55 11.3 0.22
1.64 11.3 0.24 1.51 11.2 0.20 1.54 11.4 0.17 1.67 11.8 0.20
1.15 9.0 0.18 1.12 8.7 0.15 0.93 9.4 0.17 0.70 9.1 0.16 0.66 9.2 0.16 1.89 12.2 0.22 1.28 9.3 0.17 2.01 0.9 4.09 3.17 0.9 7.92 3.0 1.5 10


'Standard data based on replicate analyses run over a several month period, each being the average of 5 or more spot analyse


Table 2-2 continued.


MgO
7.7 7.9
8.0 7.9 7.3 7.7
7.6 8.5 8.2
8.4 7.4
7.9 8.8 10.1 9.7 7.2
7.3 6.8 7.1
7.0 7.5 7.3
6.9 9.2 9.3 8.5 9.2 9.4 6.8 8.6 1.1 1.1 1.5


CaO Na20 K20 P205 Total 12.4 2.62 0.09 0.14 99.9 12.3 2.64 0.09 0.11 99.9 12.3 2.64 0.11 0.13 99.7 12.3 2.69 0.09 0.10 99.9 11.7 2.78 0.12 0.12 99.8 11.7 2.76 0.11 0.13 99.8 12.1 2.72 0.09 0.12 99.8 12.7 2.59 0.05 0.08 99.9 12.8 2.57 0.07 0.07 99.8 12.3 2.58 0.07 0.07 99.9 12.0 2.92 0.10 0.14 99.9 12.2 2.83 0.07 0.11 99.6 11.9 2.86 0.03 0.08 99.6 11.8 2.53 0.04 0.06 99.9 12.0 2.57 0.04 0.07 99.9 12.2 2.90 0.20 0.15 99.4 12.2 2.95 0.19 0.17 99.7 11.9 3.01 0.20 0.17 99.4 12.3 2.85 0.18 0.14 99.5 12.1 2.93 0.16 0.15 99.5 12.2 2.98 0.16 0.09 99.8
12,1 2.99 0.17 0.13 99.8 11.9 3.05 0.21 0.15 99.7 11.8 2.81 0.02 0.07 99.6 12.3 2.78 0.07 0.09 99.9 13.3 1.94 0.08 0.07 99.7 13.8 1.63 0.03 0.04 99.6 13.9 1.63 0.03 0.05 99.9 10.8 2,77 0.22 0.23 99.7 12.1 2.61 0.11 0.12 99.8
1.2 4.27 5.37 5.25
1.0 1.71 5.61 11.03
1.5 5 5 10









USNM. Plots of SiO2 abundance (wt.%) vs. Total (wt.%) of all oxides measured were made for each individual analytical run based on the secondary standard data acquired during that run. Where strong correlations existed, the raw abundance data for SiO2 was corrected along a linear regression after Reynolds [1995]. Individual oxides were normalized to the standard JdF-D2 (unfractionated JdF-Reynolds, 1995) in order to greater facilitate their comparison to other published MORB data.

A small subset of the major element data was analyzed by EMP at the

Smithsonian Institution by T. O'Hearn. Smithsonian data were not SiO2 corrected but were normalized to the secondary standard VG2. In order to facilitate direct comparison to our microprobe data, these data were corrected to a "JdF-D2" normalization using numerous replicate samples and secondary standards analyzed in both labs. Measured values for the MORB standard JdF-D2 and long term analytical precisions are shown in Table 2-2. The complete major and trace element set is available from the author.

2.3.3 Trace Element Analyses

2.3.3.1 X-ray fluorescence

Seventy-five powdered basalt whole rock glasses were measured at the University of Florida for the elemental abundances of Ba, K, Nb, Rb, Sc, Sr, Ti, V, Y and Zr using a fully automated ARL8420+ x-ray fluorescence (XRF) spectrometer with a Rh-target endwindow tube. Details of the operating conditions are reported in Smith et al., [1994]. Analytical precision based on long term reproducibility of standard reference materials is better than 5% for the elements K, Sr, Sc, Ti, V, Y and Zr. Precision is better than 10% for the elements Ba, Nb and Rb at values greater than 20, 3 and 2 parts per million (ppm) respectively. Precision drops to greater than 20% below these levels. Average values









and 2-sigma precisions for the standard BHVO-l are given in Rubin et al., 1998, and average measured values for our internal standard ENDV (an enriched MORB from the Endeavour Segment, JdFR) are shown in Table 2-3.

2.3.3.2 ICP-AES/MS

Abundance data for major elements, the rare earth elements (REE) and selected other trace elements were measured by inductively coupled plasma source emission spectrometry, (ICP-ES) and inductively coupled plasma source mass spectrometry (ICPMS) at the Geological Survey of Canada (GSC) in Ottawa. Elements/oxides measured by ICP-ES include major element oxides and a few trace elements (Co, Cr, Cu, Ni, and Zn), and those measured by ICP-MS include the REE and other trace elements. An additional group of samples recovered by wax core were analyzed by ICP-MS at the United States Geological Survey (USGS) in Denver. Fifty-two samples were analyzed by ICP for this study. A representative subset of these data are presented in Tables 2-3 (GSC data) and 2-4 (USGS data). Multiple internal standards analyzed by isotope dilution (lID) techniques on a thermal ionization mass spectrometer (TIMS) were included in each run to assess and correct for any inter-laboratory analytical bias. Correlation between the two laboratories was excellent. Details of analytical procedures, detection limits and precision for the GSC ICP data can be found on the internet at http://l32.156.95.172/chemistry/, the GCS Analytical Laboratory homepage. Analytical procedures, detection limits and precision for USGS ICP-MS data can be found in Arbogast [1990].










Table 2-3 Concentration (ppm) of the Rare Earth and other Select Trace Elements in Representative
Submersible Samples Analyzed by ICP-AES ,-MS ,and XRF.
Samnle * 2788-4R 2792-1R2 2794-2R 2983-R2 2983-R6 2792-4R 2792-5R 2792-7R 2948-1IR
SiO,3 es 51.1 51.3 51.5 51.3 51.1 49.9 50.3 50.1 51.4 TiO2 es 1.3 1.26 1.3 1.31 1.33 1.51 1.57 1.55 1.22 A1203 es 13.8 13.6 14 13.8 14.1 14.9 14.6 14.5 13.8 Fe203t es 12.8 12.5 12.9 12.9 13.2 12.5 12.7 12.5 11.6 Fe203 es 1.1 1.2 1.8 0.9 1.5 1.4 1.1 0.7 1.1 FeO es 10.6 10.2 10 10.8 10.5 10 10.5 10.6 9.4 MnO es 0.19 0.19 0.19 0.19 0.2 0.18 0.19 0.18 0.18 MgO es 7.14 7.05 7.18 7.2 7.39 7.4 7.35 7.23 7.37 CaO es 10.9 10.7 11 11 11.2 10.5 10.4 10.3 11.5 Na7O es 2.44 2.52 2.63 2.56 2.54 2.67 2.66 2.57 2.43 K20 es 0.12 0.13 0.14 0.1 0.11 0.19 0.17 0.17 0.15 P205 es 0.11 0.1 0.09 0.1 0.1 0.14 0.14 0.14 0.11
Total es 98.8 98.2 99.8 99.3 100.0 98.8 98.9 98.1 98.7 Co es 46 62 47 46 44 48 45 62 74 Cr es 110 120 120 120 120 220 220 230 240 Cu es 65 63 66 63 64 60 59 57 65 Ni es 61 61 61 59 61 100 100 100 64 Zn es 74 89 69 75 91 73 79 67 65 F @ 169 175 162 156 144 262 246 219 171 S @ 1263 1272 1257 1318 1305 1315 1256 1294 1151 V xrf 405 377 396 406 401 378 386 388 332 Ba xrf 32 25 25 29 28 36 33 36 29 Sc xrf 53 45 48 46 50 48 50 49 48 TiO2 xrf 1.54 1.49 1.58 1.54 1.53 1.82 1.89 1.88 1.39 K20 xrf 0.09 0.09 0.08 0.11 0.11 0.12 0.13 0.12 0.12 Cr xrf 98 85 116 83 84 223 220 241 229 Zr xrf 94 94 97 97 97 148 150 148 101 Sr xrf 90 89 91 93 93 103 99 101 94 Y xrf 38 38 39 40 40 51 50 51 35 Rb xrf 1.1 1.7 1.4 2.1 1.4 1.6 1.6 1.7 2.0 Nb xrf 3.5 3.4 3.2 4.1 3.0 4.5 5.1 4.7 4.7 Hf ms 2.2 2.4 2.5 2.5 2.4 3.5 3.8 3.6 2.5 In ms 0.11 0.12 0.13 0.09 0.13 0.1 0.14 0.11 0.09 Nb ms 4 3.5 3.3 3.9 3.3 4.7 5.2 5.6 5.1 Rb ms 1.3 1.3 1.5 1.2 1.3 1.7 1.9 1.8 1.8 Sn ms 1.3 1.5 2.7 <0.5 0.7 1.7 1.9 1.1 0.9 Ta ms 4.2 5.7 3.8 5 4.5 4.7 5.7 5 10 Th ms 0.16 0.3 0.22 0.19 0.18 1 1.4 0.31 0,26 U ms 0.08 0.07 0.1 0.1 0.07 0.14 0.13 0.12 0.11 Zr ms 81 85 87 86 83 n.a. n.a. n.a. 92 La ms 3.0 3.2 3.1 3.2 3.1 6.5 7.3 4.9 3.9 Ce ms 9.1 9.4 9.3 9.4 9.2 17.0 18.0 15.0 11.0 Pr ms 1.5 1.6 1.6 1.7 1.6 2.4 2.6 2.3 1.7 Nd ms 9.2 9.2 9.4 9.7 9.6 14.0 15.0 14.0 10.0 Sm ms 3.3 3.4 3.4 3.5 3.5 4.8 5.1 4.7 3.4 Eu ms 1.3 1.2 1.3 1.3 1.3 1.5 1.6 1.6 1.3 Gd ms 5.1 4.9 5.0 5.1 5.1 6.6 6.9 6.4 4.8 Th ms 0.90 0.94 0.97 0.95 0.94 1.20 1.20 1.20 0.84 Dv ms 6.1 6.0 6.2 6.3 6.2 7.8 8.5 8.0 5.8 Ho ms 1.3 1.3 1.3 1.4 1.4 1.8 1.9 1.8 1.3 Er ms 4.0 4.0 4.1 4.0 4.0 5.0 5.2 5.2 3.7 Tm ms 0.57 0.58 0.61 0.58 0.58 0.70 0.79 0.74 0.53 Yb ms 3.8 3.6 3.9 3.9 3.8 4.9 5.2 5.0 3.7 Lu ms 0.55 0.54 0.58 0.55 0.55 0.69 0.75 0.71 0.53 ILa/Sm(n) 0.57 0.59 0.57 0.58 0.56 0.85 0.90 0.66 0.72 Analyzed at the GSC, Ottawa, ON.; -analyzed at the Univ. of Florida, Gainesville, FL., 'values have been corrected for Si loss due to precipitation from analyzed solutions by extrapolation along a correlation between measured SiO2 concentration and Total oxide abundance (r2=0.65) to a total of 99.7 wt%.: *-method of analysis: @ byDionex Ion Chromatography; es by ICP-AES, ms by ICP-MS, xrf by x-ray fluorescence; n.a. - not analyzed. All oxide concentrations are listed as wt.% and all elemental concentrations are as ppm











Table 2-3 continued.


Samnle 2948-8R 2948-9R 2995-R2 2793-5R 2946-IR 2946-2R 2947-1R2 2948-IOR 2948-IR
SiO,3 es 51.3 51.2 51.6 51.1 50.8 50.8 50.9 51.2 50.9 TiO2 es 1.23 1.19 1.25 1.15 1.19 1.21 1.19 1.1 1.03 A1203 es 14 13.9 14 13.8 14.4 14.3 14.3 14.1 14.6 Fe203t es 11.6 11.3 11.7 11.3 11.2 11 11.1 10.8 10.8 Fe203 es 0.9 0.9 I 0.9 1.2 1.1 0.9 0.7 0.6 FeO es 9.7 9.4 9.6 9.4 9 8.9 9.2 9.1 9.2 MnO es 0.18 0.17 0.18 0.17 0.17 0.17 0.17 0.17 0.16 MgO es 7.46 7.34 7.35 7.51 7.85 7.55 7.86 7.76 8.21 CaO es 11.6 11.4 11.5 11.6 11.9 11.6 11.7 11.7 11.9 NaO es 2.44 2.46 2.56 2.56 2.49 2.5 2.51 2.3 2.16 K,0 es 0.14 0.14 0.16 0.1 0.11 0.1 0.11 0.09 0.1 P205 es 0.11 0.11 0.12 0.09 0.08 0.09 0.08 0.09 0.09 Total es 99.1 98.2 99.3 98.4 99.1 98.3 98.9 98.3 98.9 Co es 41 42 40 47 46 52 42 40 48 Cr es 240 250 230 200 300 290 310 270 320 Cu es 66 67 66 70 67 63 68 66 72 Ni es 66 66 64 68 85 80 90 78 94 Zn es 66 65 64 64 67 69 67 63 60 F @ 160 152 167 144 150 145 160 136 125 S @ 1176 1190 1197 1154 1142 1102 1107 1130 1119 V xrf 342 335 340 345 320 327 325 325 303 Ba xrf 32 25 33 21 19 19 18 18 17 Sc xrf 48 47 48 45 50 54 48 47 42 TiO, xrf 1.41 1.39 1.43 1.33 1.34 1.42 1.36 1.27 1.16 K20 xrf 0.12 0.12 0.13 0.07 0.08 0.09 0.08 0.08 0.08 Cr xrf 232 240 221 196 310 305 318 286 331 Zr xrf 102 100 105 84 90 95 91 85 74 Sr xrf 95 94 95 91 98 101 97 94 82 Y xrf 36 36 37 33 34 34 33 32 29 Rb xrf 2.1 2.1 1.7 1.0 1.2 1.4 0.8 1.3 1.5 Nb xrf 4.2 4.6 4.5 2.6 2.9 3.6 3.1 3.2 4.0 Hf ms 2.5 2.3 2.6 2.2 2.3 2.4 2.3 2 1.8 In ms 0.09 0.11 0.11 0.11 0.09 0.11 0.09 0.08 0.1 Nb ms 4.6 4.3 4.6 2.8 3.1 3.6 3.5 2.7 4 Rb ms 1.7 1.8 1.9 0.82 0.83 0.88 0.87 0.85 1 Sn ms 0.9 0.8 1 0.8 1.3 1 0.8 0.7 1.5 Ta ms 4.9 4.7 5 5 5.9 9.7 8 3.3 11 Th ms 0.27 0.24 0.25 0.2 0.13 0.14 0.18 0.16 0.15 U ms 0.12 0.11 0.15 0.07 0.07 0.07 0.08 0.07 0.06 Zr ms 90 89 99 79 87 88 87 73 69 La ms 3.8 3.8 4.0 2.6 2.8 2.7 2.9 2.7 2.6 Ce ms 11.0 11.0 11.0 7.9 8.5 8.4 8.9 8.2 7.5 Pr ms 1.7 1.7 1.8 1.3 1.4 1.4 1.6 1.3 1.2 Nd ms 10.0 9.7 10.0 8.2 9.0 9.0 9.2 8.4 7.7 Sm ms 3.4 3.2 3.4 2.9 3.0 3.2 3.1 2.9 2.6 Eu ms 1.3 1.2 1.3 1.2 1.1 1.2 1.2 1.1 0.9 Gd ms 4.9 4.7 4.9 4.4 4.6 4.5 4.5 4.2 3.9 Tb ms 0.90 0.87 0.92 0.80 0.80 0.87 0.82 0.76 0.69 Dv ms 5.9 6.0 5.9 5.4 5.3 5.4 5.5 5.4 4.7 Ho ms 1.3 1.3 1.3 1.2 1.2 1.2 1.2 1.1 1.1 Er ms 3.7 3.6 3.9 3.4 3.4 3.5 3.5 3.3 3.1 Tm ms 0.51 0.52 0.57 0.49 0.50 0.50 0.50 0.48 0.45 Yb ms 3.6 3.5 3.6 3.3 3.3 3.5 3.4 3.3 3.0 Lu ms 0.53 0.50 0.52 0.45 0.49 0.50 0.48 0.43 0.42 I alSm(n) 0.70 0.75 0.74 0.56 0.59 0.53 0.59 0.59 0.63 Analyzed at the GSC, Ottawa, ON: analyzed at the Univ. of Florida. Gainesville, FL.; values have been corrected for Si loss due to precipitation from analyzed solutions by extrapolation along a correlation between measured SiO2 concentration and Total oxide abundance (r2=0.65) to a total of 99.7 wt%.; *-method of analysis: @ byDionex Ion Chromatography, es by ICP-AES, ms by ICP-MS, xrf by x-ray fluorescence; n.a. - not analyzed. All oxide concentrations are listed as wt.% and all elemental concentrations are as ppm.










Table 2-3 continued.


Samnle * 2948-2R 2948-6R 2948-7R 2949-IR 2949-4R 2788-6R 2792-IRI 2792-2R 2792-8R
SiO, 3 es n.a. 50.5 51.1 51.2 51.3 50.2 49.9 50.0 50.8 TiO2 es n.a. 1.06 1.12 1.11 1.09 0.97 0.97 0.98 1.32 A1203 es n.a. 14.8 14.2 14.2 14.2 15.3 15.6 15.6 13.8 Fe'03t es n.a. 11.1 11.2 11.4 11.2 10.2 10 10.3 11.6 Fe203 es n.a. 0.9 1.4 0.7 0.6 1.1 1 0.5 2.3 FeO es n.a. 9.2 8.9 9.6 9.5 8.2 8.1 8.8 8.3 MnO es n.a. 0.17 0.17 0.17 0.17 0.16 0.15 0.16 0.18 MgO es n.a. 8.33 7.62 7.4 7.29 8.32 8.48 8.36 7.43 CaO es n.a. 12.1 11.9 11.7 11.6 11.7 11.5 11.8 11.5 Na2O es n.a. 2.21 2.31 2.41 2.36 2.43 2.59 2.59 2.53 K20 es n.a. 0.12 0.11 0.08 0.1 0.06 0.07 0.07 0.11 P20 es n.a. 0.09 0.09 0.08 0.07 0.05 0.04 0.06 0.1 Total es n.a. 99.5 98.9 98.6 98.3 98.5 98.4 98.9 98.4 Co es 170 46 52 40 70 45 47 49 49 Cr es 250 310 240 220 220 310 300 310 280 Cu es 67 72 64 67 67 82 77 80 61 Ni es 61 94 66 60 62 110 120 110 72 Zn es 61 62 65 66 64 53 47 53 61 F @ n.a. 120 133 126 121 132 130 137 163 S @ n.a. 1121 1126 1233 1200 1074 1020 1017 1179 V xrf 330 308 331 340 328 263 248 268 353 Ba xrf 19 21 16 17 21 14 10 13 22 Sc xrf 51 45 48 51 47 51 42 49 53 TiO1 xrf 1.25 1.18 1.28 1.27 1.25 1.10 1.11 1.11 1.55 K20 xrf 0.07 0.08 0.08 0.07 0.07 0.05 0.04 0.05 0.08 Cr xrf 227 328 245 225 224 354 319 345 301 Zr xrf 78 75 83 80 80 77 78 79 102
Sr xrf 83 81 91 85 86 95 97 95 103 Y xrf 32 29 33 33 33 31 30 31 38 Rb xrf 1.5 1.0 1.4 0.7 0.7 0.6 0.6 1.1 0.9 Nb xrf 3.9 3.8 3.5 3.0 3.3 2.3 2.1 2.3 3.2
Hf ms 1.9 2 1.9 2 2 1.8 2 2.1 2.4 In ms 0.08 0.16 0.09 0.09 0.09 0.1 0.08 0.06 0.11 Nb ms 4.2 3.6 3.3 2.6 2.9 1.7 1.7 2.2 3.6 Rb ms 0.74 1 0.78 0.82 0.77 0.61 0.56 0.68 0.86 Sin ms 1 3.3 0.5 0.8 0.7 2 1 1.7 1.1 Ta ms 17 5.8 8.6 5.3 7.2 4.3 4.7 7.2 7.5
Th ms 0.12 0.16 0.14 0.11 0.14 0.13 0.14 0.15 0.19 U ms 0.06 0.07 0.07 0.09 0.07 0.05 0.07 0.06 0.07 Zr ms 71 71 71 73 71 66 75 76 94 La ms 2.4 2.7 2.5 2.4 2.4 2.1 2.0 2.3 3.1 Ce ms 7.3 7.8 7.5 7.5 7.3 7.1 6.7 7.4 9.8 Pr ms 1.3 1.3 1.2 1.3 1.2 1.2 1.1 1.3 1.7 Nd ms 7.7 7.6 8.0 8.3 7.9 7.7 7.4 7.8 9.9
Sm ms 2.8 2.7 2.8 2.9 2.7 2.7 2.6 2.8 3.5 Eu ms 1.1 1.0 1.1 1.1 1.0 1.0 1.0 1.1 1.3 Gd ms 4.2 3.9 4.2 4.3 4.1 4.0 4.1 4.3 5.0
Tb ms 0.77 0.73 0.75 0.79 0.78 0.73 0.73 0.77 0.93 Dv ms 5.1 4.8 5.1 5.3 5.5 4.8 4.6 4.8 6.3 Ho ms 1.1 1.1 1.1 1.2 1.2 1.1 1.1 1.1 1.4 Er ms 3.5 3.2 3.4 3.4 3.4 3.2 3.1 3.2 3.9 Tm ms 0.48 0.44 0.45 0.49 0.45 0.46 0.43 0.46 0.55 Yb ms 3.2 3.1 3.4 3.2 3.2 3.1 2.9 3.1 3.7 Lu ms 0.45 0.43 0.45 0.46 0.45 0.43 0.43 0.45 0.55 1La/Sm(n) 0.54 0.63 0.56 0.52 0.56 0.49 0.48 0.52 0.56 Analyzed at the GSC, Ottawa, ON.; 2analyzed at the Univ. of Florida, Gainesville, FL.; 'values have been corrected for Si loss due to precipitation from analyzed solutions by extrapolation along a correlation between measured SiO2 concentration and Total oxide abundance (r2=0.65) to a total of 99.7 wt%.; *-method of analysis: @ byDionex Ion Chromatography; es by ICP-AES, ms by ICP-MS, xrf by x-ray fluorescence; na. - not analyzed. All oxide concentrations are listed as wt.% and all elemental concentrations are as ppm.










Table 2-3 continued.


Samnle * 2794-IR 2983-RI 2787-IR 2790-IR 2945-5R 99GTVA-la 62DR2-1 62DR2-4a Endv(n=4) 2-sigma
Sio,3 es 51.1 50.1 51.3 51.4 51.5 49.6 50.4 50.2 50.2 0.8 TiO2 es 1.33 1.2 1.24 1.21 1.25 1.35 0.62 0.6 1.39 0.02 A1203 es 13.4 15.1 13.9 13.8 13.8 14.3 14.6 14.6 15.1 0.2 Fe,03t es 12.5 11 11.7 11.5 11.7 12.3 9.9 9.9 10.3 0.2 Fe203 es 0.8 0.8 1.4 1.1 0.9 1.4 1.1 0.7 2.3 0.1 FeO es 10.6 9.2 9.3 9.4 9.7 9.8 8 8.3 7.2 0.1 MnO es 0.19 0.16 0.18 0.18 0.22 0.19 0.16 0.15 0.17 0.01 MgO es 7 8.68 7.24 7.33 7.21 7.55 9.25 9.34 7.44 0.17 CaO es 10.9 10.5 11.3 11.4 11.3 12 13.3 13.5 11.5 0.1 NaO es 2.71 2.67 2.42 2.44 2.54 2.73 1.54 1.54 2.99 0.19 K20 es 0.1 0.14 0.13 0.15 0.14 0.22 <0.05 <0.05 0.24 0.01 P20 es 0.09 0.11 0.1 0.11 0.11 0.12 0.04 0.04 0.15 0.01 Total es 98.2 98.7 98.5 98.5 98.7 99.2 99.0 99.0 n.a. n.a.
Co es 43 89 47 67 49 53 44 44 34 1 Cr es 130 300 180 190 180 270 460 450 265 6 Cu es 62 77 69 66 67 87 100 100 65 1 Ni es 57 160 62 61 62 43 93 93 76 1 Zn es 69 62 69 65 70 54 45 47 59 2 F @ 169 209 166 173 167 196 <50 <50 218 13 S @ 1313 1093 1139 1196 1287 n.a. n.a. n.a. n.a. n.a. V xrf 399 278 352 350 356 336 n.a. 228 13 290 Ba xrf 27 27 22 24 224 45 n.a. 16 2 81 Sc xrf 48 43 50 49 49 46 n.a. 46 3 42 TiO2 xrf 1.53 1.40 1.44 1.40 1.46 1.56 n.a. 0.66 0.03 1.57
K20 xrf 0.09 0.11 0.11 0.11 0.12 0.19 n.a. 0.02 0.03 0.24
Cr xrf 91 266 175 184 168 212 n.a. 426 25 236 Zr xrf 94 146 98 94 100 n.a. n.a. 22 4 119 Sr xrf 90 115 99 100 124 n.a. n.a. 29 9 165 Y xrf 38 47 35 34 36 n.a. n.a. 18 1 34 Rb xrf 1.4 1.4 1.3 1.3 1.3 n.a. n.a. 0.0 0.9 4.3
Nb xrf 3.5 4.8 3.6 5.0 4.7 n.a. n.a. 0.7 1.1 7.0
Hf ms 2.5 3.6 2.3 2.3 2.4 3 1.6 1.4 2.8 0.1 In ms 0.16 0.16 0.1 0.1 0.11 <0.05 <0.05 <0.05 0.13 0.04 Nb ms 3 5.1 4.2 4.9 4.6 5.4 0.72 0.61 7.2 1.1 Rb ms 1 1.4 1.3 1.2 1.6 2.2 0.29 0.23 4.4 0.1 Sn ms 1 4.1 0.7 1.1 1 0.7 4.8 <0.5 2.5 1.4 Ta ms 3.9 9.5 6.7 10 8.1 3.8 1.4 0.9 0.6 0.1 Th ms 0.17 0.27 0.22 0.22 0.26 0.31 0.06 0.05 0.72 0.46 U ms 0.08 0.1 0.11 0.11 0.13 0.14 0.06 0.04 0.18 0.01 Zr ms 91 n.a. 87 84 94 n.a. 56 47 n.a. n.a. La ms 2.9 4.4 3.5 3.4 3.6 4.8 0.70 0.70 6.8 1.0 Ce ms 9.4 13 10 9.9 10 14 2.3 2.4 17 2 Pr ms 1.6 2.2 1.7 1.6 1.7 2.0 0.45 0.43 2.4 0.1 Nd ms 9.8 13.0 9.4 8.9 9.9 11 3.1 3.2 13 0 Sm ms 3.6 4.5 3.0 3.1 3.3 3.3 1.3 1.4 3.8 0.1 Eu ms 1.3 1.4 1.2 1.2 1.2 1.3 0.59 0.59 1.4 0.0 Gd ms 5.1 6.0 4.8 4.4 4.7 4.6 2.4 2.5 5.0 0.1 Th ms 0.93 1.1 0.84 0.80 0.86 0.84 0.47 0.49 0.87 0.03 Dv ms 6.5 7.1 5.6 5.4 5.9 5.2 3.2 3.2 5.4 0.1 Ho ms 1.4 1.7 1.3 1.2 1.2 1.1 0.73 0.72 1.2 0.0 Er ms 4.0 4.6 3.4 3.4 3.7 3.2 2.2 2.2 3.4 0.1 Tm ms 0.59 0.63 0.54 0.49 0.53 0.45 0.32 0.31 0.46 0.01 Yb ms 4.0 4.4 3.4 3.2 3.6 3.0 2.1 2.2 3.1 0.1 Lu ms 0.56 0.65 0.47 0.48 0.53 0.47 0.33 0.35 0.44 0.01 laISm(n) 0.51 0.62 0.73 0.69 0.69 n.a. n.a. n.a. n.a. n.a.
Analyzed at the GSC, Ottawa, ON.; analyzed at the Univ. of Florida, Gainesville, FL.: values have been corrected for Si loss due to precipitation from analyzed solutions by extrapolation along a correlation between measured SiO2 concentration and Total oxide abundance (r2=0.65) to a total of 99.7 wt%.; *-method of analysis: @ byDionex Ion Chromatography; es by ICP-AES, ms by ICP-MS, xrf by x-ray fluorescence; n.a. - not analyzed. All oxide concentrations are listed as wt.% and all elemental concentrations are as ppm.









Table 2-4 Concentration (ppm) of the Rare Earth and other Select Trace Elements in Representative Rock Core Samples Analyzed by ICP-MS3

2392-9 2352-2 94RC16 94RC20 94RC22 94RC25 94RC28 94RC30 94RC32 94RC34 95RC27 95RC28 95RC35 95RC36 95RC37 Chond.4 La 3.0 3.7 3.3 4.1 1.3 4.8 4.8 3.6 3.4 5.2 4.4 5.2 2.1 4.4 4.9 0.367
Ce 9.3 11.5 10.8 12.7 4.9 13.8 13.8 11.0 10.2 15.0 12.7 13.8 7.5 12.7 13.8 0.957
Pr 1.5 1.9 1.7 1.8 0.9 1.9 1.9 1.7 1.6 2.2 2.0 2.1 1.3 1.8 1.9 0.137
Nd 8.0 10.0 9.6 9.7 5.2 10.0 10.0 9.2 8.8 12.0 11.0 11.0 7.4 9.6 10.0 0.711
Sm 2.8 3.6 3.7 3.6 2.1 3.4 3.4 3.3 3.3 4.1 3.8 3.8 2.7 3.3 3.5 0.231
Eu 1.1 1.3 1.4 1.3 0.9 1.3 1.2 1.3 1.2 1.5 1.4 1.4 1.1 1.2 1.3 0.087
Gd 4.0 4.9 5.3 4.8 3.0 4.6 4.5 4.4 4.4 5.3 5.3 5.1 3.8 4.3 4.6 0.306
Tb 0.7 0.9 1.0 0.9 0.6 0.8 0.8 0.8 0.8 1.0 1.0 1.0 0.7 0.8 0.8 0.058
Dy 4.8 5.9 6.4 5.9 3.8 5.2 5.2 5.1 4.9 6.0 6.3 6.1 4.6 4.9 5.2 0.381
Ho 1.0 1.3 1.4 1.3 0.8 1.1 1.1 1.1 1.1 1.3 1.4 1.3 1.0 1.0 1.1 0.085
Er 3.0 3.7 4.1 3.7 2.4 3.3 3.1 3.1 3.1 3.9 4.0 3.9 2.9 3.1 3.2 0.249
Tm 0.4 0.5 0.6 0.5 0.4 0.5 0.4 0.4 0.4 0.5 0.6 0.5 0.4 0.4 0.5 0.036
Yb 2.8 3.5 4.1 3.6 2.5 3.1 3.0 3.1 2.9 3.6 3.9 3.8 2.7 2.9 3.1 0.248
Ba 10 9 11 19 1 25 26 17 17 26 17 18 4 25 28 Co 41 42 51 43 52 48 46 50 47 53 47 45 51 44 45 Hf 2 3 2 2 1 2 2 2 2 3 3 3 2 2 2 Nb 2.2 4.0 4.2 5.7 0.6 6.6 7.4 4.4 4.9 7.6 6.4 5.4 1.3 6.6 7.6
Pb 3.2 2.1 5.3 <0.1 3.0 5.5 2.9 5.2 29.0 2.6 3.7 1.8 9.3 1.5 2.8
Rb 0.8 1.0 1.1 2.1 <0.1 2.2 2.2 1.6 1.5 2.3 1.6 1.7 0.3 2.1 2.2
Sc 39 41 53 44 41 46 46 44 42 55 47 46 38 43 45 Sr 120 130 110 100 100 160 160 160 140 180 130 120 160 150 160 Ta 0.1 0 0.2 0.2 <0.1 0.3 0.3 0.2 0.2 0.3 0.3 0.2 <0.1 0.3 0.3
Th 0.1 0.10 0.1 0.2 <0.05 0.3 0.3 0.1 0.2 0.3 0.2 0.3 <0.05 0.2 0.3
U 0.1 0.1 0.07 0.10 <0.05 0.10 0.20 0.07 0.08 0.10 0.1 0.1 <0.05 0.10 0.10
V 280 340 440 370 240 350 350 310 320 400 380 380 250 330 340 Y 28 34 39 35 23 31 30 30 30 36 37 37 27 28 29 Zr 70 100 88 100 50 86 100 80 73 110 110 110 70 78 98 3Analyzed at the USGS in Denver CO.; Chondrite normalizing values after Taylor, 1982.








2.3.4. Isotopic Analyses

Isotopic ratios of Sr and Nd were measured, on a representative subset of 30 and 18 samples respectively, at the University of Florida using a VG Isomass 354 thermal ionization mass spectrometer (Table 2-5). Natural glasses to be analyzed were dissolved at 120'C in sealed Teflon vials using distilled HF and a few drops of distilled HNO1. Sr and rare earth elements were separated and collected using cation exchange chromatography (Dowex 50X 12 resin), and Nd was separated from other rare earth elements (REE) using HCL elution on quartz columns packed with Teflon beads coated with bis-ethylhexyl phosphoric acid [after Richard et al., 1976]. Full procedural blanks are better than 0.2ng for Sr and 0.03ng for Nd.

Sr samples were loaded with Ta oxide on W single filaments. Sr ratio data were acquired in triple collector mode at a beam intensity of 2X 10 " A of 88Sr, with corrections for instrumental discrimination made assuming 86Sr/88Sr=O. 1194. Long-term measurement of NBS 987 yields a 87Sr/86Sr value of .710216 � 45X 10-6 (2-sigma). Nd samples were run as metals loaded on one side of a triple filament array using a Re center filament and Ta side filaments. Data were acquired at a beam intensity of 5XI0-12 A of '44Nd, with corrections for instrumental discrimination made assuming 146Nd/44Nd=0.7219. Long term measurement of the La Jolla Nd standard yields a 143Nd/'44Nd value of .511834 � 18X10-6 (2-sigma).


2.4. Regional CoAxial Segment Geochemical Characteristics


2.4.1 Major Element Trends

MORB recovered from the CoAxial Segment chemical attributes characteristic of "normal" incompatible element depleted MORB [Bryan et al., 1976; Sun et al., 1979;









Table 2-5. Sr and Nd Isotopes for Representative Samples.

Sample ID Lat. N. Long. W. 87Sr/86Sr 143Nd/144Nd Location

221-0445 46.512 129.586 0.702267 0.513214 1 2670-4 46.526 129.579 0.702262 0.513248 1 2672-2 46.523 129.581 0.702263 0.513218 1 2792-3 46.532 129.574 0.702254 n.a. 1 2672-5 n.d. n.d. 0.702325 0.513178 2 2672-6 46.521 129.573 0.702328 n.a. 2 2792-4 46.530 129.572 0.702300 n.a. 2 2985-1 46.524 129.569 0.702312 0.513179 2 2985-3 46.524 129.571 0.702306 0.513187 2 2948-11 46.300 129.720 0.702337 0.513188 3 2948-8r 46.291 129.724 0.702329 n.a. 3 2948-9r2 46.296 129.721 0.702331 n.a. 3 94rc20 46.302 129.719 0.702349 0.513127 3 221-0551 46.513 129.587 0.702266 0.513145 5 2671-2 46.509 129.589 0.702238 0.513195 5 2989-5 46.530 129.567 0.702304 n.a. 5 235-0700 46.274 129.744 0.702302 0.513215 4 2679-1T 46.313 129.704 0.702235 0.513192 4 2948-10 46.298 129.720 0.702273 0.513217 4 2995-3 46.345 129.560 0.702322 0.513165 4 2674-1 46.159 129.802 0.702390 n.a. 6 2675-2 46.178 129.797 0.702369 0.513112 6
2678-3 n.d. n.d. 0.702314 n.a. 6 2945-5 46.156 129.810 0.702321 n.a. 6 94rc25 46.038 130.013 0.702501 0.513120 9* 94rc28 46.162 129.921 0.702499 0.513090 9* 99GTVA-la 46.162 129.921 0.702505 n.a. 9* 62DR-1 46.387 129.473 0.702333 n.a. 11 62DR-2 46.387 129.473 0.702321 0.513199 11 62DR-4a 46.387 129.473 0.702308 n.a. 11 n.d. - no data; n.a. - not analyzed.









Schilling et al., 1983]. Figure 2-3 shows major element concentrations of 139 natural basaltic glasses from the CoAxial Segment analyzed by EMP. The basalt samples range in MgO content from approximately 9.6 to 6.5 weight percent (wt.%). Most of the chemical variability can be ascribed to the effects of fractional crystallization, but dispersion of data at a given MgO value is in excess of analytical uncertainty. Consequently, some variability in parental composition and/or conditions of crystallization is required.

Major element data from CoAxial lavas (Figure 2-3) form a field with a trend generally parallel to modeled low pressure fractional crystallization trends, with the exception of K20 and P205 which increase in concentration at a greater rate than predicted by simple fractional crystallization models [Weaver and Langmuir, 1990]. The apparent greater enrichment of K20 and P205 could be indicative of more complex crystallization processes such as in situ crystallization [Langmuir, 1989]. However, deviation from the modeled fractional crystallization trend also may be an artifact of the greater incompatible element depletions (relative to most CoAxial samples) present in the mafic sample used for the parent composition in the liquid line of descent (LLD) model shown in Figure 2-3.

Examination of geochemical data for MORB recovered from the CoAxial

Segment and comparison to other data from other provinces of the southern and central Juan de Fuca Ridge reveals several notable relationships. Comparison of CoAxial lava chemistry to that of the Cleft Segment and Axial Seamount Province (which includes Axial Seamount caldera, north rift zone, and south rift zone) shows Coaxial lavas to be distinct in several ways. With the exception of one andesitic sample (not shown) recently





















Figure 2-3: Variation of major element oxides (wt.%) with MgO in natural basaltic glasses from the CoAxial Segment, Juan de Fuca Ridge. Symbols are as shown in legend. Data for the 1998 Axial Seamount eruptive are after Perfit et al., 1998. Fields show compositional ranges for ridge axis basalts from the Cleft Segment (gray fill), Axial Seamount caldera (stippled fill), and Axial Seamount's north and south rift zones (white fill) for comparison. Fields are based on data from Smith et al., 1994 (Cleft Segment) and the unpublished data of Perfit et al. [in prep] (Axial Seamount). Lines show liquid lines of descent calculated for fractional crystallization at 1 kilobar (solid), 2.5 kilobars (dashed) and 5 kilobars (dotted) pressure using a recently modified version of the program of Weaver and Langmuir (1990). Parent composition used for the calculations was that of sample 221-0551, and older mafic basalt in contact with the 1993 eruption. See text for discussion.













13 12 110

10



9


8


2


6 7 8 9 106 8 9 MgO MgO













Fig. 3 continued


MgO


0.05


51 50


MgO









recovered from the North Rift Zone [Perfit et al., 1998] and some small on-axis volcanoes discussed below, glass data from the Axial Seamount display a more limited range in MgO, (-8.4 to 6.7wt%) than glasses from CoAxial (9.6 to 6.5wt%). Axial Seamount data from Rhodes et al. [ 1990] are not shown, but display a more restricted range in MgO content (6.9-7.7wt% MgO). In addition to having a greater range in MgO content, the CoAxial sample suite generally exhibits greater variation in other major element oxides at a given MgO than does the suite from Axial Seamount Province. It is, however, important to note that the Axial Seamount data set is smaller (81 glasses) than the CoAxial Segment database (139 glasses). As illustrated on MgO discrimination diagrams (Figure 2-3), there is significant overlap of CoAxial MORB with the field defined by Axial Seamount volcanics, but CoAxial lavas tend to have lower abundances of A1203 and CaO at a given value of MgO and extend to higher values of SiO and FeO. Notably, less than -10% of the CoAxial MORB have K20 or Na2O concentrations that fall within the Axial Seamount field.

Comparison of CoAxial MORB with MORB recovered from within the axis of the Cleft Segment, the southernmost segment on the Juan de Fuca Ridge [Smith et al., 1994] is useful for two reasons. Firstly, the Cleft segment represents one of the least morphologically and tectonically complex ridge segments on the Juan de Fuca spreading system, and it is somewhat spatially removed from complexities related to the Cobb melting anomaly. Secondly, the two MOR segments immediately adjacent to the CoAxial segment, the Vance segment to the south and the Cobb segment to the north, are very poorly petrologically characterized. Although data are very limited [Smith et al., 1994; Rhodes et al., 1990; Karsten et al., 1990; Delaney et al., 1981; our unpublished









data], the Cobb and Vance segments appear to have major element characteristics similar to that of the Cleft segment. Compared to Cleft samples, CoAxial MORB generally have lower abundances of A103 and higher abundances of Na2O at a comparable MgO abundance (Figure 2-3).

2.4.2 Isotopic and Trace Element Signals

Trace element concentrations in CoAxial Segment lavas have strong correlation to MgO suggesting that much of the trace element compositional variation observed within the CoAxial suite can be attributed to fractional crystallization and other differentiation processes. Additionally, covariation between trace elements that is inconsistent with crystallization or assimilation processes requires some variation in parental composition of lavas from this segment. In particular (as is discussed in greater detail in section

5.2.2), geochemical enrichments in a lava flow erupted at the Flow Site between 1981 and 1991 greatly extend the range of some trace element abundances, requiring differences in parental lava source characteristics.

Discounting these few samples, the most incompatible enriched sample in the

CoAxial Segment data field has an abundance of moderately incompatible elements (e.g. Y, Zr, Sm) 40% greater than the lowest abundance sample. Comparison of the most depleted and enriched CoAxial MORB shows that the abundances of the most highly incompatible elements (e.g. Nb, Ba, and Rb) are enriched by 250% to 300% relative to the lowest abundance sample (Table 2-3). Low-pressure fractional crystallization models [Weaver and Langmuir, 1990] indicate that -25% fractional crystallization of the most MgO-rich sample is required to produce the range in MgO values observed in Table 2-3 (note that the range of MgO values in the sample subset shown in Table 2-3 is less than









that observed in the entire suite analyzed by EMP). Maximum corresponding enrichments in trace element concentrations (by assumption of a bulk distribution coefficient equal to 0) expected from fractional crystallization are limited to a factor of

-1.3x the parent composition when 75% of the original liquid remains. Consequently, over enrichment of the most incompatible elements relative to that expected from fractional crystallization are observed. Such incompatible element over enrichments in MORB suites (relative to fractional crystallization models) have been explained via more complex crystallization processes such as in situ crystallization [Langmuir, 1989], but may also be in part due to differences in primitive parental magma compositions.

Figure 2-4a shows latitudinal variation in the 87Sr/ 86Sr isotopic ratio of Juan de Fuca Ridge lavas. Data shown are a compilation of our data (Table 2-5 and unpublished data for Axial Seamount Province and the near ridge Vance Seamounts) and previously published data [Eaby et al., 1984; Ito et al., 1987; and Hegner and Tatsumoto, 1987]. 87Sr/86 Sr data reported by Rhodes et al. [1990] (eight analyses) are not included in the compilation because they have values 10-20 X10-5 lower than the main data trend. This offset to of the Rhodes et al. [1990] data lower 87Sr/ 86Sr values is confirmed by Eaby et al. [1984] who report analyses of some of the same samples.

CoAxial MORB (latitudinal range indicated by vertical dashed lines in Figure 24) are clearly less radiogenic than MORB from the rest of the JdFR (and Pacific MORB in general). CoAxial Sr isotopic data (n=25) range from 0.70224 to 0.70239, whereas values for the rest of the Juan de Fuca are between 0.70240 and 0.70268. The less radiogenic nature of CoAxial 87Sr/86Sr relative to the rest of the JdFR is clearly illustrated in Figure 2-4b, which shows 87Sr/86Sr values averaged by region. The average CoAxial




























Figure 2-4: (A) Distribution of 875r/86Sr for Juan de Fuca Ridge basalts plotted against latitude ('N). Symbols and sources are as in the legend and text. Two-sigma error bar is shown at left. Vertical dashed lines at bottom mark CoAxial Segment boundaries, and the vertical solid line the Cobb offset. (B) 87Sr/86Sr compositions averaged by region. Error bars are � 1 standard deviation for each average. Symbols from left to right are averages of data from: the southern Cleft Segment, the northern Cleft Segment, the Vance Segment (including Vance Seamounts), Axial Seamount Province (see text), the Coaxial Segment, the Cobb Segment and the Endeavour Segment. The number of analyses for each average are shown next to the data point. Line is a linear regression through the data points excluding the CoAxial and Endeavour Segments which respectively display anomalously depleted and enriched lavas chemistries relative to other JdF data. See text for discussion.









Ito et al, 1987 Hegner and Tatsumoto, 1987 Our unpublished JdF Axis Eaby et al., 1984 Our unpublished Vance Smts


1980s Flow (n=5) 1980s Floc (n=4) 1993 CoAxial (n=4) Rogue Volcano Recent NRZ Flows


0.70270


0.70260 r# 0.70250 00
CI)
00 0.70240



0.70230 0.70220 0.70260



0.70250



0 0.70240 00

0.70230


0.70220


46 47 48 Degrees Latitude North









Segment 875r/86Sr value of 0.70231 is > 15x 105 below the trend established by the other JdFR data south of the Cobb offset. It is interesting to note that although the overlap in data from the different regions is significant; there is a trend of decreasing 87Sr/86Sr with increasing latitude. This trend is also apparent in the unaveraged data (Figure 2-4a) suggesting that this trend is not simply an effect of the averaging of 87Sr/86 Sr. Regional surveys of JdFR Sr isotopes [Eaby et al., 1984, Rhodes et al., 1990] have noted that Axial Seamount does not have a Sr isotope signature distinct from the rest of the JdFR. but a regional trend in ridge axis lavas has not previously been recognized. It is important to note that while this trend is intriguing, it should be considered with caution. Significant overlap exists in the range of 87Sr/ 86Sr values observed in lavas from different regions of the JdFR, and latitudinal variation in 143Nd/144Nd does not display a similar linear correlation (though Nd data are more limited in quantity and spatial distribution).

CoAxial Segment lavas are not only distinct from other JdFR lavas with respect to Sr isotopic ratios, but also with respect to Sr abundance. Magmatic strontium abundances are not greatly affected by typical MORB crystallization (bulk liquid-crystal distribution coefficient is near unity), and in the absence of significant disequilibrium crystal accumulation or fractionation, large differences in Sr abundance between different lavas are likely a trait acquired during melt generation. A histogram of Sr abundances for MORB recovered from the Cleft Segment, CoAxial Segment, and Axial Seamount (Figure 2-5) shows the data-ranges for each of the three regions to be quite distinct from one another. As with the Sr isotopic ratios, CoAxial Segment lavas are the most depleted (in Sr abundance) on the JdFR. Only four samples from the CoAxial Segment (all among the oldest samples recovered) overlap with the data-range for Cleft Segment lavas, and




































0\ Z 0\ I O\ - -


It- - - - - - --O t0)ItC 1tC
C' C4 " -t )W- "0-c-


Sr (ppm)





Figure 2-5: Histogram showing distribution of XRF analyzed strontium concentrations (ppm) in CoAxial Segment, Axial Seamount Province, and Cleft Segment basalts. Data used for the Cleft Segment is that of Smith et al., 1994. Data for Axial seamount province is a compilation of unpublished data of Perfit et. Al. (in prep) and that of Rhodes et al., 1990. See text for discussion.








only two Axial Seamount samples overlap with either the CoAxial or Cleft segment ranges. Strontium abundances of 96% of the CoAxial lavas analyzed fall between 79 and 104 ppm [average = 94 ppm], while 96% of Cleft Segment lavas have Sr between 105 and 121 ppm [average= 114 ppm]. Axial seamount lavas have a Sr abundance range distinct from either two, with 97% of lavas having a Sr abundance between 130 and 170 [average=147 ppm]. Regional basalt isotopic and trace element signatures will be considered further below.

2.5. Discrimination of Discrete Flow Units


2.5.1 Visual and Bathymetric Flow Discrimination

Three different, recently erupted, mappable units have been identified on the CoAxial Segment. Assessment of differences in depth-data from repeat sea beam surveys of the segment clearly shows bathymetric anomalies corresponding to three different regions of very fresh lava mapped by submersible, ROV, deep camera tow, and AMS 60 deep tow sidescan sonar [Chadwick et al., 1995, Embley et al., in press]. Figure 2-2 shows AMS60 images for the Floc (2-2a) and Flow sites (2-2b) with outlines depicting the interpreted boundaries of the recent eruptives (after Embley et al. [in press], Figures 8 and 11). Superimposed on the sidescan images are the locations of basalt samples discussed below. Timing of the emplacement of these units is constrained to two intervals by the acquisition dates of the bathymetric data, the first occurring between 1981/82 and 1991 and the second occurring between 1991 and 1993. The lava associated with the SOSUS monitored T-phase seismicity [Embley et al., 1995] spatially corresponds with a sole bathymetric feature created between the 1991 and 1993 surveys, supporting this lava flow's association with the 1993 seismic event. This lava flow,









hereafter referred to as the 1993 flow, is approximately 3.8 km long, and up to -500m wide and -30m thick. Morphology of the 1993 flow is dominated by lobate to small pillowed forms. While most of the volume resides in the main body of the flow, several small segments of 1993 flow crop out north of the main body in a small graben that extends north and south of the eruption. This graben presumably served as the eruptive vent for the 1993 flow [Chadwick and Embley, 1998; Chadwick et al., 1995; Embley et al., in press]. Detailed descriptions of the individual flows and eruptive sites (Figures 2-1 and 2-2) are presented in Embley et al. [in press].

Comparisons between the 1981 and 1991 bathymetric data (Flow Site) and the 1982 and 1991 bathymetric data (Floc Site) reveal two additional positive bathymetric anomalies generated prior to the 1993 volcanic event. Seafloor observations confirm predictions of recent volcanic activity based on bathymetric differences between surveys, and identify two additional very fresh-looking volcanic constructional features. The first, hereafter referred to as 1980s Flow, is at the Flow Site approximately 700 meters east of the 1993 lava flow. It is of a similar morphology and size to the 1993 Flow [Chadwick et al., 1995]. The second, hereafter referred to as 1980s Floc, is at the Floc Site about 500 meters west of the zone of active venting and interpreted path of the 1993 dike emplacement [Embley et al., in press]. This flow (and associated bathymetric anomalies) consists of a linear array of several elongated mounds aligned with their long axis parallel to ridge strike. These fresh lavas occur over a distance of at least 7 km, and field relations indicate that these lavas were erupted along a pre-existing topographic ridge. While bathymetric anomalies were identified coincident with the thickest part of the recent flow, absence of bathymetric anomalies where glassy lavas were observed implies









that this flow is, in places, thinner than the 5-15 meter resolution of the bathymetric differencing technique [Embley et al., in press.]. The nearly complete overlap in the temporal ranges defined by the three surveys prior to 1993, 1981-1991 for the Flow Site and 1982-1991 for the Floc Site, limits the ability to discriminate whether these two older (relative to the 1993 flow), but still recent, lava flows were emplaced during one or two separate crustal accretion events. Although field observations of sediment cover suggest that the 1980s Floc eruption may be slightly older than the 1980s Flow eruption, these data also cannot preclude the possibility that both eruptive units are related to a single crustal accretion event. An examination of the inter- and intra- flow geochemical variation further constrains the genesis of these lavas and seafloor constructional features.

2.5.2 Geochemical Flow Discrimination

Examination of the data for the three recent CoAxial Segment lava flows, as

shown in Figure 2-4a, reveals several important features. Firstly, for each of the eruptive units there is no statistically significant intra-flow variation in 875r/86Sr (or in 143Nd/144Nd, not shown), although lavas from the segment as a whole do exhibit significant variation. Secondly, the 1980s Floc and Flow site eruptives cannot be distinguished from one another using 87Sr/ 86Sr, although the 1993 Flow (number of samples analyzed (n) = 4) is distinctly less radiogenic than both the 1980s Floc (n=4) and 1980s Flow (n=5) eruptive units. There are no obvious crustal components or processes that could lower the Sr isotopic ratio of the 1980s magmas to produce the 1993 lava. Accordingly, the 1980s CoAxial lavas and the 1993 CoAxial lava cannot be easily related to a single parent, and are, therefore, neither comagmatic nor cogenetic. Further








discrimination of the three identified recent CoAxial eruptive units requires an examination of elemental abundances.

Electron microprobe data of natural glass samples recovered from each of the

three recent eruptive units, along with average compositions and standard deviations for each unit, are shown in Table 2-1. Averaged data and analytical precisions for two MORB standards measured during the analytical runs are shown in Table 2-2.

2.5.2.1 The 1980s Floe Site Eruptive

Glasses from the three recent CoAxial flows exhibit both inter-flow and intra-flow major element chemical variability (Figure 2-6). The 1980s Floc Site eruptive, which is an -7 km long linear array of elongate mounds, exhibits the greatest overall variation in major element concentrations. Data from samples collected in situ range from 7.5 to 7 wt.% MgO. The inclusion of wax core sample 94RC 19, which targeted the southernmost mound of this flow (Figure 2-2a), extends the data to 8wt% MgO. This intra-flow variation, while not great (-12% relative to concentration), is well outside of analytical error, which is less than 2%. Perhaps more significant than the overall extent of intraflow MgO variability is the co-variation between the major element oxides occurring in samples from this unit.

Figure 2-6a shows liquid lines of descent (LLDs) calculated for fractional

crystallization at lkbar pressure. A key to symbols is given in the figure caption. The models were generated using both the original program of Weaver and Langmuir (dashed line) [Weaver and Langmuir, 19901, and a more recently modified version of the program that includes the oxides K20 and P2O5 in the calculations (solid line) [Reynolds, 1995]. Each of the algorithms of Weaver and Langmuir utilize a combination of experimental














15 A 120315

14 14

13 CaO 13

12 12


10 ,10

2.75 2.75


2.25 2.25


1.75 1.75


1.25 T 2 1.25

6.8 7 7.2 7.4 7.6 7.8 8 8.2 6.7 6.8 6.9 7 7.1 7.2 7.3
MgO MgO


Figure 2-6: MgO concentrations plotted against concentration of A1203 (squares), FeO (triangles), CaO (circles), Na20 (diamonds) and TiO2 (crosses) for the 1980's Flow Site eruptive (open symbols), 1980's Floc Site eruptive (black filled symbols) and 1993 Flow
Site eruptive (gray filled symbols). Error bars are shown for the highest MgO content
sample in each of the three flows, and represent � 2 standard deviations (2-sigma) of our
standard data. In general, 2-sigma precisions (relative to concentration) are equal to or better than: 0.5% for SiO2, 1.5% for FeO* (total Fe as FeO), MgO, CaO and A1203, 3% for TiO2, 5% for Na20 and K20, and 10% for P205 and MnO. (A) Comparison of glass data for the two flows erupted between 1981 and 1991. Trend lines show theoretical 1 kilobar liquid line of decent (LLD) calculations using the original program of Weaver and Langmuir, 1990 (dashed line) and a recently updated version of the same program
designed to include K20 and P205 (solid line) (See text). (B) Comparison of the two
flows recently erupted at the Flow Site. Trend lines (generated using the updated LLD
program) show the effects of pressure on modeled LLD paths with trend lines calculated
at both I (solid) and 2.5 (dashed) kilobars pressure.









and thermodynamic data to predict the chemical evolutionary path, or liquid line of descent (LLD), of MORB liquids as they cool and crystallize mineral assemblages in chemical equilibrium with their host liquids. The algorithms are designed to model liquid chemical evolution during both closed system (equilibrium crystallization) and open system (fractional crystallization) crystallization processes. However, over the limited range of crystallization relevant to the intra-flow chemical variation observed, there are no significant differences between the LLDs generated using either the equilibrium or fractional crystallization models. The parent lava composition used for the calculations shown in Figure 2-6a is 94RC 19, the southernmost and most mafic (MgO-rich) sample recovered from the 1980s Floc site eruptive. Five oxides plotted against MgO (A1203, CaO, FeO*, Na2O and TiO2) show covariation trajectories parallel to and bounded by the two model LLDs. The primary difference between the two models occurs because plagioclase's arrival onto the liquidus is suppressed in the model that considers effects of K20 and P205 in calculating the LLD path.

The model shown in Figure 2-6a calculated using the original Weaver and

Langmuir program (dashed line) has crystallization of clinopyroxene (cpx) as the sole liquidus phase beginning at a temperature of 1192�C. After only 4' of cooling (1 188�C) both olivine (ol) and plagioclase (pl) begin crystallizing as the liquid becomes triply saturated. Using the same data to run the newer model (solid line) produces a liquid that once again begins with cpx as the sole liquidus phase, but crystallization now begins at a temperature 10 degrees higher (1 202�C) than in the older (no K20 or P205) model. Cpx remains the sole liquidus phase through 10 degrees of cooling at which point olivine joins the liquidus. Plagioclase begins crystallizing at the same temperature as in the older









model (1 188'C), but by that time approximately 5% crystallization has already occurred. The effects of delayed plagioclase crystallization during the first 5% of crystallization are most dramatically seen in the LLDs for FeO (FeO enrichment is initially suppressed) and A1203 (A1203 shows greater initial enrichment). In general, the newer model does a better job of reproducing the natural variability observed in the 1980s Floc glasses, although our A1203 data suggest that initial increases in A1203 concentration are not as great as are predicted from the model. TiO2 is the exception to the other major element oxides, showing better agreement with the older model.

The models' suggestion of clinopyroxene as a sole liquidus phase at low pressures in such relatively differentiated MORB is to be questioned. There is no experimental evidence that supports cpx occurring as a sole liquidus phase in such evolved compositions. Though small amounts of water may have the capability to suppress plagioclase crystallization [Michael and Chase, 1987], the algorithms of Weaver and Langmuir do not incorporate H20 concentrations as a variable. Moreover, the 10 degree maximum interval of sole cpx crystallization is near the limit of experimental reproducibility [Grove et al., 1992]. It is, therefore, likely that sample 94RC19 was multiply saturated at its liquidus temperature.

The close agreement of the data shown in Figure 2-6a with the predicted LLD trends suggests that elemental variations in glasses from the 1980s Floc eruption are consistent, within the limitations of current experimental models, with evolution along a liquid line of descent resulting from progressive cooling and crystallization. Thus, the chemistry of the sample recovered from the southernmost of the 1980s Floc mounds (94RC 19) is consistent with it being cogenetic, comagmatic, and parental to the rest of








the samples from the 1980s Floc eruptive. Mineral phases observed in the lavas are consistent with the LLD models, and the total range of MgO compositions observed in the 1980's Floc Site eruptive corresponds to approximately 240 of cooling and 20% crystallization, as predicted by the LLD model (more recent version).

Trace element data are not from pure glasses, but instead are of "whole rock

glasses" (which include a small amount of crystals which may not be a truly equilibrium assemblage). Accordingly, it is difficult to use our trace element data to rigorously constrain the crystallization conditions modeled, but it is nonetheless clear that some degree of liquid evolution has occurred within this flow unit. Whether this variation truly represents the process of fractional crystallization or is simply reflecting differences in the evolution of the liquid during equilibrium cooling and crystallization cannot be adequately assessed utilizing the glass data alone.

Trace element data gathered by XRF on whole rock glasses can be used to help constrain to what extent crystallization processes responsible for liquid differentiation within the 1980s Floc unit were closed (equilibrium crystallization) or open (fractional crystallization) system processes. If liquid differentiation observed within the glass data results from fractional crystallization processes, then composition of the whole rock glasses (crystals+liquid) should evolve with the liquid composition. Comparison of covariation of glass MgO and "whole rock glass" Zr concentration with latitude (Figure 2-7) reveals some interesting relationships. In this case, because the trend of the long axis of the 1980s Floc lava emplacement is aligned roughly parallel to ridge strike at

-022, latitude is proportional to distance along strike of emplacement. A strong correlation exists between MgO content of glasses and distance along strike of the flow,






























Figure 2-7: Latitudinal variation in MgO (upper) and Zr (lower) concentrations for basalt samples from the 1980s Floc Site eruptive. Symbols are keyed to method of recovery and are as shown in the legend. Error bars (�2-sigma) are shown, as are r2 values for linear regression lines. Linear regressions are based on all the data shown. Though latitudinal increase in whole rock glass Zr concentrations (XRF) are similar to trends seen in the EMP data (upper) supporting the hypothesis of geochemical differentiation along the strike of flow emplacement, Zr increase is insufficient to be the result of pure fractional crystallization processes.











































46.28 46.29 46.3 46.31
Degrees Latitude North


0


7

110 105


100


95


90
46.27


46.32









with MgO decreasing (cooling and crystallization increasing) from south to north. A similar, though less well defined, latitudinal covariation seems to exist in the Zr data, though the lack of "whole rock" data for the wax core 94RC 19 limit the total variation observed. Nonetheless, if one extrapolates along the Zr trend (which over this limited range approximates a linear trend) to arrive at a Zr value for this sample, one would postulate it to have approximately 97 ppm Zr. Using this value as the parent composition, it becomes clear that the increase in Zr from 97 to 106 ppm (< 10% increase relative to concentration) is not sufficient to invoke pure fractional crystallization as the differentiation process. This small an increase would require a liquid-crystal bulk distribution coefficient (bulk D) of approximately 0.57, a value far too high for an element as incompatible during MORB crystallization as Zr (a "bulk D" equal to zero would yield a 25% increase in concentration after 20% crystallization.) Based on the variations of "whole rock" glass Zr data (and other highly incompatible trace elements not shown) it is likely that some intra-flow magmatic differentiation along the strike of emplacement exists. The differentiation process, however, was inefficient and cannot be accurately modeled via true Rayleigh fractionation [Shaw, 1970]. More likely, liquid differentiation was a process intermediate between closed system equilibrium crystallization and pure fractional crystallization.

2.5.2.2 The 1980s and 1993 Flow Site Eruptives

As discussed above, two discrete lava flows erupted between 1982 and 1993 have been identified at the Flow Site. Major element compositions of glasses recovered from the 1980s Flow unit (Figure 2-6 a and b, open symbols) and 1993 Flow unit (Figure 2-6b, gray filled symbols) show compositional variations similar to, but more restricted than,








those from the 1980s Floc unit (shown only in Figure 2-6a). LLD models depicted by the trend lines in Figure 2-6b show the affects of varying crystallization pressure between 1 (solid line) and 2.5 kilobars (dashed line). In both models the starting composition is that of sample 2792-7, the most MgO-rich sample recovered from the 1980s Flow Site eruptive. Observed mineral phases and major element covariation with MgO in both of the Flow site units are consistent with LLD models and small amounts of liquid evolution (-9�C of cooling and 2-5% crystallization). Although the total intra-flow chemical variation (4-5% relative to concentration for MgO) for these two units is close to analytical error, the consistent behavior between the different oxides and the LLD parallel trends are compelling. That being said, there is an apparent decrease in MgO and increase in Zr with latitude in the 1993 Flow (not shown), similar to the recent Floc Site eruptive, although only the northernmost and southernmost samples are completely distinguishable from one another considering the 2-sigma uncertainties. Neither of the two recent Flow Site eruptive units occurs over as great a distance along strike as does the 1980s Floc unit. The limited length of the Flow Site units, combined with lesser overall geochemical variability and restricted sample coverage for the 1980s Flow eruptive, inhibits the ability to define along strike chemical variability for these flows. Examination of Figure 2-6 does, however, clearly suggest that the two 1980s units are not cogenetic to one another, a conclusion that could not be drawn from the bathymetric or isotopic data.

Relative to the 1980s Floc unit, the 1980s Flow unit exhibits markedly higher TiO2 and FeO* and lower CaO. In fact, as can be seen in Figure 2-3, TiO2 and CaO values for the 1980s Flow site unit appear to be rather anomalous compared to lavas from








the southern JdFR in general. Additionally, an examination of Table 2-3 reveals that, compared to the other recent CoAxial lavas that are both more and less geochemically differentiated, this basalt flow exhibits relatively high abundances in Zr, Hf, and to a lesser extent Nb. Elevated rare earth element (REE) abundances (relative to similarly differentiated samples from the segment) are also seen in samples from the 1980s Flow unit (Figure 2-8b, open circles). For example, REE patterns from the 1980s Flow unit are similar to, but have slightly higher chondrite-normalized abundances than those in either the 1993 Flow unit (Figure 2-8b, gray filled diamonds) or the 1980s Floc unit (Figure 28a). Abundances of the middle REE are about 20X chondrites, on average, in the 1980s Flow Site unit. That is approximately 30% higher than those from the Floc site unit (-1 5X chondrites) which has an average whole rock glass MgO content similar to that of the 1980s Flow unit. In addition to the elevated REE abundances, 1980s Flow samples have chondrite-normalized La/Sm (La/Sm) values (0.66-0.68) that are high relative to other Flow site lavas (0.48-0.62; n=22), but well within the CoAxial segment range in general (0.48-0.75, n=52).

Examination of the major and trace element data for the 1993 flow unit (Figures 2-6 and 2-8, Tables 2-1 and 2-3) reveal it to be the most depleted among the recent flows in incompatible elements, and among the most depleted along the entire CoAxial segment as well. While the 1993 flow is the most geochemically evolved (lowest MgO content) of the three recent CoAxial flows discussed, it has similar Na20 contents to the more primitive 1980's lavas and the lowest overall concentrations of the elements Zr, Nb, Hf, Sr, Rb, K20, U and REE. Correspondingly, the 1993 flow samples are among the most light-REE depleted MORB recovered from the CoAxial Segment axis, having La/Smn



























Figure 2-8: Chondrite-normalized (Taylor, 1982) rare earth element (REE) concentrations for MORB glasses from the 1980s Floc Site eruptive (upper figure; 6 samples analyzed), the 1980s Flow Site eruptive (lower fig.-open circles; 4 samples analyzed) and the 1993 Flow Site eruptive (lower fig.-gray triangles; 7 samples analyzed). All data were analyzed by ICP-MS at the Geological Survey of Canada with the exception of two wax core samples from the Floc Site (upper fig.-dashed lines) which was analyzed by ICMMS at the USGS (Denver). Note the y-axis is an expanded exponential scale ranging from 5X to 40X chondrites. La and Ce are anomalously high in two of the 1980s Flow Site samples (lower figure-dashed lines), and may be suspect. Excellent intra-flow consistency is seen in the REE data, but variability occurs between different flows. Samples from the 1993 flow have the lowest overall REE concentrations and La/Smn (0.56-0.59), despite being the most geochemically evolved.

















I I I I I I I I I I I I I I


10 F


I I I I I I I I I I I I I I


I - I I I I I I I I I I i


La Ce Pr Nd Sm Eu Gd Th Dy Ho Er Tm Yb Lu


Element


B









ratios between 0.56 and 0.59. In addition, as shown in Figure 2-4a, the 1993 eruptive also has the least radiogenic Sr isotopic signature of the three recent CoAxial lavas sampled. All of these observations suggest that the primary magma parental to the 1993 lava flow was more depleted in mantle magmatophile elements than other recent CoAxial lavas sampled.

2.6. Distinction of Regional Petrochemical Provinces


The comparison of minor and trace element characteristics of CoAxial lavas to those from other geologically identifiable terrains along the central JdFR suggests the CoAxial Segment is a magmatic province distinct from other segments of the JdFR. Further, examination of basalt geochemical data from the different geologically identified terrains allow us to test several hypotheses put forth by Embley et al. [in press] regarding their origin. In addition, these data can be used to investigate what relationship, if any, these central JdFR terrains have to the CoAxial Segment.

Firstly, it has been hypothesized that a highly reflective and unfractured area of seafloor south of -46�1 8'N along Axial Seamount's NRZ delineates the extent of Axial Seamount related volcanism (boundary is shown as a solid black outline in Figure 2-1). Secondly, Embley et al. [in press] propose that the ridge north of -46'18'N, termed the western fault block ridge (WFBR), is not tectonically or magmatically related to Axial Seamount as has been previously suggested [Sohn et al., 1998]. Direct comparison of the geochemistry of historical CoAxial flows to the geochemistry of lavas recently erupted within Axial Seamount caldera, lavas recently erupted along Axial Seamount's north and south rift zones (including the 1998 eruption detailed in Embley et al., 1998; Perfit et al.,









1998 and others) [Smith et al., 1997], and lavas recovered along the WFBR allow for the evaluation of these two hypotheses.

2.6.1 Axial Seamount Province

Samples recovered from Axial Seamount NRZ and the WFBR are limited to those recovered by wax core and one video guided grab (Figure 2-1). NRZ samples (all within the region of highly reflective and unfractured sea floor) include 4 samples from a very fresh eruptive unit(s), as well as 9 other apparently older samples (Figure 2-1, Tables 2-2, 2-3, 2-4). With the exception of two samples recovered from small seamounts discussed in section 2.6.3. all of the samples recovered from this region bear strong geochemical affinity to samples recovered from within the caldera (Figures 2-3, 2-4, 2-9 and 2-10). Similarly, only one sample recovered from the SRZ [Perfit et al., 1998] shows any geochemical distinction from lavas recovered from within the caldera, and it too was recovered from a small seamount. Isotopically, all of the NRZ and SRZ samples analyzed to date fall within the field defined by samples from Axial Seamount caldera (Figure 2-4). The geochemical similarities between fresh lavas recovered from within Axial Seamount caldera and those recovered from the north and south rift zones extend beyond the Sr isotopic data.

Lavas sampled from Axial Seamount's rift zones are indistinguishable from

caldera lavas with respect to the major elements (Figure 2-3), K20/TiO2 ratios (Figure 29), REE (Figure 2-10) and other trace elements. The ratios of elements with similar distribution coefficients are unaffected by crystallization processes. Accordingly, regional differences in primitive magma composition are well illustrated by plots showing the ratio of two highly incompatible elements against latitude (e.g. Figure 2-9).













o 1998 Axial Smt. + CoAxial Segment
9 Recent NRZ 0 Rogue Volcano
<> Axial N+S RZ 0 WFBR On-Axis Volcanoes


Axial Seamount Proper
15

9 + Recent Floc and > t> + Flow Site lavas
T

+4+
C+ ++0
#, ; +++ +o ,+ .
C + +
(+* ++ +


NRZ+SRZ * Volcanoes 0 '- End of NRZ
0- I I

45.5 46 46.5 Degrees Latitude North





Figure 2-9: Regional variation in EMP K2O/TiO2 shown plotted against latitude. Symbols are as in the legend, and data fields are as labeled. All data are electron microprobe analyses of natural MORB glasses. Maximum possible variation in the ratio x/y due to analytical precision (2-sigma) governed by the equation [(x+2ax)/(y-2oy)]-[(x2o,)/(y+2Oy)] equals 16% (shown for a K20TiO2= 10). The field for Axial Seamount caldera, data points for its SRZ and the lava erupted in within the caldera and SRZ in early 1998 (Perfit et al., 1998) are all based on 73 unpublished EMP analyses (Perfit et al., in prep). The "Axial Proper" field encompasses data from all but one of the samples reported in Rhodes et al., 1990. Vertical dashed line marks NRZ boundary discussed in text.




























Figure 2-10: (Upper) Chondrite-normalized (Taylor, 1982) REE patterns for regional basaltic lavas. Open diamonds show data for recent lavas from Axial Seamount caldera, and black-filled circles the recent NRZ flow discussed in text. Other symbols are as in Figure 2-9. Gray field marks the range of data shown in Figure 2-8 for the 1993 lava flow. Regional differences in LREE depletion are observed in lavas, with La/Sm ranging from -0.9 in Axial Seamount province samples to -0.3 in samples from Rogue Volcano. Samples from the WFBR (see text) have closer geochemical affinity to CoAxial Segment lavas than to lavas from Axial Seamount Province. (Lower) Covariation between EMP K20/TiO2 and ICP La/Sm, in regionally sampled lavas. Data from the Cobb-Eickelberg Seamount chain (Desonie and Duncan, 1990) are shown as open triangles. Strong positive correlation confirms that K20/TiO2 in sampled lavas preserves a mantle derived geochemical signal and it is not dominated by crustal processes.











40

�-20
-o

0 7







La Ce Pr Nd Sm Eu Gd Tb Dy HoEr Tm Yb Lu



2.0
A

1.5 A
1993 Flow A AA A A A
1.0 AA


A Axial Seamount
0.5 +Province


0.0 I I

0 10 20 30 K20/TiO2 100









Due to the limited number wax core samples analyzed for trace elements by either XRF or ICP techniques, the ratio K20/TiO2 (microprobe data) serves as the best available elemental parameter relatively unaffected by crystallization, to discriminate regional geochemical differences. Good regional correlation of K20/TiO2 with La/Sm,, Zr/Y and 87Sr/86 Sr (Figures 2-10, 2-11 and 2-12) supports the assumption that the geochemical signal seen in K20,/TiO2 variations is related to differences in mantle-derived magmatic composition, and not to subsequent shallow level processes such as fractional crystallization and/or crustal assimilation. K20/TiO2 values for Axial Seamount Province range from -0.1 to 0.17, and samples recovered from the north and south rift zones (right and left pointing triangles respectively) are indistinguishable from samples recovered from Axial Seamount caldera.

Chondrite normalized REE patterns are similarly indistinguishable between the rift zone lavas and those from Axial Seamount caldera (Figure 2-10a). REE patterns from four samples of the youngest north and south rift zone lavas are shown (filled circles) and completely overlap in shape and abundance with data from the three caldera lavas analyzed for REE (open diamonds). As with the NRZ lavas for which data is shown, data for caldera lavas plotted are from the youngest identified (visually) in the caldera and have MgO contents comparable to the rift zone lavas. Seven fresh looking samples were analyzed from the Axial Seamount Province, and have a total range in La/Smn of 0.88-0.92. Consideration of the three older rift zone lavas (not shown) analyzed for REE extends the range of La/Smn down to 0.8. All of these observations, combined with the distinct chemistry of lavas recovered outside of the zone defined by Embley et al. [in press] (discussed below), support the hypothesis that the highly





67










40

A
C) 30
30 A


O Crystalline Whole . 20 Rock Analysis

A
A .......::::::::: A
++
10 0

0 0 Axial Smt. Province
0~ ~ , 0 , . . ..

1 2 3 4 5 Zr/Y




Figure 2-11: Regional covariation of MORB K20/TiO2 (EMP) with ZrIY (XRF/ICP). Strong positive correlation suggests that the assertions made in Figure 2-10 hold true for trace element groups other than the REE. Symbols and fields are as in Figures 2-9 and 210. See text for further discussion.





























Figure 2-12: Correlation of the long lived radiogenic isotope 8Sr/Sr with both K20/TiO2 (upper) and Sr concentration (lower). Symbols are as in the legend. Fields for the Cleft Segment (grey field) and Axial Seamount Province (stippled field) are based on our unpublished data. Data for Axial Seamount samples analyzed for both isotopic and elemental abundances are limited, and, as such, the field shown encompases the full range of values for all samples in our Axial Seamount database. The range is similar to other published data, with the exception of 875r/ 86Sr values reported in Rhodes et al. (1990) and discussed in text. Calculated mixing curves (Langmuir et al., 1978) are shown (lower) for mixes between a CoAxial Segment lava (221-0551) and lavas from Axial Seamount caldera (dashed line) and Warwick Seamount (DH4-4-solid line). Symbols on mixing-curves mark 10% mixing increments.












Cleft


Smt. Province


A AA


+ CoAxial Segment + A C-E Smts.
o Rogue Volcano * Axial Province


K20/TiO2* 100


0.70260 Cleft Seg.


0.70250
0.020A A

" 0.70240 + Axial Smt. Province 0- +
0.70230 _]+ +

+
0.70220 + ' I 50 100 150 200 250 Sr ppm


300


0.70260 0.70250


0.70240


0.70230 0.70220


4#-









reflective and unfractured region of seafloor north of Axial Seamount caldera is blanketed with lavas that emanated from a magma source similar to the source that feeds Axial Seamount caldera.

2.6.2 The Western Fault Block Ridge

The topographic ridge extending north from NRZ (west of the CoAxial segment axial valley) becomes distinctly less acoustically reflective and more fissured north of approximately 46' 18' N latitude [Embley et al., in press]. This has been interpreted to reflect the greater age and sediment cover of the seafloor surface in this region, which apparently lacks recent volcanic overprinting of seafloor structures. Embley et al. [in press] conclude that the terrain between Axial caldera and 46' 18'N is related to recent constructional volcanism associated with NRZ volcanic activity, but north of -46' 18' they propose the topographic ridge is unrelated to Axial Seamount magmatism and refer to it as the western fault block ridge (WFBR).

Five lava samples recovered by wax core from the WFBR are, on average, more MgO-rich than Axial Seamount Province lavas. These WFBR lavas have MgO concentrations from 10.1-7.4 (ave=8.8). These samples (shown as open circles on Figure 2-9) have K20/TiO2 between 0.3 and 0.7 and fall outside of the field defined by Axial Seamount Province (K20/TiO2 = 0.1-0. 17). The low K2OiTiO2 values are more characteristic of the range defined by CoAxial Segment lavas (shown as crosses on Figure 2-9), for which 90% of 98 samples analyzed have a K20/TiO2 < 0.9. Two of the WFBR lavas (94JdFRC22 and 95JdFRC35) have been analyzed for trace elements, and these data confirm their more incompatible element depleted nature relative to Axial Seamount Province lavas (Table 2-4). The two WFBR lavas have La/Smn values of 0.38








and 0.48 respectively compared to recent Axial Seamount lavas that range from 0.880.92. Zr/Y values of less than 2.6 in these samples are much more characteristic of CoAxial Segment lavas which average 2.7 and have 90% of samples less than 3.0 (101 out of 11l analyzed). By comparison, Axial Seamount lavas have Zr/Y greater than 3 (Figure 2-11).

While the sample base for the WFBR is small, and the two samples analyzed for trace elements do not uniformly exhibit the Sr elemental depletions characteristic of CoAxial lavas, these lavas nonetheless seem to bear a much stronger geochemical affinity to those from the CoAxial Segment than to those from Axial Seamount Province. WFBR lavas' elementally depleted nature relative to Axial Seamount lavas supports the hypotheses that the WFBR is a feature unrelated to Axial Seamount magmatism.

Samples recovered from the ridge that bounds the eastern side of the CoAxial Segment over a similar latitude range as the WFBR, are limited to six basalt samples from one dredge haul (95DR2). Though the data are limited, this topographic feature also seems to have geochemical signatures similar to both the CoAxial Segment and WFBR. These samples have XRF measured whole rock glass Zr/Y values between 2.4 and 2.8 (average = 2.6) compared with an average value of 2.7 for the CoAxial segment lavas. XRF K2OTiO2 values are also similar to CoAxial, ranging between 0.054 and

0.096 (average = 0.077) compared with the average CoAxial value of 0.065.

2.6.3 Small NRZ Volcanoes

Two small seamounts within the NRZ terrain were sampled by rock (Figure 2-1, Tables 2-2 and 2-4, samples 94RC32 and 33) and both have very mafic compositions (greater than 9 wt.% MgO) and exhibit K2O/TiO2 values well below those measured in









lavas from Axial Seamount Province (Figure 2-9). A similar volcanic feature sampled from the Axial Seamount SRZ (Figure 2-9) also has similar mafic and depleted characteristics. Incompatible element depletions in these features appear not to be limited to the major elements. Sample 94RC32 was analyzed for trace elements (Table 2-4), and has both La/Smn and Zr/Y that are significantly lower than in lavas from Axial Seamount Province and comparable or lower then WRBF and CoAxial lavas (Figures 210 and 2-11). Two opposing hypotheses arise from consideration of the above observations and the data in Figures 2-9, 2-10 and 2-11. Firstly, these two small volcanoes may represent topographically high eruptive centers associated with the WFBR, and have survived subsequent burial by more recent volcanism along Axial Seamount NRZ. Alternatively, these two volcanic cones might have distinct magmatic plumbing systems similar to the many other small to medium sized near-ridge volcanoes that exist throughout the region (Figure 2-1, dark gray fill).

On the one hand, if the WFBR terrain is distinct from Axial Seamount Province as is suggested above and by Embley et al. [in press], then the presence of a volcanic feature geochemically similar to the NRZ volcanoes within the SRZ is difficult to explain via the first hypothesis. If, on the other hand, the alternate hypothesis is true and these small volcanoes have magmatic sources distinct from either Axial Seamount Province or the WFBR, then one might expect to see similar geochemical depletions in lavas from other near axis volcanoes in the region (Figure 2-1). Although there are numerous small near axis volcanoes in the CoAxial Segment region, particularly on the Pacific plate (Figure 2-1), only Rogue Volcano (on the JdF plate) has been sampled (Tables 2-2 and 23). Basalts analyzed from this volcano (6 for major elements, 2 for trace elements and Sr









isotopes) are all from one dredge (Figure 2-1). Similar to the small volcanoes sampled from the NRZ, lavas recovered from Rogue Volcano tend to be relatively MgO-rich and incompatible element depleted having average values for MgO (wt.%), Zr/Y, La/Smn and KzO/TiO2 equal to 9.1, 1.5, 0.33 and .058 respectively.

The association of near-axis volcanoes with unusually mafic and depleted basalt compositions has been previously recognized on the JdFR [Leybourne and Van Wagoner, 1991; Finney, 1989] and other regions of the MOR system [e.g., Fornari et al., 1988; Allan et al.]. In particular, lavas sampled from the Heck and Heckle Seamount chains (located west of the Endeavor Offset -225 km north of the Rogue and NRZ volcanoes) have very similar geochemical characteristics to the lavas recovered from Rogue and the NRZ volcanoes. Similar to the samples discussed above, basaltic glasses recovered from the Heck and Heckle seamount chains [Leybourne and Van Wagoner, 1991] tend to be mafic, averaging greater than 8.4wt% MgO and reaching concentrations as high as 9.3 wt.%. Heck and Heckle lavas are also depleted in incompatible elements, having Zr/Y less than 2.4, K20/TiO2 less than 6 and La/Smn less than 0.42. The above elemental considerations support a hypothesis whereby the small NRZ seamounts result from magmatic activity unrelated to either Axial Seamount or the WFBR, possibly having origins similar to other near-axis volcanoes in the central JdFR region.

2.6.4 The CoAxial Segment

CoAxial Segment lavas are perhaps most uniquely characterized by their depleted strontium abundances and isotopic ratios relative to other axial JdFR basalts (Figures 2-4 and 2-5). CoAxial segment lavas also tend to be quite incompatible element-depleted having Zr/Y, K20/TiO2 and La/Sm ratios ranging from 2.4-3.2, 0.03-0.1, and 0.48-0.75









and averaging 2.7, 0.07 and 0.6 respectively. Lavas from the 1980s Floc unit and the Source site (though not well represented in our sample suite) tend to be the least LREE depleted of any CoAxial lavas analyzed, consistently having the highest La/Smn (greater than 0.68). These samples also tend to have among the most radiogenic Sr isotopic signatures observed for the segment. With the exception of the 1980s Flow unit discussed in detail in section 5.2.2, samples from the flow site tend to be the most radiogenically and elementally depleted along the segment. This effect is exemplified by a trend of generally decreasing K20/TiO2 with increasing latitude along the segment (Figure 2-9). Although CoAxial lavas are elemental-depleted, their compositions do not display the extreme of elemental depletions evident in the "near-axis" volcanoes discussed above (Figures 2-9, 2-10 and 2-11).


2.7. Discussion


2.7.1 Recent CoAxial Segment Volcanism

The volcanic eruption that occurred on the CoAxial Segment in 1993 at the Flow Site was a phenomenon unique in the study of mid-ocean ridges because it was an actively monitored crustal accretion event [Fox et al., 1995]. Real time seismic monitoring of the JdFR recorded the diking event that fed the 1993 eruption. The locations and migration of the epicenters of T-phase seismicity recorded during the event led to the initial conclusion that the dike intruded along strike and likely originated from the north rift zone of Axial Volcano [Dziak et al., 1995]. Although it was recognized that topographic effects could have caused a westward bias in the locations of epicenters, Axial Seamount could not be ruled out as a source for the eruption. Subsequent studies of post-emplacement seismicity on the CoAxial Segment concluded that the dike path









was along the CoAxial spreading center axis, and suggested that T-phase epicenters associated with the 1993 seismic event were mislocated [Sohn et al., 1998]. The documented geochemical differences between Axial Seamount Province lavas and CoAxial Segment lavas in general, and the 1993 flow specifically, make it highly unlikely that Axial Seamount served as the source for the 1993 dike. The geochemical distinction of the two regions is consistently seen in the major element, trace element and isotopic characteristics of lavas from each region. A "CoAxial" origin for the 1993 eruption is further supported by geologic and hydrologic observations [Embley et al., in press], making a compelling argument against Axial Seamount as a source for the 1993 event.

Two other eruptive units have been emplaced along the CoAxial Segment since 1981 [Embley et al., in press]. Differences in the major and trace element geochemistry between these two flows strongly suggest that lavas from these two units are neither comagmatic nor are they cogenetic. This leads to the conclusion that at least three separate eruptive events have occurred along this ridge segment during a span of twelve years or less. This is the first such area of the MOR system where such detailed constraints on eruptive frequency and distribution have been established, allowing for a more detailed investigation of the temporal and spatial aspects of crustal accretion at this medium rate spreading center than would otherwise be possible. Relative location of the three lava flows establishes the "neovolcanic zone" for this ridge segment to be greater than 1 km wide at the decadal time scale and over lengths of less than 25 km along strike.

The eruption of the 1980s Floc unit along a preexisting constructional ridge

[Embley et al., in press] and the in situ sampling of relatively fresh and geochemically









distinct lava from a kipuka in the 1980s Floc unit (sample 2948-10R) is intriguing. The presence of relatively fresh, non comagmatic lavas comprising the pre-existing ridge suggests that at least two eruptions have occurred at this location in the recent past (less than 100's of years), some 500m west of the interpreted ridge-axis and locus of current geologic and hydrothermal activity. Accordingly, although CoAxial eruptive activity might not be as well focussed as in some other regions of the Juan de Fuca Ridge and northern East Pacific Rise studied in comparable detail [Embley and Chadwick, 1994; Fornari et el., 1998], there is likely some control of preexisting structure on the location of dike emplacement along this segment. This is consistent with the observation that the 1993 dike was emplaced along the trend of a preexisting fissure system that was apparently reactivated during the intrusive event.

Perhaps as intriguing as the temporal/spatial variation in lava emplacement is the variation observed in lava chemistry. None of the flows which have been emplaced within the CoAxial Segment axis over a twelve year period appear to be cogenetic and each has a genetic origin uniquely distinguished from the other by elemental concentrations and ratios and/or Sr isotopic ratios. In addition, it has been noted that samples recovered from the CoAxial Segment axis show a greater dispersion in elemental concentrations at a given MgO value than do other Segments of the southern JdFR. These observations suggest that there is not a well-mixed, long-lived magma chamber serving as the source for CoAxial Segment volcanism. Inter-flow geochemical variability is more suggestive that the magmatic plumbing underneath the CoAxial Segment is poorly developed, with individual melt bodies being relatively small and short lived. In fact, differences in mantle derived geochemical signals observed between the two recent









units at the Flow site (Sr and Nd isotopes and the ratios of elements highly incompatible with mantle mineral assemblages) show that mantle geochemical heterogeneity can be preserved by seafloor crustal accretion here on temporal and spatial scales of a decade and 500m.

The hypothesis that CoAxial Segment volcanic activity (at least that part north of the poorly geochemically characterized Source site) is not related to a large, long-lived, controlling magma chamber is consistent with the observed geology and geophysics. Seismic data for the upper three km of the crust show no evidence of a low velocity zone associated with an axial magma chamber reflector [Sohn et al., 1997], and seafloor compliance measurements conducted on the segment failed to detect any magma bodies in the crust that comprises the CoAxial Segment axial valley [Crawford, 1994]. Additionally, Sohn et al. [ 1997] have interpreted the crustal seismic structure of the northern portion of the CoAxial Segment in terms of a dominantly amagmatic extensional terrain.

The CoAxial Segment's morphology and structure is not like that of some

magmatically robust segments such as the Cleft segment to its south [Embley et el., 1991] or the N. East Pacific Rise (EPR) [Haymon et al., 1993]. Along much of the Cleft Segment and the 9�-10�N segment of the EPR, axial volcanism appears to be dominantly focussed within a relatively narrow region. These areas, presumably underlain by well developed magma chambers [Smith et al., 1994; Fornari et al., 1998], are associated with phenomenon such as high temperature hydrothermal venting, formation of lava lakes and large regions of volcanic collapse [Embley et al., 1991; Haymon et al., 1993; Fornari et al., 1998; Kent et al., 1993]. None of these features have been observed in the northern









portion (Floc and Flow Sites) of the CoAxial Segment. CoAxial Segment volcanism occurs over a zone greater than 1 km wide, and significant portions of the CoAxial terrain north of the Source Site are highly fissured giving the impression that amagmatic extension has played an important role in the development of axial structure, despite the evidence for recent volcanic activity.

The spatial patterns of recent volcanism and abundant fissures within the

neovolcanic zone of the northern CoAxial segment are consistent with seismic data that suggest no shallow magma chamber underlies this portion of the segment. These observations are consistent with a southerly source for recent volcanism, and support the hypothesized lateral propagation of the 1993 dike, thought to have originated south of the Floc Site.

2.7.2 Intra-flow Geochemical Variability.

Although no direct evidence for crustal magma bodies exists, it is clear that the recently erupted CoAxial lava flows have undergone substantial differentiation prior to eruption. One situation where chemical fractionation could have taken place is during the lateral emplacement of the dike through the crust. In the case of the 1993 eruption, the dike is thought to have migrated some 40 km along axis [Dziak et al., 1995] over a period of -2 days. While evidence of intra-flow fractional crystallization in the 1980s Floc unit along the strike of its emplacement (Figure 2-7) is intriguing, a closer examination of major element data shows this effect to be insufficient to generate the observed fractionated compositions of the 1993 lava flow. Basaltic glasses show a nearly 1 wt.% variation in MgO over the 7 km long axis of the flow. This change in MgO corresponds to -20% fractional crystallization of the most mafic portion of that flow based on LLD








calculations. On the surface then, it would seem that a fractional crystallization rate of

-3% per km would be sufficient, over a lateral emplacement distance of -40km, to account for the approximately 75% fractional crystallization required to differentiate a moderately primitive MORB magma (-9.5wt% MgO) to the composition of the 1993 Flow (-6.7wt% MgO). Two observations argue that chemical differentiation of the 1993 flow could not have entirely occurred during dike intrusion and eruption.

Firstly, comparison of the intra-flow whole rock glass data to the EMP data for the 1980s Floc unit shows that the liquid differentiation process was no where near as efficient (by -50%) as pure fractional crystallization predicts. In addition, subaerial analogues tot he 1993 diking event do not exhibit this degree of magmatic differentiation along the path of the dike injection. This is not surprising, as it is difficult to envision efficient separation of crystals from liquid in the dynamic environment of a dike emplacement that occurs over the course of only a few days. Nonetheless, some bulk chemical differentiation does appear to have occurred along the 7 km strike of the 1980s Floc Site flow.

Secondly, the petrography of samples from the 1993 flow suggests that the magma likely had some residence in a crustal magma body prior to dike intrusion. Glassy rims of the 1993 lava samples tend to be quite crystalline (visual estimates of up to 10-20 modal percent). Much of the crystal assemblage consists of small, quenched microphenocrysts of plagioclase and clinopyroxene consistent with a phase of rapid cooling and crystallization prior to eruption. There are, however, a second population of larger phenocrysts (up to 2 mm) of olivine, clinopyroxene and plagioclase, some of which show signs of resorption and chemical disequilibrium. It is unlikely that this









second population of larger grains could have formed during the time frame of the intrusion event, and it is more likely that they crystallized from a cooling crustal magma body. This conclusion in no way, however, detracts from the importance of the intraflow chemical variation that does exist within the 1980s Floc unit, but does infer the presence of a CoAxial crustal magma body that has yet to be imaged.

Recently a positive correlation between intra-flow chemical variability and

eruption volume has been suggested [Rubin et al., 1998; Perfit and Chadwick, 1998]. Additionally, others have suggested an inverse correlation between spreading rate and eruptive flow volume [Sinton, 1997; Perfit and Chadwick, 1998]. The strong correlation of glass geochemistry along the strike of emplacement for the 1980s Floc unit, and to a lesser extent the 1993 Flow unit, may imply a causal mechanism relating these two observations. Intra-flow geochemical variation observed in the three recent CoAxial lava flows are all consistent with small amounts of magmatic differentiation along low pressure LLDs. This is most easily explained via small degrees of cooling and crystallization occurring during lava emplacement. This hypothesis is supported by the petrography of the 1993 CoAxial flow which has a large proportion of microlites and microphenocrysts in the outer most glassy selvage of the lava flow, a characteristic that is likely related to a rapid phase of cooling prior to eruption. The elongate lineated nature of the 1980s Floc unit and 1993 Flow unit, their systematic intra-flow geochemical variation along ridge strike, and the observation of lateral dike propagation during the 1993 event all suggest that the 1980s Floc unit was emplaced via a similar south to north propagating dike. Does the process of lateral propagation of intruding (and eventually erupting) dikes in the shallow ocean crust aid in the generation of intra-flow geochemical









heterogeneity? These data would support that conclusion. Is a dike more likely to propagate long distances laterally in a slower spreading environment? Intuitively it would seem that an increasingly tensile stress regime would promote such propagation, but factors such as ridge-axis structure, magma density, eruptive pressure and topographic considerations are all factors to be considered. It is intriguing that several investigations of JdFR volcanism have led to conclude that lateral dike propagation served as the mechanism of lava emplacement [Embley et al., 1991 and in press]. Conversely, investigations of the northern East Pacific Rise have concluded that recent eruptions were emplaced through mostly vertical dike intrusion [Wright et al, 1995].

2.7.3 Regional Petrochemical Characteristics

This study recognizes four petrochemically distinct magmatic provinces in the Axial Seamount-CoAxial Segment region. Axial Seamount Province and its south and north rift zones comprise one. Another is comprised of the western fault block ridge. The third is the Coaxial Segment, and the last is comprised of several small axial and near axis volcanoes.

2.7.3.1 Axial Seamount

Samples recovered from Axial Seamount north and south rift zones are generally geochemically indistinguishable from lavas recovered from within the caldera. Data presented here support the boundaries of NRZ volcanism proposed in Embley et al. [in press]. Axial Seamount Province lavas are the least geochemically depleted of the four provinces discussed. They are enriched in potassium and sodium relative to other southern and central JdFR lavas (Figure 2-3), and compared to the other three aforementioned magmatic provinces, Axial Seamount lavas are less depleted in the light









REE (La/Sm, of 0.88-0.92), Sr (Figure 2-5), 87Sr/ 86Sr, and elemental ratios such as Zr/Y and K20/TiO2. Each of these observations is indicative of geochemical enrichment in mantle-incompatible elements in Axial Seamount lavas relative to CoAxial or WFBR lavas. Lavas from Axial Seamount are not, however, obviously enriched in isotopic and incompatible elements relative to all JdFR lavas south of the Cobb offset [Rhodes et al., 1990; Smith et al., 1994]. Nonetheless, they do generally have higher values of Zr/Y, La/Sm and Sr/Zr, and, as such, the Axial data only partly overlap the geochemically enriched end of the JdFR data-field for each of these parameters [Rhodes et al., 1990]. Previous authors have modeled enrichments in Sr, Na and Ca observed in Axial Seamount lavas to be the result of a greater depth of initial melting of a source similar to that of the ridge-axis MORB [Rhodes et al., 1990]. The more depleted nature of the CoAxial Segment and WFBR provides further constraints on the degree of regional mantle geochemical heterogeneity and the depleted mantle end-member.

2.7.3.2 The Coaxial Segment

While a detailed petrogenetic analysis is beyond the intended scope of this paper, and will be dealt with elsewhere [Smith et al., in prep], a few general observations are relevant to the discussion presented here. Qualitatively, depletions in A1203 and Na2O in the CoAxial MORB suite relative to the Axial Seamount lavas, coupled with enrichments in FeO, could be explained via increased degree of melting of an "Axial Seamount" melt source [Klein and Langmuir, 1987 and 1989, Langmuir et al., 1992]. This hypothesis is not, however, consistent with the depletion of CaO in CoAxial MORB relative to Axial Seamount lavas. In fact, depletion in CaO of CoAxial lavas relative to Axial Seamount lavas is supportive of smaller degrees of melting, unless melting proceeds well past the









elimination of CPX as a residual phase in the source mantle [Klein and Langmuir, 1987]. Additionally, the inverse correlation between Na2O and FeO contents between the two regions corresponds to the trend associated with the "global" variability vector of Klein and Langmuir [1987, 1989], while the inverse correlation between NaO and SiO, would be more easily explained via the "local trend vector" and sampling of instantaneous melts from different depths in a melting column.

The inability to consistently account for differences in the suite of major element oxides observed between Axial Seamount and CoAxial Segment lavas via differences in the extent and depth of melting (i.e. via either the local or global trends of Klein and Langmuir, 1989) argues against the hypothesis that differences in major element characteristics between the MORB from the two regions are due to different conditions of melting of an "Axial Seamount" source in the generation CoAxial lavas.

As in the above arguments, comparison of CoAxial MORB data to the Cleft Segment MORB data also suggests that differences in major element lava chemistry between the two regions cannot be simply explained by different melting conditions of a similar and homogenous source. Lower relative Na2O abundances in Cleft segment lavas would require a higher extent of melting relative to CoAxial lavas, but higher A1203 contents suggest a lower extent of melting. Clearly these two observations are incompatible with the above hypothesis. In summary, regional differences in major element chemistry of CoAxial lavas relative to other southern Juan de Fuca lavas cannot be generated by melting of a homogeneous mantle, but must instead be related to mantle chemical heterogeneity and or differences in crystallization and differentiation processes.









Additionally, trace element and isotopic constraints prohibit the generation of CoAxial lavas through melting of a homogeneous "Axial Seamount" source mantle.

An alternate hypothesis to that discussed above (differences in primitive melt composition relating to differences in the extent and depth of melting of similar mantle compositions) is that the source mantle sampled by Coaxial lavas is different from the other two regions either by large scale mantle heterogeneity or finer scale heterogeneity sampled differentially through variable conditions of melting. CoAxial Segment lavas exhibit characteristics of long term incompatible element depletion (radiogenically depleted 87Sr/86 Sr and enriched 143Nd/144Nd). Regional correlation between incompatible element ratios (Figures 2-9, 2-10 and 2-11), and between long-lived isotopes of strontium and incompatible element ratios (Figure 2-12), indicate that the range of mantle source characteristics observed can be accounted by mixing of very depleted and slightly enriched components.

Figures 2-10, 2-11 and 2-12 show regional correlations between samples from the CoAxial Segment, Axial Seamount Province, Rogue and the NRZ volcanoes, the WFBR and the Cobb-Eickelberg Seamount Chain, of which Axial Seamount is the youngest member [Desonie and Duncan, 1990]. All three diagrams show generally hyperbolic trends characteristic of mixing curves on ratio-ratio diagrams [Langmuir et al., 1978]. Figure 2-12 shows mixing of 87Sr/86Sr with both K20/TiO2 (upper) and Sr (lower). Calculated trends are shown for mixes between a depleted CoAxial lava and lavas from Axial Seamount caldera (dashed curve) and Warwick Seamount, a member of the CobbEikelberg chain (solid line). Trend symbols mark 10% mixing-increments. Examination of Figure 2-12 shows the CoAxial data trend to be elongated in the direction of the









mixing curves, with the most enriched CoAxial samples showing up to a 40% mix of "Axial Seamount melt component" or a 20% mix of a "Warwick Seamount melt component" with the depleted CoAxial end-member. Axial Seamount lavas themselves fall in the range of 40% mixing of the enriched "Warwick Seamount" end member with the depleted CoAxial end-member. While it is not the intention or within the scope of this paper to model in detail exact nature of heterogeneities occurring in the upper mantle, some interesting observations and hypothesis arise from the examination of Figures 2-10, 2-11 and 2-12.

2.7.4 Regional Mantle Heterogeneity and the Origins of CoAxial Segment
Geochemical Depletion.

Regional correlation of MORB Sr isotopic values with incompatible trace element abundances implies some antiquity to mantle geochemical heterogeneity in this region. While some aspect of this heterogeneity must be related to long-term depletion events, significant variability of incompatible elements/ratios about the mixing curves at given values of 87Sr/86 Sr implies there has also been more recent enrichments/depletions (Figure 2-12). This regional mantle heterogeneity is likely the result of recent (relative to the half-life of 87Rb) upper mantle melting, but the cause(s) of the more ancient heterogeneity is more difficult to assess [Smith et al., in prep]. Distinct Isotopic depletions observed in CoAxial lavas relative to all other Juan de Fuca MORB (Figure 24) imply a depletion of ancient mantle geochemical enrichments that is unique on the Juan de Fuca Ridge. Two hypotheses are recognized to explain these depletions.

The first hypothesis is that melting around the outer portions of the Cobb melting anomaly has left "CoAxial Segment" upper mantle depleted in enriched source component. Evidence suggests that much of the northern CoAxial Segment appears









relatively amagmatic, and that the sources of recent eruptions probably have an origin south of the Floc Site. If this were in fact the case, and CoAxial lavas were being depleted by previous source melting due to the Cobb melt anomaly, then one might expect source depletions to show a radial pattern about the melt anomaly. Little evidence exists for comparably depleted lavas south of the Cobb anomaly, although sampling in this region is currently insufficient to truly test this hypothesis.

An alternate hypothesis is that recent rapid north westward migration of the ridgeaxis in this area (> 35km in the past 0.5 Ma and up to 100 km in the past 7 Ma) as the ridge has approached the Cobb melting anomaly [Karsten and Delaney, 1989] has placed the current ridge-axis above a region of upper mantle already depleted through the process of MORB magma genesis. This hypothesis of rapid ridge-axis migration was called upon by Karsten and Delaney [1989] to explain the asymmetric distribution of near-axis seamounts occurring on the Pacific plate (Figure 2-1). If both the asymmetric seamount distribution and CoAxial Segment source depletions are due to the rapid northwestward migration of the ridge in this area, then one might expect two conditions to be true. Firstly, one might expect similar depletions in Vance segment just south of the melt anomaly, providing it was equally affected by ridge migration. Although existing data is sparse, this does not appear to be true. Secondly, one might expect that seamounts on the Pacific plate outside of the current ridge-axis zone of melting to have, on average, geochemical signatures more enriched than either the ridge-axis MORB or the near-axis volcanoes to their south east.

Whether the depleted character of CoAxial lavas is the result of effects associated with the Cobb melting anomaly, ridge-axis migration, or some other phenomenon, it is









likely that CoAxial lavas were produce by a mantle previously "conditioned" by at least small amounts of previous melt extraction. It is hypothesized that this previous melting preferentially sampled whatever enriched mantle component existed in the JdFR MORB source giving CoAxial lavas a distinct geochemical character relative to other JdFR lavas. Broad scale regional correlation of source characteristics and finer scale variation within individual magmatic provinces suggests that the spatial scale of mantle heterogeneity in this region is smaller than the individual magmatic provinces discussed.

2.8. Conclusions


An integrated approach to the study of MORs involving the use of geophysical, geologic and geochemical mapping methods is necessary to address, in detail, questions of spatial and temporal variability in MOR crustal accretion. At least three different eruptive events have occurred on the CoAxial Segment over a span of less than 12 years. Data presented above leads to several important conclusions regarding the process of crustal accretion at the CoAxial Segment and the spatial association of distinct melting regimes on the central JdFR in general. These conclusions include the following:

(1) Spatial relations of these units establish a neovolcanic zone greater than 1 km wide over a decadal time scale.

(2) Recent CoAxial Lavas are geochemically distinct from recent Axial Seamount lavas

and are believed to have originated from within the CoAxial Segment, not from

Axial Seamount as has been previously suggested.

(3) Inter-flow chemical variability in CoAxial lavas suggests that the CoAxial Segment

has a poorly developed magmatic plumbing system, in accord with the lack of any geophysically identified axial magma chamber within the upper 3 km of the crust.









(4) Intra-flow geochemical variability of up to I wt.% MgO is observed in CoAxial

lavas, and is best explained by small amounts of crystallization and differentiation

along low pressure liquid lines of decent.

(5) Four distinct magmatic provinces have been identified in the Axial SeamountCoAxial Segment region. The range of mantle source characteristics is consistent

with mixing trends established using "source" characteristics of variably

depleted/enriched lavas recovered from within the region.

(6) Correlation between long lived Sr isotopes and incompatible element ratios imply
that some aspect of local mantle heterogeneity has antiquity, but scatter of
incompatible element ratios at a given isotopic value further suggest that a more
recent component of elemental enrichment/depletion contributes to the geochemical
heterogeneity of the upper mantle in this region.















CHAPTER 3
SUBMARINE INVESTIGATIONS OF A THIRD-ORDER OSC AT 9 37' N:
ESTABLISHING A CAUSE AND EFFECT RELATIONSHIP BETWEEN OSC
PROPOGATION AND MAGMATIC ACTIVITY


3.1. Introduction


The origins, development, and causal mechanisms of disruptions in the continuity and linearity of the East Pacific Rise (EPR) ridge-axis have been the subject of much investigation, discussion, and speculation in recent years [e.g. Macdonald and Fox, 1983 and 1988; Macdonald et al., 1984, 1986, 1987, 1988; Lonsdale, 1983 and 1986; Langmuir et al., 1986; Toomey et al., 1990; Perram and Macdonald, 1990; Haymon et al., 1991; Carbotte and Macdonald, 1992; Kent et al, 1993a, 1993b, Carbotte et al., in press, etc.]. Four different "orders" of discontinuity have been recognized on the basis of their spatial dimensions [Macdonald et al., 1988b], subdividing the northern EPR into discrete segments of varying length and temporal continuity [Macdonald and Fox, 1983; Macdonald et al., 1984; Langmuir et al., 1986; Perram and Macdonald, 1990; Toomey et al., 1990; Haymon et al., 1991; Carbotte and Macdonald 1992; Macdonald et al., 1992; etc.].

Several criteria have been used to define first through fourth-order ridge segments and associated first through fourth-order discontinuities that bound and define them [Macdonald et al., 1988b]. First-order spreading segments are tectonically defined and bounded by large transform offsets or propagating rifts whose offsets are such that plate boundaries behave rigidly, juxtaposing crust of greater than approximately 0.5-1 Myr age









difference across the offset boundary. Offsets are typically greater than 50 km and partition the ridge-axis at intervals of 100-1000 km. Second-order segments have a length scale of 50-300 km and are bounded by smaller, non-rigid discontinuities such as large overlapping spreading centers (OSC) with an offset greater than 3-5 km or small offset (less than 20 km) non-rigid transform faults. Third-order segmentation divides the ridge-axis into characteristic lengths of 30-100 km, and the segments are bounded by smaller (0.5 to 3-5 km) OSCs characterized by small (10's of meters) increases in axial depths. Lastly, fourth-order segments have typical length scales of 10-50 km. The fourth-order discontinuities are not typically associated with axial depth anomalies, and the segment boundaries are defined by small non-overlapping offsets of the ridge-axis and/or small (1-5�) bends in the ridge-axis, referred to as deviations in axial linearity (DEVALS) [Langmuir et al., 1986]. Although first-order segments and discontinuities are thought to be persistent on the order of millions of years, smaller ridge offsets may not be as temporally stable [Macdonald et al., 1988b]. The tendency of smaller ridgeaxis discontinuities to migrate and evolve has led to much speculation regarding their causative mechanism(s).

While in some instances far-field plate stresses and/or changes in the direction of spreading have been identified as the driving mechanism for migration of second and third-order discontinuities [Hey et al., 1980, 1982, 1988; Carbotte and Macdonald, 1992], magmatic activity (or lack there of) has also been postulated as the active mechanism for the formation and evolution of similar features [Lonsdale, 1983; Sempere et al., 1984; Macdonald et al., 1984, Langmuir, 1986; Mutter et al., 1988; Toomey et al., 1990 and 1994; Smith et al., 1998; etc.]. The association between magmatic activity and fine-scale









ridge segmentation has been largely based on correlation between the location of discontinuities and factors such as axial depth, continuity and depth of the seismically imaged axial magma chamber (AMC), magnetic, seismic and structural characteristics of the crust, and the geochemistry of mid-ocean ridge basalt (MORB) dredged from either side of offsets.

While many such discontinuities in the eastern Pacific spreading axis have been imaged bathymetrically [e.g. Macdonald et al., 1984], seismically [Toomey et al., 1990 and 1994, Kent et al., 1993a, Kent et al, 1993b; Carbotte et al, in press], and magnetically [Sempere et al., 1984], relatively few have been optically imaged or petrologically sampled [Langmuir et al., 1986; Sinton et al., 1991; Batiza and Niu, 1992]. Those that have been sampled for basalt geochemistry have been only sparsely sampled by dredge, without having the sample density or positional accuracy to reconstruct in detail the geometry of geochemical variation across the ridge-axis discontinuity. Here the first detailed report of in situ observational and geochemical data recovered by submersible at a small third-order discontinuity is presented. Submersible data is supplemented by dense wax core sampling of basalt glass from the crestal region surrounding the OSC.

3.2. Previous Studies and Regional Geology


3.2.1 Previous Studies

The study area lies between -9' 35' and 90 37' N latitude, within a second-order spreading segment of the fast spreading EPR bounded to the south at -90 03' N by an OSC and to the north at - 100 10' N by the intersection of the ridge-axis with the Clipperton transform fault (Figure 3-1) [Haymon et al., 1991]. This second-order segment spreads at approximately 1 l cm/yr full-rate [Klitgord and Mammerickx, 1982],
































Figure 3-1: Regional bathymetric map of the second-order EPR segment between -9' 03' N and I 0N latitude. Regional map-inset and scale are shown at left. Entire figure is after Haymon et al. [ 1991]. Contour interval for rise-crest bathymetry is 20 m and columns to right to show fourth-order segment boundaries and geologic, hydrothermal and biologic observations based on ARGO deep tow data gathered in 1989.















104"30'W 9*50'N


From Haymon et al., 1991 Axial Summit Caldera
(ASC)

C3 Dive Areas

0 1991 Eruption Area

0 5 10
kms




Full Text

PAGE 1

GEOCHEMISTRY OF EASTERN PACIFIC MORE: IMPLICATIONS FOR MORE PETROGENESIS AND THE NATURE OF CRUSTAL ACCRETION WITHIN THE NEOVOLCANIC ZONE OF TWO RECENTLY ACTIVE RIDGE SEGMENTS By MATTHEW C. SMITH A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1999

PAGE 2

ACKNOWLEDGMENTS There are many that have contributed to this research and deserve aclcnowledgement. I would like to thank Dr. Michael Perfit for helping to put me in a position to take advantage of numerous opportunities to participate on the fascinating oceanographic cruises that have led to this research. Without his contacts, guidance, mentorship, and friendship, none of this would have been possible. Drs. Robert Embley and Daniel Fomari served as chief scientists on many of the cruises to the Juan de Fuca and East Pacific Rise respectively. Their example of how to be an effective sea going scientist made a lasting impression on me, and this research benefited from many discussions of the structure and volcanic morphology of eastern Pacific spreading centers. The crews of the ALVIN, Atlantis II, ROPOS and Discoverer deserve thanks for their willingness to pursue the excellence we required, even in the wee hours of the morning. Dr. William Chadwick graciously prepared the sample location maps in chapter 2 and provided AMS -60 data to aid in interpretation of regional geologic associations. H. Paul Johnson generously donated splits of basalt samples collected during three cruises to the CoAxial Segment. Dr. Ann Heatherington assisted greatly in the isotopic analyses and I thank her. Drs. Tim O'Heam and W. I. Ridley analyzed data from the Smithsonian and USGS respectively, and Dr. Ian Jonasson provided ICP data from the GSC at no cost to this project. I owe them all a debt of gratitude. Dr. Karen Von Damm, Dr. John Lupton and Dr. Marvin Lilley graciously provided unpublished data for vent fluids discussed in chapter 3. D. John Chadwick also graciously provided his unpublished data for basalts ii

PAGE 3

from Axial Seamount south rift zone used in chapter 2 discussions. Lastly, I would like to thank Robin Milam for her assistance in the formatting and preparation of this manuscript, and my family for their continual love and support. iii

PAGE 4

TABLE OF CONTENTS page ACKNOWLEDGMENTS ii ABSTRACT vi CHAPTERS 1 . GENERAL INTRODUCTION 1 1 . 1 The CoAxial Segment 7 1 .2 The East Pacific Rise between 9 and 10°N 8 2. GEOCHEMICAL CHARACTERISTICS AND PETROGENETIC RELATIONSHIPS OF RECENTLY ERUPTED MORE FROM THE CENTRAL JUAN DE FUCA RIDGE: SEGMENT SCALE, INTERFLOW AND INTRAFLOW VARIATIONS 11 2.1. Introduction 11 2.1.1 Field Investigations and Previous Studies 1 1 2.1.2 Regional Geology 12 2.2. Data Collection and Sample Distribution 16 2.2. 1 Alvin and ROV Based Sample Acquisition 16 2.2.2 Surface Based Sample Acquisition 16 2.2.3 Sidescan Sonar and Near Bottom Photographic Data Acquisition 20 2.3. Analytical Methods 20 2.3. 1 Sample Preparation 20 2.3.2 Major Element Analyses 22 2.3.3 Trace Element Analyses 26 2.3.4. Isotopic Analyses 33 2.4. Regional CoAxial Segment Geochemical Characteristics 33 2.4. 1 Major Element Trends 33 2.4.2 Isotopic and Trace Element Signals 40 2.5. Discrimination of Discrete Flow Units 46 2.5. 1 Visual and Bathymetric Flow Discrimination 46 2.5.2 Geochemical Flow Discrimination 48 2.6. Distinction of Regional Petrochemical Provinces 61 2.6. 1 Axial Seamount Province 62 2.6.2 The Western Fault Block Ridge 70 2.6.3 Small NRZ Volcanoes 7 1 iv

PAGE 5

2.6.4 The CoAxial Segment 73 2.7. Discussion 74 2.7. 1 Recent CoAxial Segment Volcanism 74 2.7.2 Intra-flow Geochemical Variability 78 2.7.3 Regional Petrochemical Characteristics 81 2.7.4 Regional Mantle Heterogeneity and the Origins of CoAxial Segment Geochemical Depletion 85 2.8. Conclusions 87 3. SUBMARINE INVESTIGATIONS OF A THIRD-ORDER OSC A 9° 37' N: ESTABLISHING A CAUSE AND EFFECT RELATIONSHIP BETWEEN OSC PROPAGATION AND MAGMATIC ACTIVITY 89 3.1. Introduction 89 3.2. Previous Studies and Regional Geology 91 3.2.1 Previous Studies 91 3.2.2 General Second-Order Scale Observations 99 3.2.3 Previous Geologic Observations of the EPR Axis Between 9° 49.9' N and 9° 37.1' N 102 3.3. Data Acquisition and Alvin Observations 104 3.3.1 In Situ Geologic Observations 104 3.3.2 Hydrothermal and Biologic Activity 109 3.4. Basalt Geochemistry 1 1 1 3.4. 1 Basalt Sample Recovery and Analysis Procedures Ill 3.4.2 Second-Order Segment Scale Geochemical Observations 118 3.4.3 Fourth-Order Segment Scale Geochemical Observations of Segment D 130 3.5. Discussion 131 3.5.1 Southward Aging of Volcanic Activity 131 3.5.2 The 9° 37' OSC as a Magmatic and Hydrothermal Divide 133 3.5.3 Evidence for a Magmatic Perturbation at 9° 37' 136 3.5.4 Association of E-Type MORB 138 3.6. Conclusions 144 4. CONCLUSIONS 147 APPENDIX FREQUENTLY USED ACRONYMS AND DESCRIPTIONS OF SAMPLING TECHNIQUES 151 REFERENCES I53 BIOGRAPHICAL SKETCH 164 V

PAGE 6

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Doctor of Philosophy GEOCHEMISTRY OF EASTERN PACIFIC MORE: IMPLICATIONS FOR MORE PETROGENESIS AND THE NATURE OF CRUSTAL ACCRETION WITHIN THE NEOVOLCANIC ZONE OF TWO RECENTLY ACTIVE RIDGE SEGMENTS By Matthew C. Smith December 1999 Chairman: Michael R. Perfit Major Department: Geological Sciences A detailed investigation of two recently volcanically active mid-ocean ridge (MOR) segments in the N.E. Pacific Ocean has resulted in a better understanding of crustal accretion within the neovolcanic zones of mediumto fast-spreading ridges. The use of submersible and other near-bottom systems, in conjunction with rock sampling by wax core, provides excellent spatial control, resulting in a database of precisely located in situ geologic observations and samples. The combination of dense high-resolution sampling, temporal control gained by recent eruptive activity, and multi-year monitoring efforts makes these investigations unique in MOR research. The CoAxial Segment of the Juan de Fuca Ridge (JdFR) is a medium-spreading rate ridge that has experienced at least three volcanic eruptions between 1981 and 1993. Evidence of amagmatic extension and chemical heterogeneity between different lava flows indicates that the CoAxial Segment has behaved as a magma limited system, despite the occurrence of recent eruptive activity. The current CoAxial Segment neovolcanic zone is greater than 1 km wide, and spatial focussing of eruptive activity is vi

PAGE 7

poor relative to the more magmatically robust Cleft Segment of the JdFR and 9-10 N segment of the East Pacific Rise (EPR). A broader survey of the central JdFR reveals that, on a regional scale, geochemically different melt regimes can be associated with distinct structural, morphologic and volcanic provinces. Detailed study of ridge-axis morphology, structure, and chemistry of the faster spreading, magmatically robust, 9°-10° N segment of the EPR shows a different relationship between ridge-axis structure and magmatic activity than that observed at the CoAxial Segment. Data from a third-order overlapping spreading center (OSC) located at 9° 37' N suggests that there is a close association between magmatic activity and axial discontinuities, and further suggest that this OSC is a magmatic and hydrothermal divide between adjoining third-order ridge segments. Temporal constraints correlate a recent magmatic event to southward propagation of the eastern OSC limb. These data establish a direct causal relationship between magmatic activity and evolution of ridge-axis discontinuities, and confirm that segmentation of melt delivery to the ridge-axis may affect ridge-axis structure along the fast spreading EPR. vii

PAGE 8

CHAPTER 1 GENERAL INTRODUCTION The mid-ocean ridge (MOR) system is a globe-encircling volcanic mountain chain more than 60,000 km long. It extends throughout all of the world's major ocean basins, rising some 1000-3000 m above the abyssal sea floor. Mid-ocean ridges are the site of oceanic crustal accretion and lithospheric genesis and, as such, are important to the study of plate tectonics and Earth's geological processes. While an understanding of magmatic processes occurring at MORs is an essential part of understanding plate tectonics, the significance of magmatic activity at ocean ridges extends far beyond geological interests. Magmatic activity along MORs provides the source of heat that drives extensive hydrothermal circulation within the oceanic crust which, in turn, exerts a strong influence on ocean chemistry. Not only does this volcanically driven hydrothermal system help to control seawater chemistry, but it also provides the energy that supports a rich and diverse abyssal ecosystem that utilizes chemosynthetic reduction as the basis of its food web. Surprisingly, relatively little of this immense volcanic system has been studied in detail. Difficulties in investigating such a remote and extreme environment limited direct examination until the 1960's when advances in technology allowed man to travel to and freely access the mid-ocean ridges, which typically range from 2000-4000 meters below sea level. Early studies sampled the MOR system with coarse spatial resolution, providing insights on some of the first-order differences between different ridges. The 1

PAGE 9

2 advent of swath mapping systems that allow for rapid underway bathymetric mapping of large areas of seafloor greatly advanced the understanding of MOR morphology and structure. Discontinuities and offsets in the strike of the ridge-axis separate the MOR system into segments of varying scale, depending on the magnitude of the axial discontinuity. Macdonald et al. (1988) established four different orders of ridge axis discontinuity classified on the basis of the discontinuity's spatial dimensions. In this categorization, which is today generally recognized as the accepted nomenclature, first-order spreading segments are tectonically defined and bounded by large transform offsets or propagating rifts whose offsets are such that plate boundaries behave rigidly, juxtaposing crust of greater than approximately 0.5-1 Myr age difference across the offset boundary. Offsets are typically greater than 50 km and partition the ridge-axis at intervals of 100-1000 km. Second-order segments have a length scale of 50-300 km, and they are bounded by smaller, non-rigid discontinuities such as large overlapping spreading centers (OSC) with an offset greater than 3-5 km or small offset (less than 20 km) non-rigid transform faults. Third-order segmentation divides the ridge-axis into characteristic lengths of 30-100 km, and these segments are bounded by smaller OSCs (0.5 to 3-5 km). These smaller OSCs are often associated with small (lO's of meters) increases in axial depths. Lastly, fourthorder segments have typical length scales of 10-50 km and are not commonly associated with increased axial depth near the segment ends. These fourth-order discontinuities are defined by small non-overlapping offsets of the ridge axis and/or small (r-5°) bends in the ridge-axis referred to as deviations in axial linearity (DEVALS) [Langmuir et al., 1986]. Although first-order segments and their associated axial discontinuities are

PAGE 10

3 thought to be persistent on the order of milhons of years, smaller ridge offsets may not be as temporally stable [Macdonald et al., 1988b]. As investigations of the MOR system were undertaken at finer and finer spatial scales, it became apparent that the mid-ocean ridge environment was the product of complex interactions between magmatic, tectonic, hydrologic, and biologic processes. Only by compiling detailed investigations of processes occurring at different spreading segments would a more comprehensive understanding of this environment be achieved. To date, only a small percentage of the mid-ocean ridge system has been subjected to detailed multi-disciplinary investigation at the first to fourth-order segment scale. Oceanic ridges exhibit a range of plate divergence or "spreading rates". Full divergence rates vary by over an order of magnitude, and to a first-approximation ridge morphology, structure, and magmatic activity are strongly influenced by this spreading rate [Perfit and Chadwick, 1998]. Typically, slow-spreading ridges (full plate-divergence rate on the order of 1-4 cm/year), exemplified by ridges in the Atlantic and Indian oceans, rise steeply from the abyss relative to faster-spreading MORs. They are generally rifted by an axial graben that can be lO's of kilometers across and greater than 1500 meters in relief. Slow-spreading ridge terrains are generally more greatly affected by extensional tectonics and faulting and, as such, tend to be much rougher than at faster-spreading ridges. Volcanic morphology of the axis floor, which is the site of active crustal accretion, tends to be dominated by discrete constructional features that may coalesce to form mounds or ridges. Lava morphology is often pillowed, a characteristic associated with relatively slow lava effusive rates. The zone of crustal accretion, or "neovolcanic zone," is thought to be relatively wide and unfocussed (2-12 km wide), with relatively

PAGE 11

4 less control on the location of magma emplacement in the crust relative to faster spreading regimes. Contrary to the above, fast spreading ridge systems (-8-16 cm/yr full-rate), exemplified by the East Pacific Rise, exhibit very different characteristics. Characteristic morphology of fast spreading ridges is that of a broad swell or "rise" in the abyssal seafloor. Fast spreading ridges generally lack a large fault bounded axial valley, and when an axial graben is present, it is of a much smaller scale than those present at slow spreading ridges. Overall ridge structural fabric is less rough, and melt focussing appears to be greater, with a majority of volcanic activity being restricted to a much narrower zone of accretion (100-200 m wide). Magmatic activity appears to be much less episodic than in slower-spreading environments, and overall ridge fabric appears to be more greatly dominated by magmatic processes rather than tectonic processes. Many sections of fast-spreading ridge can be associated with a seismically detectable axial magma chamber in the shallow crust, while these features are largely absent from seismic imaging of slow-spreading ridge-axes. Medium spreading-rate ridges (-4-8 cm/yr fullrate), exemplified by the Juan de Fuca Ridge in the northeast Pacific, exhibit characteristics intermediate between the two end-member axial morphologies (neovolcanic zone 200-2000 m wide). While recognition of the first-order characteristics has allowed for great advancing in the understanding of mid-ocean ridges, details of the process of oceanic crustal accretion has been hampered by a lack of any high resolution temporal control to the origins of observable phenomenon. The first documented recent eruption was not until the 1990s when anomalies between repeat bathymetric surveys were able to show the

PAGE 12

5 existence of a seafloor volcanic constructional terrain that was not present 6 years prior (time of eruption constrained to being between 1981 and 1987) [Chadwick et al., 1991; Embley et al., 1991]. Since then, fewer than 10 additional eruptions have been confirmed and investigated, several of which are reported on in this dissertation. Numerous questions fundamental to understanding MOR crustal accretion and the linkages between geologic activity and important biologic and oceanographic processes still exist. Outstanding questions, among the as-of-yet unsatisfactorily addressed geologic issues include (but are not limited to): (1) How frequently and with what spatial consistency do volcanic eruptions occur on a given segment of MOR? (2) What is the dependence of eruptive frequency on spreading rate and melt supply? (3) How does spreading rate and melt supply affect melt focussing and the physical configuration of the neovolcanic zone? (4) How does magmatic activity affect ridge axis structure and segmentation or viceversa? (5) It has been observed that different segments of ridge spreading at similar overall rates can have very different axial morphologies. Is this observation best explained by a temporal cyclicity to the intensity of magmatic and tectonic activity at discrete spreading segments, or do discrete segments have inherent morphologies resulting from differences in volcanic activity that are constant over longer time scales? Answers to these and other questions that relate to temporal variability and evolution of magmatic, hydrothermal and biological systems can only be answered through the integration of detailed, long-term studies and active monitoring of many different portions of the global MOR system. While this dissertation does not provide holistic answers to any of the above questions, it does present the results of detailed

PAGE 13

6 geologic and petrologic studies of two recently active ridge segments, each occurring in different spreading-rate and apparent melt-supply regimes. Documentation of recent eruptive activity along two second-order segments of the eastern Pacific mid-ocean spreading system provides a rare opportunity to examine the geometry and petrogenetic associations of magmatic events that form the basis for crustal accretion along medium to fast spreading divergent plate boundaries. The CoAxial Segment of the medium spreading-rate Juan de Fuca Ridge and the 9°-10° N Segment of the fast-spreading East Pacific Rise were the sites of eruptive activity in 1993 and 199192 respectively [Fox et al., 1995; Haymon et al., 1993]. The Juan de Fuca Ridge located approximately 500 kilometers west of the Washington/Oregon/British Columbia coast. The 9-10° N latitude segment of the East Pacific Rise is located west of the Central american coast, between the Clipperton and Siqueros fracture zones. Detailed maps of these two regions are provided in chapters 2 and 3 respectively. Not only are these two second-order spreading segments [Macdonald et al., 1988] different in terms of spreading rate, but they also differ greatly in axial morphology, and possibly current melt supply rates from the upper mantle. The 9°-10° N segment of the East Pacific Rise has an axial morphology typical of faster spreading ridge segments dominated by magmatic activity. Contrary to this, the CoAxial segment has an axial morphology intermediate in nature between the fastand slow-spreading end-members, and is much more indicative that amagmatic extension plays a prominent role in the development of axial morphology. Additionally, these two recently active ridge segments provide a good contrast to the Cleft Segment, another comparably studied portion of the Juan de Fuca Ridge that although spreading at a similar rate to the CoAxial

PAGE 14

7 segment, displays many of the volcanic features observed on the EPR [Smith et al., 1994]. The two investigations presented in this dissertation, and the results that arose from them, are briefly summarized below in "Abstract form" for the convenience of the reader, providing a bit more detail than is presented in the overall dissertation abstract. 1.1. The CoAxial Segment Recent investigations along the central portion of the Juan de Fuca Ridge have identified at least three different eruptive units that were emplaced along the CoAxial Segment between 1981 and 1993 [Chadwick et al., 1995]. Major element, trace element and isotopic compositions of mid-ocean ridge basalt (MORB) samples recovered by submersible, ROV and wax core are examined in order to critically assess the spatial and temporal chemical variability on regional and local scales. The data allow us to quantify the chemical variances within and between each of the mapped flow units, identify and define genetic relationships between the flows, and to document the relationship of these recent eruptives to older basalts recovered along the CoAxial and other nearby segments of the Juan de Fuca Ridge. Geochemical data show that each of the three recent eruptives are distinct from one another, supporting the hypothesis that at least three separate volcanic units have been emplaced along the CoAxial segment since 1981. Further, when compared to fresh looking eruptives sampled from within Axial Seamount caldera and along its associated north and south rift zones, recent CoAxial eruptives have isotopic and elemental characteristics distinct from lavas associated with Axial Seamount volcanism. This suggests that magmas parental to lavas sampled from each of these regions have unique origins, and that it is highly unlikely that Axial Seamount was the source of the 1993

PAGE 15

8 CoAxial eruption, as has been suggested. CoAxial segment lavas are geochemically very depleted and show no geochemical evidence for influence of Axial Seamount hotspot. The integration of these geochemical data with observational, photographic and sidescan sonar data defines several different tectono-magmatic provinces in the region between the Vance and Cobb segments of the Juan de Fuca Ridge. The presence of these different provinces suggests quite distinct sub-axial melting regimes exist along this medium rate spreading ridge. 1.2. The East Pacific Rise between 9° and 10°N Alvin based investigations of a small ridge-axis discontinuity, located at the crest of the East Pacific Rise crest between -9° 35-37'N, provide strong evidence that the feature is actually a small overlapping spreading center (OSC) that divides two magmatic and hydrothermal systems. Two overlapping axial summit collapse troughs (ASCT) morphologically express this third-order ridge-axis discontinuity. The eastern ASCT that extends from the north has an approximately 0.4 km right-lateral offset from the western ASCT that extends from the south and overlaps it by at least 1 .2 km. Changes in volcanogenic morphology and increased hydrothermal and biologic activity observed between surveys of the area in 1989 and 1991 suggest that a magmatic event affected this region during this time-period. These changes combined with geochemical data from basalt samples and hydrothermal fluids recovered during three Alvin dives suggest that the eastern limb of the OSC is actively propagating to the south. Observational data that support this hypothesis include: (1) the occurrence in 1991 of extensive areas of bacterial bloom and diffuse venting of 39°C, low-chlorinity hydrothermal fluids along the southern extension of the eastern ASCT; (2) regions of local collapse of drained out young-

PAGE 16

9 looking lava flows, the presence of lava pillars, and hydrothermal fluids venting directly out of glassy, young-looking lava flows along the southward extension of the trend of the eastern ASCT; and (3) the presence of several extinct sulfide edifices within the western ASCT, and venting of high chlorinity hydrothermal fluids at temperatures of greater than 1 13°C in 1991 and 1994 within that region. MORB geochemistry lends further support to the hypothesis that southward propagation of the eastern OSC limb at 9° 37' N is related to recent magmatic activity in the 9° 37-52' N area. Basalt samples with the highest MgO (most primitive, highest temperature) are typically the freshest-looking lavas and tend to be spatially associated with the ASCT of the eastern OSC limb, and its extension to the south. More fractionated (lower temperature) basalt samples are found along the western limb ASCT and in the zone of overlap between the two summit troughs. The more evolved MORB western ASCT MORB are consistently older in appearance than the fresher, younglooking high-MgO lavas to the east. Recently enhanced magmatic activity of the OSC's eastern limb is apparent by MgO content of the most recent eastern ASCT lavas which have Mg#'s that are anomalously high relative to the trend observed in the regional MORB geochemical data. These lavas have major element chemistry similar to the 1991 eruptive at 9° 46-52' N and may be related to the same phase of magmatic activity. Additionally, the chemistry of hydrothermal fluids recovered both within the region of the OSC and along segments of the EPR crest proximal to it, indicates that the OSC marks a hydrographic divide. Hydrothermal fluids generally have higher chlorinity than seawater south of the OSC, and a lower chlorinity than seawater north of it (including fluids from the eastern OSC limb). Investigation of hydrothermal fluids between 9° 17-

PAGE 17

10 54N have shown that low-chlorinity fluids are associated with either eruptive/intrusive events, or occur in areas with very shallow and robust heat sources. With time, the fluids often increase in chlorinity. It is only at the 9°37'N OSC and further south that extensive extinct sulfides and fluids with chlorinities -1.5 times seawater have been sampled, suggesting that hydrothermal systems north of the OSC have been more recently perturbed by a magmatic event than those on the failing rift and further south. It is proposed that the OSC's southward propagation is locally driven, resulting from a combination of renewed and enhanced magmatic activity north of the propagator, and a waning magmatic system south of it. Local occurrence of E-MORB on the Pacific plate near the OSC may reflect lower degrees of melting in the region of the retreating OSC limb.

PAGE 18

CHAPTER 2 GEOCHEMICAL CHARACTERISTICS AND PETROGENETIC RELATIONSHIPS OF RECENTLY ERUPTED MORE FROM THE CENTRAL JUAN DE FUCA RIDGE: SEGMENT SCALE, INTERFLOW, AND INTRAFLOW VARIATIONS 2.1. Introduction 2.1.1 Field Investigations and Previous Studies The 1993 detection of a seismic swarm thought to be related to dike intrusion along what is now known as the CoAxial Segment of the Juan de Fuca Ridge [Fox et al., 1995] triggered an extensive series of investigative cruises to this portion of the midocean ridge system. A total of three rapid response cruises undertaken between July and October of 1993 (the seismic event lasted from June 26 until July 10, hydrocasting began on July 3 and the first direct sea floor observations began on July 9) investigated the remotely detected seismic events. Investigators participating on these cruises sampled hydrothermal event plumes and visually confirmed that a basalt flow had indeed erupted onto the ocean floor in the location where the most intense seismic swarms had been located [Embley et al., 1995; Baker et al., 1995]. These cruises concentrated on mapping and sampling lavas, biota, and hydrothermal waters associated with the new flow and along the path of proposed dike emplacement. Subsequently, at least 7 cruises to this region have expanded the sampling and mapping efforts to include other portions of the CoAxial Segment as well as Axial Seamount and its associated north and south rift zones. 11

PAGE 19

12 Prior to the aforementioned events, very little was known from this section of the Juan de Fuca Ridge (with the exception of Axial caldera). The CoAxial Segment had not been recognized as a distinct second-order MOR segment [Macdonald and Fox, 1983; Embley et al., 1995], and geochemical data from only four relatively poorly located dredges had been reported in the literature (Eaby et al., 1984; Rhodes et al., 1990). 2.1.2 Regional Geology The geology and morphology of the CoAxial Segment and its surrounding terrain is discussed in detail by Embley et al. [in press], and, as such, will only be discussed here in sufficient detail to provide basic background information. This paper is, in many ways, complimentary to and an extension of a recent paper by Embley et al. [in press]. The Juan de Fuca Ridge (JdFR) is a first-order, medium spreading rate ( ~6 cm/yr full rate) ridge segment located approximately 500 km off of the OregonAVashington/ British Columbia coast (Figure 2-1). It extends some 490 km along strike (-020°) between the Blanco and Sovanco Fracture Zones which intersect the ridge-axis at ~44°30' and 48°45' N latitude respectively. The JdFR has been subdivided by various authors into at least 6 different second-order ridge segments ranging from 50-150 km in length. More detailed discussions of the JdFR and its second-order subdivisions have been published [Delaney et al., 1981; Johnson and Holmes, 1989; Kappel and Ryan, 1986; Kappel and Normark, 1987; Hey, 1982]. Axial Seamount, a large ridge-axis centered volcano that is the latest expression of the Cobb melting anomaly [Desonie and Duncan, 1990], is located in the middle portion of the JdFR at -45° 55'N latitude and 130° W longitude (Figure 2-1). This large edifice rises more than 1000 m from the surrounding abyssal terrain and has produced a large summit caldera and two major rift

PAGE 20

Figure 2-1: Bathymetric map of the CoAxial Seamount region taken after Figure 4 of Embley et al., in press. True north is towards the top of the page. Regional NE Pacific map is shown in inset (upper left). Shading depicts interpretation of volcanic terrain based on sidescan sonar and bathymetric data (Embley et al., in press). Basalt sample locations are shown as crosses (wax core samples), filled circles (ROV/Submersible samples) and filled squares (starting/ending points of dredges). Solid black outline depicts the interpreted aerial extent of Axial Seamount north rift zone volcanism (see text). Dashed lines mark major rift zones of Axial Seamount and the CoAxial Segments. Rectangular outlines mark the Floe and Flow Sites which are shown in Figures 2-2a (southern outline) and 2-2b (northern outline) respectively. Regional shading is based on interpretation of sidescan and bathymetric data and is keyed to the legend (lower right).

PAGE 21

14 46" OO'N life m-^!^-: O Volcanic Ediflce O Lava Flow Q Volcano + Rock Core « Sub/ROV sample /' Rift Zone Kilometers I I 1 1 0 5 10 15 130° 20'W 130° OO'W 129°40'W 129° 20'W

PAGE 22

15 zones extending north (to -46° 15' N) and south (to at least 45° 30' N) from the main edifice, which is some 20 km in diameter [Johnson and Holmes, 1989]. Though this edifice is the youngest volcano of the Cobb-Eickelberg chain, a greater than 300 km long seamount chain. Axial Seamount lavas do not display geochemical enrichments typically associated with hot spot volcanism [Rhodes et al., 1990]. Together, Axial Seamount and its north and south rift zones comprise the Axial Seamount Province. The southern end of the Coaxial Segment exists in an overlapping relationship with the Axial Seamount' s north rift zone (NRZ) to its west, overlapping it by at least 25 km. The southern terminus of the CoAxial segment is at Helium Basin (-46° N and 129°55' W) which lies at the base of the steep, faulted NE flank of Axial Seamount [Lupton, 1990]. The CoAxial Segment extends north-northeast along a strike of -022° to -46° 40' where it overlaps some 10 km with the Cobb segment to its east [Embley et al., in press]. The ridge is shallowest at its southern end, and deepens northward along the strike of the ridge by approximately 425 m to a depth of -2550 m at its northern terminus. The CoAxial Segment's axial valley is quite wide compared with other JdFR second-order segments (up to 2 km), as is the apparent current zone of active crustal accretion, which will be discussed further below. It is intriguing that the only other segment on the JdFR with as wide of an axial valley is the Vance Segment which lies immediately south and with Axial Volcano's south rift zone (SRZ) [Embley et al., in press]. Based upon our sampling regimen and the criteria presented in Embley et al. [in press], we have designated three different sites along the CoAxial segment as the primary focus of our investigations. These sites listed from south to north are the "Source Site",

PAGE 23

16 the "Floe Site" and the "Flow Site" (Figure 2-1). Each of these sites is described in detail in Embley et al. [in press], and the reader is referred to that paper for a detailed discussion of their geology and morphology. 2.2. Data Collection and Sample Distribution 2.2.1 Alvin and ROV Based Sample Acquisition A total of 136 site-specific in situ basalt samples have been recovered from the CoAxial segment since the first rapid response cruises in late 1993 (Figures 2-1 and 2-2). These samples were collected utilizing a variety of submersible (104 samples) and remotely operated vehicle (ROV) platforms (32 samples) employed during 7 different oceanographic expeditions conducted between 1993 and 1996. The majority of ROV and submersible sampling concentrated around the three primary study sites (described above and in Embley et al. [in press]), with an emphasis placed on sampling lavas of different apparent age and flow morphology. In addition, special attention was paid to recovering a denser and well distributed suite of samples for areas where discrete eruptive units could be identified and mapped on the basis of visual observations and depth differences between repeat bathymetric surveys [Chadwick et al., 1995]. Relative ages and flow contact relationships were deduced from stratigraphic relationships, flow morphology, amount of sediment cover and presence or absence of sessile organisms. 2.2.2 Surface Based Sample Acquisition The suite of basalt samples recovered by submersible and ROV is supplemented by 57 samples recovered by wax corer (out of 66 attempts), two dredge hauls and one TV Grab (Figures 2-1 and 2-2). Surface based sampling was generally performed at night or

PAGE 24

Figure 2-2: Expanded view of the Floe (A) and Flow (B) sites depicted by rectangle outlines in Figure 21 . Figures are after figures 1 1 and 8 of Embley et al., in press. Basalt sample locations, mapped boundaries of recent flows discussed in text, major fissures and low temperature hydrothermal venting locations are superimposed on images generated by deep tow AMS 60 sidescan sonar (Embley et al., in press). Flow boundaries were determined in consort with visual and photographic mapping, repeat bathymetic surveying and sidescan data. Wax core samples are shown as crosses, and submersible/ROV collected samples as circles. Fissures marked by bold white lines (22A) depict the current zone of hydrothermal activity, and interpreted ridge-axis, which extends northward passing directly underneath the 1993 Flow (2-2B).

PAGE 26

19 129°35'W 129°34'W

PAGE 27

20 during submersible/ROV down time. Wax cores primarily sampled the Axial Seamount NRZ (15) and regions along the CoAxial segment outside of the major sites of submersible investigation. Dredges sampled Rogue Volcano and a topographic high bounding the east side of the CoAxial Segment. The video guided grab sample recovered basalt from a fresh flow outcropping along Axial Seamount NRZ. This flow was also sampled by wax core. 2.2.3 Sidescan Sonar and Near Bottom Photographic Data Acquisition Photographic data gathered by submersible, ROV and near bottom camera tow. as well as sidescan imagery acquired using SeaMARC II and AMS-60 sidescan sonar systems, have been incorporated in our regional geologic interpretations. Submersible, ROV and camera tow data were generally located using long-baseline transponder navigation that provides positional accuracy to within approximately 20 meters. AMS-60 sidescan data have a spatial resolution of Im. Details of the distribution of both the photographic data and the sidescan sonar imaging used to place samples in a geologic context are presented in Embley et al. [in press]. 2.3. Analytical Methods 2.3.1 Sample Preparation All basalt samples were described and catalogued at sea. Geologic context and the location for each sample were carefully documented, and glassy rinds sub-sampled when present. Shoreside, a representative suite of samples were sectioned for petrographic analyses. Natural basaltic whole rock glasses (quenched glass ± crystals) were cleaned and prepared for analysis by x-ray fluorescence. Glasses were rinsed in 2X distilled water and the cleanest portion (generally 5-50 grams for ALV7A^/ROV recovered

PAGE 28

21 samples) of the glass was hand picked under binocular microscope for further preparation. Picked glass subsamples were subsequently coarsely crushed and sonicated first in a solution comprised of equal parts 2.5N HCL and 30% H2O2 (15-30 min) and then in 2X distilled water. Samples were then dried and picked free of any visible surficial alteration or precipitated coating using a binocular microscope. Glass splits to be analyzed by electron microprobe, thermal ionization mass spectrometry (TIMS) or laser ablation ICP-MS (LA ICP-MS) were then further hand picked to assure that only the most crystal-free and pristine glass was analyzed. Glass splits to be analyzed by XRF or solution ICP-MS were ground using either a tungsten-carbide disk mill, a SPEX agate ball mill, or some combination of the two. Use of the tungsten-carbide disk mill was kept to a minimum, but it is likely that in at least some of the samples the elements Ta and W may be to some degree compromised by tungsten-carbide contamination [Thompson and Bankston, 1970]. Samples (< 100 mg) analyzed by TIMS were subjected to a second leaching (-10 minutes) using warmed distilled 6N HCL and digested in a solution of distilled HF and ultra pure HNO3. Dissolved glasses were then chlorinated using distilled 6N HCL and Sr and Nd separated and collected using standard ion chromatographic techniques. 2.3.2 Major Element Analyses The major element abundances of 153 natural basaltic glasses are reported here and a representative subset of the data are presented in Tables 2-1 and 2-2. The majority of these analyses were measured using a JEOL 8800 electron microprobe (EMP) at the USGS in Denver. Multiple standards were frequently measured during the course of each analytical run, including the mid-ocean ridge basalt (MORE) glass standards JdF-D2 and

PAGE 29

22 a On NO NO NO 00 00 u-i in NO NO NO NO oo m 100.0 ON c NO Tot Ov ON On ON On On a\ ON ON ON ON ON ON On ON ON ON ON ON ON ON ON ON ON o^ 0\ ON ON ON ON ON as ON ON ON ON ON On ON On ON On o^ On a\ On On ON ON On ON ON ON On On ON ON On ON On o rO ON ON o (N o On NO NO m NO _ _ tN d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d O CM — NO in NO NO in in NO NO NO 00 o 00 oo 00 ON oo ON o 00 d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d O a m SD NO rNO NO NO NO 0\ NO ON NO in NO NO (N NO NO 00 NO 00 in ON NO ON NO 00 NO in 00 oo oo o Z (N cnI rj CNi oi IN fN tN CN) CNi tN tN tN tN tN tN tN fN tN tN tN fN tN tN tN fN o rn m rn <»N ON 00 00 ON oo On On On O On 00 00 fN NO fN Ca d d d d d d d tN fN fN tN fN (N tN fN O 00 O o\ ON NO in NO NO o ON ON IT) ON m 00 o fN NO q NO o fN o in ON rC ON in 00 m ON 00 ON NO On NO in 00 vd NO NO NO NO NO vd NO NO NO NO NO NO r-^ [-^ o c rn n m in m in in m in m m 00 ON o O ON o oo in m oo 0\ m m m m in m tN tN tNtNtNfNfNfNtNr<-)r<^rn in r-; fN NO NO q NO rj in d r"r ^CNON— .fNTj-m — TtinmON ^. ^•^^^ NqoqoOTfinfNoooomooqp — r^l^r^ododofjodododotiod f*^ — — — — — — 00 oo r-~ 1— P„ ?„ ° ° ° ° o ° o o o o o o o o o nONDnOnOND NOnONONDnONOnOnOnOnOnOvO OS OS Oi fN NO fN tN Oi °r d tN 1 oo 1 fN rlm oo ON ON 00 OO ON NO NO 1^ On On On fN fN fN (N tN fN f^l fN T+^V-N /i^. no —.1 —.1 ' ' '_ I I I Di m a: fN n as On On 00 1^ ON fN fN (N fN fN(N^(N(NrslrNl_, rNlfNfNfNtNfNtNtNfN

PAGE 30

23 oo NO NO r00 NO NO o m O m oo m 00 00 NO rin NO ON On ON ON ON ON ON ON ON ON ON ON ON ON ON ON ON ON ON ON ON On On On O^ On o^ ON ON On ON ON ON On as ON On ON o\ ON OS ON as ON as ON O^ ON o^ ON as ON ON ON ON o tN NO 00 OO o o rtN m tN o (N oo m NO in t^ m in (N o tN ; — ; in m d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d o o q o o q o q o o o o o o o o o — o o in O d d d d d d d d d d d d d d d d d d d d d d d d d d d d CNl oo tN rNO NO tN t — t^i 00 ON m OO CO in in in NO NO NO tN in NO tN in in NO ON m in m in NO NO in ND ON tN CN tN tN oj tN (N (N tN tN tN tN tN tN tN tN tN tN tN tN tN tN tN tN tN d tN d tN d tN O (N ON 00 lO tN ON in m NO O ON 00 o ON ON in d tN ri tN tN tN tN tN d tN tN d (N OOOOooONoooNddododoNoddONddoNONdodddddd_:rf; TrininininTfTj-TfTtininTrrfTfTf5=S^inin^^5^^S^^in^==? tNcnminNomTj-^cr-, ininminininminininininin in m in in in 'd-Ti-TtTtTtininininminininmininininininininmNONONONO tN ON O q in o o On on (N (N IN in — : q en CO o o On on tN (N in Tt (N On in — rn in m m o o On On tN CN — oo NO m d o o ON ON (N tN o o ON ON (N tN Ol — — tN On o o ON On (N (N On O o o On on tN On in O On O o o On on (N (N ON On O ^ tN NO NO o o ON On (N (N — tN OO (N (N NO CO o o ON ON (N (N tNooinrtinONoorsi tN m 00 tN — OO 00 m oo — in NO in in in 00 oo oo oo r<^ Tjre : > II u c Q .185' .232' .102' .218' .599' .807' .591' ,778' .955' .354' .024' ,624' in ON ZLL .960 994' 071' OO ON 280' 280' .086 ,890 404' 137' NO rr~ ro tN o CO m m m (N r^f m (N 00 tN On (N od tN tN m tN d <^ On O ON o On O o NO o NO 0 NO o NO o NO o NO o NO o NO 0 NO o NO o NO o NO o NO o NO o NO o NO o NO o NO o NO 0 NO 0 NO NO o NO o NO o NO o NO OSClSDioS^inQiQiQSalQj NOr^ — TTDsmNO — tNoo — oooooNONin?oo(N(N(NTr 3"3'^'^'^"~'^OnOnONOn ?ly^oNONONtNr^[^[ — r^r^ in m NO a: OS °T m as OfN OS oo oo oo OO oo ON ON On On ON (N (N (N (N tN MP.tNtNCsKi^hh^KSSSagigjaagJgJ^g5f:i?^^^ tN m tN 5 a; 6 O d 1 On On On ON ON ON On On On On On ON (N tN tN tN tN (N ^3 ONOnOnOnONONOn (NtNtNtNtNtNtN 5^(N >„ °1 NO -J o — > II u c Q 2 " tN tN . — > II 1> c Q ^ (N OS QJ O 1 in tu t-^ O 6 in oo ON ON I* On r~ ON On tN tN tN tN nit D Un a a o o c/: O O 00 oo ON ON

PAGE 31

24 p~ so 00 rr~ in rNO rr-NO ON 00 NO oo r00 NO 00 in NO oo 00 oc 00 OC' ON 00 oo 1^ 00 Tot OS as as On as ON On On On On On On On ON ON o^ ON On On O^ ON On' ON ON ON o^ ON ON ON C3N ON ON ON ON ON On ON ON ON ON ON ON ON ON ON ON ON ON ON On On o^ ON ON ON ON ON ON On ON ON ON ON O Os tN in m 00 O m ON m in m (N OC m ro 00 o o O 0. d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d O NO o 00 o 00 00 m NO On O On O ON o in tN NO o NO o CN ON O On O o On O On O d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d O ca t — r-i On NO oo in CN OO q ON NO 00 ON m CN [ — 00 NO CN tN NO On NO 00 CO oo oo On In Np ON in cn NO NO 00 ON Z ri oi (N C NO NO NO NO NO 00 o<: r~-' 0C3 r-^ oo' r--' 00 O ON CN CN m o ON CO O o ON CN CN tN CN 00 o CN CN o o CN o c CN CN CN Ol CN CN CN tN CN CN CN CN CN tN CN tN CN CN CN CN CN CN tN CN CN OI CN d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d O CO NO t CN in CO CO in CO CO in CO NO On CN NO ON NO 00 ON q 00 Fe d d d oi CN CN CN CN d d d d d d d d d d d d O NO NO in NO in ON rON o CO CN NO CN ON NO ON 00 t m (N ON ro o 00 tN 0C5 00 NO NO CO CO CN CO CO tN in in NO CO CO CO in NO H d in CN NO OO On 00 ON 00 CO o NO in o CN NO NO CO NO in ro ro < in CO CO ro CO in m NO o d d d in m in CN CN d d d d ON in in in in NO NO o 00 00 ro NO ON m o d d d d d d d d d d d d ON ON d d d On d ON d d d d in m in m in in in m in in in m Tiin m in m m m m .127' .280' .107' .107' o in NO .318' .987' .825' .231' .684' .540' .163' .052' .703' .197' ,236' ,824' 127' 261' 653' 519' 386' ,060 ,860 ,600 .600 858' 321' oin ON 500' 374' ro TtrNO NO NO NO •o;o;o;o^ -rt -rt -rf Tt U U U U OS OS OS OS u u OS a: in m m in in CN CN CO CN u U u U u OS OS OS OS OS m in ON ON ON ON ON .5 an

PAGE 32

25 o ON o o o (N (N o d d d d d d Tj— O o o o o o o o o — o — — d d> di d> d> (N ON 00 VO fN vo vo vo \d rr~r-^ ^ ro tN tN (N fN ON p r-^ r-^ 00 ON rn ON q m >n r~q On q q q q tN d d d d d d d d d d d d d d d d d d d d d d d in o o o o (-~o o o o o IN ON o CN 00 NO NO tN tN o q 00 o CO O o tN tN — d d d d d d d d d d d d d d d d d d d d d d ON IT) IT) 00 >o tN ON m °° NO 00 o On in ON q in 00 ON 00 ON ON ON q 00 00 ON NO NO NO tN tN tN tN CN tN tN tN tN tN tN tN CN tN tN tN tN r~ 00 O tN ON OC) O tN tN On — tN — ON OO m oo ON 00 — tN tN tN tN tN ts tN tN oi tN IN tN (N rn d (N tN ON 00 p~ tN 00 q >n m ON tN m tN <* oo NO 00 00 00 00 d ON NO NO ON ON ocj ON ON NO oo" 00 ON tN (N ON 00 ON tN tN tN > > ON tN III cd CO > > ^ 4) OJ •a -a — _, 00 NO ii: CM d ON NO NO ON NO o in o NO in m ro oo NO in tN NO tN tN NO ro tN 00 in ^ oti oi 00 0\ On d d in ON cn o m O O o O o NO o o o o o o o o o o o o NO NO NO NO NO NO NO NO NO NO NO NO m ro tN tN o o NO NO I ( E B
PAGE 33

USNM. Plots of Si02 abundance (wt.%) vs. Total (wt.%) of all oxides measured were made for each individual analytical run based on the secondary standard data acquired during that run. Where strong correlations existed, the raw abundance data for SiOi was corrected along a linear regression after Reynolds [1995]. Individual oxides were normalized to the standard JdF-D2 (unfractionated JdF-Reynolds, 1995) in order to greater facilitate their comparison to other published MORB data. A small subset of the major element data was analyzed by EMP at the Smithsonian Institution by T. O'Hearn. Smithsonian data were not Si02 corrected but were normalized to the secondary standard VG2. In order to facilitate direct comparison to our microprobe data, these data were corrected to a "JdF-D2" normalization using numerous replicate samples and secondary standards analyzed in both labs. Measured values for the MORB standard JdF-D2 and long term analytical precisions are shown in Table 2-2. The complete major and trace element set is available from the author. 2.3.3 Trace Element Analyses 2,3.3.1 X-ray fluorescence Seventy-five powdered basalt whole rock glasses were measured at the University of Florida for the elemental abundances of Ba, K, Nb, Rb, Sc, Sr, Ti, V, Y and Zr using a fully automated ARL8420+ x-ray fluorescence (XRF) spectrometer with a Rh-target endwindow tube. Details of the operating conditions are reported in Smith et al., [1994]. Analytical precision based on long term reproducibility of standard reference materials is better than 5% for the elements K, Sr, Sc, Ti, V, Y and Zr. Precision is better than 10% for the elements Ba, Nb and Rb at values greater than 20, 3 and 2 parts per million (ppm) respectively. Precision drops to greater than 20% below these levels. Average values

PAGE 34

27 and 2-sigma precisions for the standard BHVO-1 are given in Rubin et al., 1998, and average measured values for our internal standard ENDV (an enriched MORB from the Endeavour Segment, JdFR) are shown in Table 2-3. 2.3.3.2 ICP-AES/MS Abundance data for major elements, the rare earth elements (REE) and selected other trace elements were measured by inductively coupled plasma source emission spectrometry, (ICP-ES) and inductively coupled plasma source mass spectrometry (ICPMS) at the Geological Survey of Canada (GSC) in Ottawa. Elements/oxides measured by ICP-ES include major element oxides and a few trace elements (Co, Cr, Cu, Ni, and Zn), and those measured by ICP-MS include the REE and other trace elements. An additional group of samples recovered by wax core were analyzed by ICP-MS at the United States Geological Survey (USGS) in Denver. Fifty-two samples were analyzed by ICP for this study. A representative subset of these data are presented in Tables 2-3 (GSC data) and 2-4 (USGS data). Multiple internal standards analyzed by isotope dilution (ID) techniques on a thermal ionization mass spectrometer (TIMS) were included in each run to assess and correct for any inter-laboratory analytical bias. Correlation between the two laboratories was excellent. Details of analytical procedures, detection limits and precision for the GSC ICP data can be found on the internet at http://l 32. 1 56.95. 1 72/chemistry/ , the GCS Analytical Laboratory homepage. Analytical procedures, detection limits and precision for USGS ICP-MS data can be found in Arbogast [1990].

PAGE 35

28 Table 2-3 Concentration (ppm) of the Rare Earth and other Select Trace Elements in Representative Submersible Samples Analyzed by ICP-AES'.-MS'.and XRF'. Samnle * 2788-4R 27921R2 2794-2R 2983-R2 2983-R6 2792-4R 2792-5R 2792-7R 2948-1 IR SiO,' es 51.1 51.3 51.5 51.3 51.1 49.9 50.3 50.1 51.4 TiOj es 1.3 1.26 1.3 1.31 1.33 1.51 1.57 1.55 1.22 Al.O, es 13.8 13.6 14 13.8 14.1 14.9 14.6 14.5 13.8 FeoOjt es 12.8 12.5 12.9 12.9 13.2 12.5 12.7 12.5 11.6 FejO, es 1.1 1.2 1.8 0.9 1.5 1.4 1.1 0.7 1.1 FeO es 10.6 10.2 10 10.8 10.5 10 10.5 10.6 9.4 es n 1 Q n 1 0 \j. ly \j. ly L/.Z n 1 8 u. 1 0 \j. ly n 1 8 u. 1 0 fl 1 8 u. 1 0 MgO es 7. 14 7.05 7.18 7.2 7.39 7.4 7.35 7.23 7.37 CaO es 10.9 10.7 11 11 11.2 10.5 10.4 10.3 11.5 NajO es 2.44 2.52 2.63 2.56 2.54 2.67 2.66 2.57 2.43 K.O es 0.12 0.13 0.14 0.1 0.11 0.19 0.17 0.17 0.15 r 2W5 es 0.1 1 0.1 0 09 n I W. i n I A \A A 1 1 Total es 98.8 98.2 99.8 99.3 100.0 98.8 98.9 98.1 98.7 Co es 46 62 47 46 44 48 45 62 74 cr es 1 in 1 zu 1 zu OTA zzu OOA ZZU 1 1 A ZJU z4U Cu es 65 63 66 63 64 60 59 57 65 Ni es 61 61 61 59 61 100 100 100 64 Zn es 74 89 69 75 91 73 79 67 65 F @ 169 175 162 156 144 262 246 219 171 S @ 1263 1272 1257 1318 1305 1315 1256 1294 1 151 V xrf 405 377 396 406 401 378 386 388 332 Ba xrf 32 25 25 29 28 36 33 36 29 Sc xrf 53 45 48 46 50 48 50 49 48 TiO, xrf 1.54 1.49 1.58 1.54 1.53 1 .82 1 .oy i .00 1 'IQ 1 .jy V n xri f\ HQ u.uy u.uy A AO All 0.1 1 0.1 1 0.12 0.13 0.12 0.12 Cr xrf 98 85 1 16 83 84 223 220 241 229 Zr xrf 94 94 97 97 97 148 150 148 101 Sr xrf 90 89 91 93 93 103 99 101 94 Y xrf JS 38 39 40 40 51 50 51 35 Rb xrf 1.1 1.7 1 .4 2.1 1.4 1.6 1.6 1.7 2.0 Nb xrf 3.5 3.4 3.2 4.1 3.0 4.5 5.1 4.7 4.7 Hi ms z.z Z.4 2.5 2.5 2.4 3.5 3.8 3.6 2.5 In ms r\ 1 1 U. 1 1 U.12 0.13 0.09 0.13 0.1 0.14 0.11 0.09 Nb ms 4 3.5 3.3 3.9 3.3 4.7 5.2 5.6 5.1 Rb ms 1 .J 1.3 1.5 1.2 1.3 1.7 1.9 1.8 1.8 Sn ms 1 .3 1.5 2.7 <0.5 0.7 1.7 1.9 1.1 0.9 Ta ms 4.2 5.7 3.8 5 4.5 4.7 5.7 5 10 Th ms 0.16 0.3 0.22 0.19 0.18 1 1.4 0.31 0.26 U ms 0.08 0.07 0.1 0.1 0.07 0.14 0.13 0.12 0.11 Zr ms 81 85 87 86 83 n.a. n.a. n.a. 92 La ms 3.0 3.2 3.1 3.2 3.1 6.5 7.3 4.9 3.9 Ce ms 9.1 9.4 9.3 9.4 9.2 17.0 18.0 15.0 11.0 Pr ms 1.5 1.6 1.6 1.7 1.6 2.4 2.6 2.3 1.7 Nd ms 9.2 9.2 9.4 9.7 9.6 14.0 15.0 14.0 10.0 Sm ms 3.3 3.4 3.4 3.5 3.5 4.8 5.1 4.7 3.4 cu ms 1 .J 1 ./ 1 .3 1.3 1.3 1.5 1.6 1.6 1.3 Gd ms 5.1 4.9 5.0 5.1 5.1 6.6 6.9 6.4 4.8 Tb ms 0.90 0.94 0.97 0.95 0.94 1.20 1.20 1.20 0.84 Dv ms 6.1 6.0 6.2 6.3 6.2 7.8 8.5 8.0 5.8 Ho ms 1.3 1.3 1.3 1.4 1.4 1.8 1.9 1.8 1.3 Er ms 4.0 4.0 4.1 4.0 4.0 5.0 5.2 5.2 3.7 Tm ms 0.57 0.58 0.61 0.58 0.58 0.70 0.79 0.74 0.53 Yb ms 3.8 3.6 3.9 3.9 3.8 4.9 5.2 5.0 3.7 Lu ms 0.55 0.54 0.58 0.55 0.55 0.69 0.75 0.71 0.53 I.a/Smrn^ 0.57 0.59 0.57 0.58 0.56 0.85 0.90 0.66 0.72 Analyzed at the GSC, Ottawa. ON.; "analyzed at the Univ. of Florida, Gainesville. FL.; 'values have been corrected for Si loss due to precipitation from analyzed solutions by extrapolation along a correlation between measured Si02 concentration and Total oxide abundance (r2=0.65) to a total of 99.7 wt%.; ""-method of analysis: @ byDionex Ion Chromatography; es by ICP-AES ms by ICP-MS xrf by x-ray fluorescence; n.a. not analyzed. All oxide concentrations are listed as wt.% and all elemental concentrations are as ppm '

PAGE 36

29 Table 2-3 continued. 2948-8R 2948-9R 2995-R2 2793-5R 2946IR 2946-2R 29471R2 2948-1 OR 2948IR SiO.' es 51.3 51.2 51.6 51.1 50.8 50.8 50.9 51.2 50.9 TiOj es 1.23 1.19 1.25 1.15 1.19 1.21 1.19 1.1 1.03 AI2O3 es 14 13.9 14 13.8 14.4 14.3 14.3 14.1 14.6 FejOjt es 11.6 11.3 11.7 11.3 11,2 11 11.1 10.8 10.8 Fe,0, es 0.9 0.9 1 0.9 1.2 1.1 0.9 0.7 0.6 FeO es 9.7 9.4 9.6 9.4 9 8.9 9.2 9.1 9.2 IVlllW es n 1 8 u. 1 0 n 17 u. 1 / n 1 9. u. 1 0 n 1 7 n 1 7 n 17 w. 1 / n 17 U. i D iVlgU es / Ab 7.34 7.33 TCI 7.85 7.55 7.86 1.1b 8.21 CaO es 11.6 11.4 1 1.5 1 1.6 11.9 11.6 1 1.7 11.7 11.9 Na.O es 2.44 2.46 2.56 2.56 2.49 2.5 2.51 2.3 2.16 K2O es 0.14 0.14 0.16 0.1 0.11 0.1 0.11 0.09 0.1 es 0.1 1 0.1 1 0.12 0.09 0.08 0.09 0.08 0.09 0.09 1 Olal es QO 1 yy. 1 Vo.Z yy .J Ofi A V5.4 QQ t yy. i QC 1 yQ.D OC Q yo.i Pn es 41 42 40 47 46 47 Cr es 240 250 230 200 300 290 310 270 320 Cu es 66 67 66 70 67 63 68 66 72 Ni es 66 66 64 68 85 80 90 78 94 Zn es 66 65 64 64 67 69 67 63 60 F @ 160 152 167 144 150 145 160 136 125 S @ 1176 1190 1197 1154 1142 1102 1107 1130 1 1 19 V xrf 342 335 340 345 320 327 325 325 303 Ba xrf 32 25 33 21 19 19 18 18 17 Sc xrf 48 47 48 45 50 54 48 47 42 TiO, xrf 1.41 1.39 1.43 1.33 1.34 1.42 1.36 1.27 1.16 Yrf Al 1 w. 1 z n n7 U.U/ U.Uo A An A AO U.UO 0.08 0.08 Cr xrf ZiZ z4U 221 196 310 305 318 286 331 Zr xrf 102 100 105 84 90 95 91 85 74 Sr xrf C\A 94 95 91 98 101 97 94 82 Y xrf 36 36 37 33 34 34 33 32 29 Rb xrf Z. 1 z.l 1.7 1.0 1 .2 1.4 0.8 1.3 1.5 Nb xrf A 1 4.6 4.5 2.6 2.9 3.6 3.1 3.2 4.0 Hf ms Z.D 2.3 2.6 2.2 2.3 2.4 2.3 2 1.8 In ms U.U9 A 1 1 0. 1 1 0.1 1 0.1 1 0.09 0.1 1 0.09 0.08 0.1 Nb ms A a. 4.0 4.3 4.6 2.8 3.1 3.6 3.5 2.7 4 Rb ms 1 . / 1.8 1.9 0.82 0.83 0.88 0.87 0.85 1 Sn ms u.y A 0 0.0 1 0.8 1.3 1 0.8 0.7 1.5 Ta ms 4.y 4.7 5 5 5.9 9.7 8 3.3 11 Th ms U.Z / 0.24 0.25 0.2 0.13 0.14 0.18 0.16 0.15 U ms U. 1 Z 0.1 1 0.15 0.07 0.07 0.07 0.08 0.07 0.06 Zr ms 90 89 99 79 87 88 87 73 69 La ms 3.8 3.8 4.0 2.6 2.8 2.7 2.9 2.7 2.6 Ce ms 1 1 n 1 1 .U 1 1 A 1 1 .0 1 1 .0 7.9 8.5 8.4 8.9 8.2 7.5 Pr ms 1 . / 1 .7 1.8 1.3 1.4 1.4 1.6 1.3 1.2 Nd ms 1 U.U 9.7 10.0 8.2 9.0 9.0 9.2 8.4 7.7 Sm ms 3.4 3.2 3.4 2.9 J. 1 2.6 Eu ms 1.3 1.2 1.3 1.2 1.1 1.2 1.2 1.1 0.9 Gd ms 4.9 4.7 4.9 4.4 4.6 4.5 4.5 4.2 3.9 Tb ms 0.90 0.87 0.92 0.80 0.80 0.87 0.82 0.76 0.69 Dv ms 5.9 6.0 5.9 5.4 5.3 5.4 5.5 5.4 4.7 1.1 Ho ms 1.3 1.3 1.3 1.2 1.2 1.2 1.2 1.1 Er ms 3.7 3.6 3.9 3.4 3.4 3.5 3.5 3.3 3.1 Tm ms 0.51 0.52 0.57 0.49 0.50 0.50 0.50 0.48 0.45 Yb ms 3.6 3.5 3.6 3.3 3.3 3.5 3.4 3.3 3.0 Lu ms 0.53 0.50 0.52 0.45 0.49 0.50 0.48 0.43 0.59 0.42 0.63 I .a/.Smfnl 1 » 1 J 0.70 0.75 0.74 0.56 0.59 0.53 0.59 • '"^ "I I iuuua, uainesviiie, rL.; values have been correcled for Si loss due to precipitation from analyzed solutions by extrapolation along a correlation between measured Si02 concentration and Total oxide abundance (r2=0.65) to a total of 99.7 wt%.: --method of analysis: @ byDionex Ion Chromatography; es by ICP-AES ms by ICP-MS xrf by x-ray fluorescence; n.a. not analyzed. All oxide concentrations are listed as wt.% and all elemental concentrations are as ppm '

PAGE 37

30 Table 2-3 continued. Samnle * 2948-2R 2948-6R 2948-7R 2949IR 2949-4R 2788-6R 2792-1 Rl 2792-2R 2792-81 SiOo ' es n.a. 50.5 51.1 51.2 51.3 50.2 49.9 50.0 50.8 Ti02 es n.a. 1.06 1.12 1.11 1.09 0.97 0.97 0.98 1.32 AI2O3 es n.a. 14.8 14.2 14.2 14.2 15.3 15.6 15.6 13.8 FeoOjt es n.a. 11.1 11.2 11.4 11.2 10.2 10 10.3 11.6 Fe^O, es n.a. 0.9 1.4 0.7 0.6 1.1 1 0.5 2.3 FeO es n.a. 9.2 8.9 9.6 9.5 8.2 0. 1 0.0 8 O.J IVIIIW es n.a. u. 1 / u. 1 / u. i / A 1 7 U. I / A 1 A U. 1 0 A 1 C U.l J A 1 A O.lo 0.18 MgU es n.a. 8.33 7.62 7.4 7.29 8.32 8.48 8.36 7.43 CaO es n.a. 12.1 11.9 11.7 11.6 11.7 11.5 11.8 11.5 NajO es n.a. 2.21 2.31 2.41 2.36 2.43 2.59 2.59 2.53 KjO es n.a. 0.12 0.11 0.08 0.1 0.06 0.07 0.07 0.1 1 es n.a. 0.09 0.09 U.UO n 07 U.U / U.UJ A C\A U.U4 A AA U.UO A 1 U. 1 I Ota! es n.a. 99.5 98.9 98.6 98.3 98.5 98.4 98.9 98.4 Co es 170 46 52 40 70 45 47 49 49 es 1 1 n jlU Z4U Ton zzu 01 A zzu 1 1 A 310 300 310 280 Cu es 67 72 64 67 67 82 77 oU 01 Ni es 61 94 66 60 62 1 10 120 1 10 1 1 u 7? Zn es 61 62 65 66 64 53 47 53 U 1 F @ n.a. 120 133 126 121 132 130 137 S @ n.a. 1121 1126 1233 1200 1074 1020 1017 1 1 7Q 11/7 V xrf 330 308 331 340 328 263 248 268 Ba xrf 19 21 16 17 21 14 10 13 22 Sc xrf 51 45 48 51 47 51 42 49 53 TiOj xrf 1.25 1.18 1.28 1.27 1 .25 1.10 1.11 1 1 1 1 K2O xrf n 07 U.UO U.Uo U.U / A A7 U.U/ 0.05 0.04 0.05 0.08 Cr xrf ZZ / "1 AC 245 225 224 354 319 345 301 Zr xrf TO 75 83 80 80 77 78 79 102 Sr xrf OJ 0 1 81 91 85 86 95 97 95 103 Y xrf il 29 33 33 33 31 30 31 38 Til* Kb xrf 1 A 1 .0 1 .4 0.7 0.7 0.6 0.6 1.1 0.9 Nb xrf 3.9 3.8 3.5 3.0 3.3 2.3 2.1 2.3 3.2 Hf ms 1 .9 2 1.9 2 2 1.8 2 2.1 2.4 In ms U.Uo 0.16 0.09 0.09 0.09 0.1 0.08 0.06 0.11 Nb ms 4.2 3.6 3.3 2.6 2.9 1.7 1.7 2.2 3.6 Rb ms 0.74 1 0.78 0.82 0.77 0.61 0.56 0.68 0.86 Sn ms 1 3.3 0.5 0.8 0.7 2 1 1.7 1.1 Ta ms 17 5.8 8.6 5.3 7.2 4.3 4.7 7.2 7.5 Th ms 0.12 0.16 0.14 0.1 1 0.14 0.13 0.14 0.15 0.19 u ms U.Uo A A"T U.U7 0.07 0.09 0.07 0.05 0.07 0.06 0.07 IS ms / 1 / 1 71 73 71 66 75 76 94 La ms 2.4 1.1 2.5 2.4 2.4 2.1 2.0 2.3 3.1 Ce ms 7.3 7.8 7.5 7.5 7.3 7.1 6.7 7.4 9.8 Pr ms 1 .3 1 .3 1.2 1.3 1.2 1.2 1.1 1.3 1.7 Nd ms 1.1 7.6 8.0 8.3 7.9 7.7 7.4 7.8 9.9 Sm ms 2.8 1.1 2.8 2.9 2.7 2.7 2.6 2.8 3.5 Eu ms 1.1 1.0 1.1 1.1 1.0 1.0 1.0 1.1 1.3 Gd ms 4.2 3.9 4.2 4.3 4.1 4.0 4.1 4.3 5.0 Tb ms 0.77 0.73 0.75 0.79 0.78 0.73 0.73 0.77 0.93 Dv ms 5.1 4.8 5.1 5.3 5.5 4.8 4.6 4.8 6.3 Ho ms 1.1 1.1 1.1 1.2 1.2 1.1 1.1 1.1 1.4 3.9 0.55 Er ms 3.5 3.2 3.4 3.4 3.4 3.2 3.1 3.2 Tm ms 0.48 0.44 0.45 0.49 0.45 0.46 0.43 0.46 Yb ms 3.2 3.1 3.4 3.2 3.2 3.1 2.9 3.1 3.7 0.55 A "^A Lu ms 0.45 0.43 0.45 0.46 0.45 0.43 0.43 0.48 0.45 n "i? 0.54 0.63 0.56 0.52 0.56 0.49 • ... ^ • .uMua, vjaiiicbviue, PL.: values nave been corrected for Si loss due to precipitauon from analyzed solutions by extrapolation along a correlation between measured Si02 concentration and Total oxide abundance (r2=0.65) to a total of 99.7 wt%.; .-method of analysis: @ byDionex Ion Chromatography; es by ICP-AES ms by ICP-MS xrf by x-ray fluorescence; n.a. not analyzed. All oxide concentrations are listed as wt.% and all elemental concentrations are as ppm '

PAGE 38

31 Table 2-3 continued. .Samnle I /y41 K z9o3-Kl TTOT t D Z/5/-IK i/9UIK 2945-5R 99GTVA-Ia 62DR2-I 62DR2-4a Endv(n=4i 2-siema SiO^ ^ es CI 1 CI 1 J 1.4 51.5 49.6 50.4 50.2 50.2 A 0 U.6 Ti02 es 1.33 1.2 1.24 1.21 1.25 1.35 0.62 0.6 1.39 0.02 AljO, es 13.4 15.1 13.9 13.8 13.8 14.3 14.6 14.6 15.1 0.2 Fe^Ojt es 12.5 11 1 1.7 1 1.5 11.7 12.3 9.9 9.9 10.3 0.2 FejO, es 0.8 0.8 1.4 1.1 0.9 1.4 1.1 0.7 2.3 0.1 FeO es 10.6 9.2 9.3 9.4 9.7 9.8 8 8.3 7.2 0.1 MnO es 0.19 0.16 0.18 0.18 0.22 0.19 0.16 n 1 s U.I/ 0.01 MgO es 7 8.68 7.24 7.33 1 0 1 /.zl T C C n OC 9. j4 1 A A /.44 0.17 CaO es 10.9 10.5 11.3 1 1.4 1 1.3 12 13.3 13.5 11.5 0.1 NaiO es 2.71 2.67 ? 4"? 9 44 2.54 2.73 1.54 1.54 2.99 n 1 Q KoO es 0.1 0.14 0.13 0.15 0.14 0.22 <0.05 <0.05 0.24 0.01 P2O5 es 0.09 0.11 0.1 0.11 0.1 1 0.12 0.04 n 04 0.01 Total es 98.2 98.7 98.5 98.5 Vs. / yy.z yy.u 99.0 n.a. n.a. Co es 43 89 47 67 49 /I /I 1 Cr es 130 300 180 190 180 no 460 450 265 6 Cu es 62 77 69 66 67 87 100 100 65 1 Ni es 57 160 62 61 62 43 93 93 76 1 Zn es 69 62 69 65 70 54 45 47 59 2 F @ 169 209 166 173 167 196 <50 <50 218 13 S @ 1313 1093 1 139 1 196 1287 n.a. n.a. n.a. n.a. n.a. V xrf 399 278 352 350 356 336 n.a. 228 13 290 Ba xrf 11 27 22 24 224 45 n.a. 16 2 81 Sc xrf A 0 43 50 49 49 46 n.a. 46 3 42 TIOt xrf 1.53 1.40 1.44 1.40 1.46 1.56 n.a. 0.66 0.03 1.57 K2O xrf 0.09 0.11 0.11 0.11 yj. ly n.a. U.Uz U.Uj 0.24 Cr xrf 91 266 175 184 168 212 n.a. 426 25 236 Zr xrf 94 146 98 94 n.a. n.a. 22 4 1 19 Sr xrf 90 115 99 100 1 1 A 124 n.a. n.a. 29 9 165 Y xrf 38 47 35 34 36 n.a. n.a. 18 1 34 Rb xrf 1.4 1.4 1.3 1.3 1.3 n.a. n.a. 0.0 0.9 4.3 Nb xrf 3.5 4.8 3.6 5.0 4.7 n.a. n.a. 0.7 1.1 7 0 Hf ms 2.5 3.6 2.3 2.3 2.4 3 1.6 1.4 2.8 0 1 In ms 0.16 0.16 0.1 0.1 0.1 1 <0.05 <0.05 <0.05 0.13 0.04 Nb ms 3 5.1 4.2 4.9 4.6 5.4 0.72 0.61 7.2 11 Rb ms 1 1.4 1.3 1.2 1.6 2.2 0.29 0.23 4.4 f) 1 Sn ms 1 4.1 0.7 1.1 1 0.7 4.8 <0.5 2.5 1.4 Ta ms 3.9 9.5 6.7 10 0 1 0.1 3.8 1.4 0.9 0.6 0.1 Th ms 0.17 0.27 0.22 0.22 0.26 0.31 0.06 0.05 0.72 0.46 U ms 0.08 0.1 0.11 0.11 0.13 0.14 0.06 0.04 0.18 0.01 Zr ms 91 n.a. 87 84 94 n.a. 56 47 n.a. n 3. La ms 2.9 4.4 3.5 3.4 3.6 4.8 0.70 0.70 6.8 1 0 Ce ms 9.4 13 10 9.9 10 14 2.3 2.4 17 L Pr ms 1.6 2.2 1.7 1.6 1.7 2.0 0.45 0.43 2.4 \J. 1 Nd ms 9.8 13.0 9.4 8.9 9.9 11 3.1 3.2 13 0 Sm ms 3.6 4.5 3.0 3.1 3.3 3.3 1.3 1.4 3.8 0.1 Eu ms 1.3 1.4 1.2 1.2 1.2 1.3 0.59 0.59 1.4 A A O.U Gd ms 5.1 6.0 4.8 4.4 4.7 4.6 2.4 2.5 5.0 0.1 Tb ms 0.93 1.1 0.84 0.80 0.86 0.84 0.47 0.49 0.87 0.03 Dv ms 6.5 7.1 5.6 5.4 5.9 5.2 3.2 3.2 5.4 0.1 Ho ms 1.4 1.7 1.3 1.2 1.2 1.1 0.73 0.72 1.2 0.0 Er ms 4.0 4.6 3.4 3.4 3.7 3.2 2.2 2.2 3.4 0.1 Tm ms 0.59 0.63 0.54 0.49 0.53 0.45 0.32 0.31 0.46 0.01 Yb ms 4.0 4.4 3.4 3.2 3.6 3.0 2.1 2.2 3.1 0.1 0.01 n a Lu ms 0.56 0.65 0.47 0.48 0.53 0.47 0.33 0.35 n.a. 0.44 n a 0.51 0.62 0.73 0.69 0.69 n.a. n.a. 01 norma, uainesville, f-L.; values have been corrected for Si loss due ,0 precipuatmn from analyzed solutions by extrapolation along a correlation between measured Si02 concentration and Total oxide abundance (r2=0.65) .0 a total of 99.7 wt%.; '-method of analysis: @ byDionex Ion Chromatography; es by ICP-AES ms by ICP-MS xrf by x-ray fluorescence; n.a. not analyzed. All oxide concentrations are listed as wt.% and all elemental concentrations are as ppm '

PAGE 39

rU c< in U o\ U o< lo o\ oo u ON fN u OS ^ m U OS U OS ON o m U Oi t o\ 00 (N U oi. as m (N U Qi as (N U ai Os O u as so u OS CO as cn r-p-r— — r-vooo — inoNvooo mo — r~-CNtomo<^OcNO(N dcddddddddddd O ^ _ O — CN IT) — • OO m d m tN v£) oo (N CM CNi _ m m O O jj, oc C O ~ ^>-)':t•r-^--oor-;^00^*p~ _ ^^i^--^— r~fN— md-^— •tNdcN'n m m m 00 — ON d o ^ o o o o o V V V — PoOTf— -p— ;mc>>noOoo — — " — OOOro n — o o P^OOTfropmTtpvDONr^ cs„m — iri — vO— 'rfdrn — ''^lilifNjriTt — u-i — \d — r-idrnfN — oom — rj-'d"*— 'rndtNrv£> t~~ 2 m ^ :2 2 ^, o\ P m (N 2 ^ ^a ^ 2: n <^ g O O O m ^ O O ^ § r'l <-i n tN ^ _i m O O d d g ;c! ° 2 ^ ^ o o ^ ^ m 00 — -t ONfN — ovpvooooo-^Ttin o>nfNdf<^dmdcNdcN fN so o •J^ ~" d m — o ~ o ^ o o >n in p O O d d ^ fN in V V f^ o oct^sDrnooasOsmr^yr-tsDQ^ — ONm — Tj-din — r^d(^ — CN ^ „ O O o o — i~-;^r^-^mC'*-^ — so — — Osr'^--Kn~\0—^-rfci-rt-~ — ^, ^ 1^ ^ ^ i^ <;::;2f^-S2o>oo — _: o ON ON ON m in in d in — m d m f^m'^--0_2 '^ — p in o 00 m On — 06 fN — Or-;OtDppTtOOo — -^d-T)— mdtN-lUa,Zc/2UJOHQ >> o E XI c3 I tU E>CO -:cs'^'^°°aNS gooo ^ (Nf-, of^::!oddS'~^" u I g £ S " -'^ 00 ON > 00 c o u > c u Q o u N ca c <

PAGE 40

33 2.3.4. Isotopic Analyses Isotopic ratios of Sr and Nd were measured, on a representative subset of 30 and 18 samples respectively, at the University of Florida using a VG Isomass 354 thermal ionization mass spectrometer (Table 2-5). Natural glasses to be analyzed were dissolved at 120°C in sealed Teflon vials using distilled HF and a few drops of distilled HNO3. Sr and rare earth elements were separated and collected using cation exchange chromatography (Dowex 50X12 resin), and Nd was separated from other rare earth elements (REE) using HCL elution on quartz columns packed with Teflon beads coated with bis-ethylhexyl phosphoric acid [after Richard et al., 1976]. Full procedural blanks are better than 0.2ng for Sr and 0.03ng for Nd. Sr samples were loaded with Ta oxide on W single filaments. Sr ratio data were acquired in triple collector mode at a beam intensity of 2X10"" A of ^*Sr, with corrections for instrumental discrimination made assuming ^^Sr/^^Sr=0.1 194. Long-term measurement of NBS 987 yields a ^^Sr/^Sr value of .710216 ± 45X10"^ (2-sigma). Nd samples were run as metals loaded on one side of a triple filament array using a Re center filament and Ta side filaments. Data were acquired at a beam intensity of 5X10"'" A of Nd, with corrections for instrumental discrimination made assuming '''^Nd/''*^Nd=0.7219. Long term measurement of the La Jolla Nd standard yields a '^^Nd/'^Nd value of .51 1834 ± 18X10"^ (2-sigma). 2.4. Regional CoAxial Segment Geochemical Characteristics 2.4.1 Major Element Trends MORE recovered from the CoAxial Segment chemical attributes characteristic of "normal" incompatible element depleted MORE [Eryan et al., 1976; Sun et al., 1979;

PAGE 41

Table 2-5. Sr and Nd Isotopes for Representative Samples. Sample ID Lat. N. Long. W. ^'Sr/''Sr Nd/ Nd Location Zz 1 -U44 J A £i CIO 46.512 ion c 0 zr 129.586 0.513214 1 Z07U-4 46.526 129.579 0.702262 0.513248 1 2672-2 46.523 129.581 0.702263 0.513218 1 2792-3 46.532 129.574 0.702254 n.a. 1 2672-5 n.d. n.d. 0.702325 0.513178 2 2672-6 46.521 129.573 0.702328 n.a. 2 2792-4 AH C f\ 46.530 129.572 0.702300 n.a. 2 2965-1 A H CO A 46.524 129.569 0.702312 0.513179 2 A £i CO /I 46.524 129.571 0.702306 0.513187 2 OH/l Oil zy4o-i 1 A c 0 r\f\ 46.300 129.720 0.702337 0.513188 3 on/I 0 0^ 294o-or A H. orv 1 46.291 129.724 0.702329 n.a. 3 294o-9r2 AH '^r\/' 46.296 129.721 0.702331 n.a. 3 94rc20 46.302 129.719 0.702349 0.513127 3 221-U551 AH £"10 46.513 129.587 0.702266 0.513145 5 26/1-2 A H C f\C\ 46.509 129.589 0.702238 0.513195 5 AC c orv 46.530 129.567 0.702304 n.a. 5 Zjj-U/UU 46.274 129.744 0.702302 0.513215 4 z6/y-l 1 /I ^ 0 1 0 46.313 129.704 0.702235 0.513192 4 OQ/IQ in zy4o-lU A £i or\o 46.298 129.720 0.702273 0.513217 4 2995-3 AH ^ A C 46.345 129.560 0.702322 0.513165 4 2674-1 46.159 129.802 0.702390 n.a. 6 2675-2 46.178 129.797 0.702369 0.513112 6 2678-3 n.d. n.d. 0.702314 n.a. 6 2945-5 46.156 129.810 0.702321 n.a. 6 94rc25 46.038 130.013 0.702501 0.513120 9* y^i Lzo /I A 1 AO 40. 1 0/ 1 on no 1 129.921 f\ A f\r\ 0.702499 0.513090 9* 99GTVA-la 46.162 129.921 0.702505 n.a. 9* 62DR-1 46.387 129.473 0.702333 n.a. 11 62DR-2 46.387 129.473 0.702321 0.513199 11 62DR-4a 46.387 129.473 0.702308 n.a. 11 n.d. no data; n.a. not analyzed.

PAGE 42

35 Schilling et al., 1983]. Figure 2-3 shows major element concentrations of 139 natural basaltic glasses from the CoAxial Segment analyzed by EMP. The basalt samples range in MgO content from approximately 9.6 to 6.5 weight percent (wt.%). Most of the chemical variability can be ascribed to the effects of fractional crystallization, but dispersion of data at a given MgO value is in excess of analytical uncertainty. Consequently, some variability in parental composition and/or conditions of crystallization is required. Major element data from CoAxial lavas (Figure 2-3) form a field with a trend generally parallel to modeled low pressure fractional crystallization trends, with the exception of K^O and P2O5 which increase in concentration at a greater rate than predicted by simple fractional crystallization models [Weaver and Langmuir, 1990]. The apparent greater enrichment of K2O and P2O5 could be indicative of more complex crystallization processes such as in situ crystallization [Langmuir, 1989]. However, deviation from the modeled fractional crystallization trend also may be an artifact of the greater incompatible element depletions (relative to most CoAxial samples) present in the mafic sample used for the parent composition in the liquid line of descent (LLD) model shown in Figure 2-3. Examination of geochemical data for MORE recovered from the CoAxial Segment and comparison to other data from other provinces of the southern and central Juan de Fuca Ridge reveals several notable relationships. Comparison of CoAxial lava chemistry to that of the Cleft Segment and Axial Seamount Province (which includes Axial Seamount caldera, north rift zone, and south rift zone) shows Coaxial lavas to be distinct in several ways. With the exception of one andesitic sample (not shown) recently

PAGE 43

c3 O 3 a S .2i E ^ 60 . on ctn :2 >< r ^3 U n 2: -2 (N ^ Hh =3 J5 J T3 60 ^ ~ c3 j= CS E 00 5 > 00 c/3 3 o o X3 3 a c 3 a< c -S Oj "O C cS "O CJ 00 c3 0 l> c 1 E.S CS 2 — to 60 CS L. ^ CS

PAGE 45

38 AI2O3 CaO

PAGE 46

recovered from the North Rift Zone [Perfit et al., 1998] and some small on-axis volcanoes discussed below, glass data from the Axial Seamount display a more limited range in MgO, (-8.4 to 6.7wt%) than glasses from CoAxial (9.6 to 6.5wt%). Axial Seamount data from Rhodes et al. [1990] are not shown, but display a more restricted range in MgO content (6.9-7.7wt% MgO). In addition to having a greater range in MgO content, the CoAxial sample suite generally exhibits greater variation in other major element oxides at a given MgO than does the suite from Axial Seamount Province. It is, however, important to note that the Axial Seamount data set is smaller (81 glasses) than the CoAxial Segment database (139 glasses). As illustrated on MgO discrimination diagrams (Figure 2-3), there is significant overlap of CoAxial MORE with the field defined by Axial Seamount volcanics, but CoAxial lavas tend to have lower abundances of AI2O3 and CaO at a given value of MgO and extend to higher values of SiOi and FeO. Notably, less than -10% of the CoAxial MORE have K2O or NaaO concentrations that fall within the Axial Seamount field. Comparison of CoAxial MORE with MORE recovered from within the axis of the Cleft Segment, the southernmost segment on the Juan de Fuca Ridge [Smith et al., 1994] is useful for two reasons. Firstly, the Cleft segment represents one of the least morphologically and tectonically complex ridge segments on the Juan de Fuca spreading system, and it is somewhat spatially removed from complexities related to the Cobb melting anomaly. Secondly, the two MOR segments immediately adjacent to the CoAxial segment, the Vance segment to the south and the Cobb segment to the north, are very poorly petrologically characterized. Although data are very limited [Smith et al., 1994; Rhodes et al., 1990; Karsten et al., 1990; Delaney et al., 1981; our unpublished

PAGE 47

40 data], the Cobb and Vance segments appear to have major element characteristics similar to that of the Cleft segment. Compared to Cleft samples, CoAxial MORB generally have lower abundances of AI2O3 and higher abundances of Na20 at a comparable MgO abundance (Figure 2-3). 2.4.2 Isotopic and Trace Element Signals Trace element concentrations in CoAxial Segment lavas have strong correlation to MgO suggesting that much of the trace element compositional variation observed within the CoAxial suite can be attributed to fractional crystallization and other differentiation processes. Additionally, covariation between trace elements that is inconsistent with crystallization or assimilation processes requires some variation in parental composition of lavas from this segment. In particular (as is discussed in greater detail in section 5.2.2), geochemical enrichments in a lava flow erupted at the Flow Site between 1981 and 1991 greatly extend the range of some trace element abundances, requiring differences in parental lava source characteristics. Discounting these few samples, the most incompatible enriched sample in the CoAxial Segment data field has an abundance of moderately incompatible elements (e.g. Y, Zr, Sm) 40% greater than the lowest abundance sample. Comparison of the most depleted and enriched CoAxial MORB shows that the abundances of the most highly incompatible elements (e.g. Nb, Ba, and Rb) are enriched by 250% to 300% relative to the lowest abundance sample (Table 2-3). Low-pressure fractional crystallization models [Weaver and Langmuir, 1990] indicate that -25% fractional crystallization of the most MgO-rich sample is required to produce the range in MgO values observed in Table 2-3 (note that the range of MgO values in the sample subset shown in Table 2-3 is less than

PAGE 48

41 that observed in the entire suite analyzed by EMP). Maximum corresponding enrichments in trace element concentrations (by assumption of a bulk distribution coefficient equal to 0) expected from fractional crystallization are limited to a factor of ~1.3x the parent composition when 75% of the original liquid remains. Consequently, over enrichment of the most incompatible elements relative to that expected from fractional crystallization are observed. Such incompatible element over enrichments in MORB suites (relative to fractional crystallization models) have been explained via more complex crystallization processes such as in situ crystallization [Langmuir, 1989], but may also be in part due to differences in primitive parental magma compositions. Figure 2-4a shows latitudinal variation in the ^^Sr/^^Sr isotopic ratio of Juan de Fuca Ridge lavas. Data shown are a compilation of our data (Table 2-5 and unpublished data for Axial Seamount Province and the near ridge Vance Seamounts) and previously published data [Eaby et al., 1984; Ito et al., 1987; and Hegner and Tatsumoto, 1987]. ^^Sr/^^Sr data reported by Rhodes et al. [1990] (eight analyses) are not included in the compilation because they have values 10-20 XIO"'' lower than the main data trend. This offset to of the Rhodes et al. [1990] data lower ^^Sr/^^Sr values is confirmed by Eaby et al. [1984] who report analyses of some of the same samples. CoAxial MORB (latitudinal range indicated by vertical dashed lines in Figure 24) are clearly less radiogenic than MORB from the rest of the JdFR (and Pacific MORB in general). CoAxial Sr isotopic data (n=25) range from 0.70224 to 0.70239, whereas values for the rest of the Juan de Fuca are between 0.70240 and 0.70268. The less radiogenic nature of CoAxial ^'Sr/^^Sr relative to the rest of the JdFR is clearly illustrated in Figure 2-4b, which shows *'Sr/^^Sr values averaged by region. The average CoAxial

PAGE 49

Figure 2-4 : (A) Distribution of Sr/ Sr for Juan de Fuca Ridge basalts plotted against latitude (°N). Symbols and sources are as in the legend and text. Two-sigma error bar is shown at left. Vertical dashed lines at bottom mark CoAxial Segment boundaries, and the vertical solid line the Cobb offset. (B) *^Sr/^Sr compositions averaged by region. Error bars are ± 1 standard deviation for each average. Symbols from left to right are averages of data from: the southern Cleft Segment, the northern Cleft Segment, the Vance Segment (including Vance Seamounts), Axial Seamount Province (see text), the Coaxial Segment, the Cobb Segment and the Endeavour Segment. The number of analyses for each average are shown next to the data point. Line is a linear regression through the data points excluding the CoAxial and Endeavour Segments which respectively display anomalously depleted and enriched lavas chemistries relative to other JdF data. See text for discussion.

PAGE 50

43 < Ito et al, 1987 1980s Flow (n=5) 0 Hegner and Tatsumoto, 1987 1980s Floe (n=4) 0 Our unpublished JdF Axis 1993 CoAxial (n=4) Eaby et al., 1984 O Rogue Volcano X Our unpublished Vance Smts © Recent NRZ Flows 0.70270 I — • — r 0.70220 ' — ' — ' — • — ' — • — • — ' — — ^ I , . 44 45 46 47 48 Degrees Latitude North

PAGE 51

44 Segment ^'Sr/^Sr value of 0.70231 is > 15x10"^ below the trend established by the other JdFR data south of the Cobb offset. It is interesting to note that although the overlap in data from the different regions is significant; there is a trend of decreasing '^^Sr/^^Sr with increasing latitude. This trend is also apparent in the unaveraged data (Figure 2-4a) suggesting that this trend is not simply an effect of the averaging of ^^Sr/^^Sr. Regional surveys of JdFR Sr isotopes [Baby et al., 1984, Rhodes et al., 1990] have noted that Axial Seamount does not have a Sr isotope signature distinct from the rest of the JdFR, but a regional trend in ridge axis lavas has not previously been recognized. It is important to note that while this trend is intriguing, it should be considered with caution. Significant overlap exists in the range of ^^Sr/^^Sr values observed in lavas from different regions of the JdFR, and latitudinal variation in ''^''Nd/''"Nd does not display a similar linear correlation (though Nd data are more limited in quantity and spatial distribution). CoAxial Segment lavas are not only distinct from other JdFR lavas with respect to Sr isotopic ratios, but also with respect to Sr abundance. Magmatic strontium abundances are not greatly affected by typical MORE crystallization (bulk liquid-crystal distribution coefficient is near unity), and in the absence of significant disequilibrium crystal accumulation or fractionation, large differences in Sr abundance between different lavas are likely a trait acquired during melt generation. A histogram of Sr abundances for MORE recovered from the Cleft Segment, CoAxial Segment, and Axial Seamount (Figure 2-5) shows the data-ranges for each of the three regions to be quite distinct from one another. As with the Sr isotopic ratios, CoAxial Segment lavas are the most depleted (in Sr abundance) on the JdFR. Only four samples from the CoAxial Segment (all among the oldest samples recovered) overlap with the data-range for Cleft Segment lavas, and

PAGE 52

45 40 00 ^ 30 c < O 20 no»r)'^' in o in o in O OO OO a\ ON Sr (ppm) Figure 2-5: Histogram showing distribution of XRF analyzed strontium concentrations (ppm) in CoAxial Segment, Axial Seamount Province, and Cleft Segment basalts. Data used for the Cleft Segment is that of Smith et al., 1994. Data for Axial seamount province is a compilation of unpublished data of Perfit et. Al. (in prep) and that of Rhodes etal., 1990. See text for discussion.

PAGE 53

46 only two Axial Seamount samples overlap with either the CoAxial or Cleft segment ranges. Strontium abundances of 96% of the CoAxial lavas analyzed fall between 79 and 104 ppm [average = 94 ppm], while 96% of Cleft Segment lavas have Sr between 105 and 121 ppm [average=l 14 ppm]. Axial seamount lavas have a Sr abundance range distinct from either two, with 97% of lavas having a Sr abundance between 130 and 170 [average=147 ppm]. Regional basalt isotopic and trace element signatures will be considered further below. 2.5. Discrimination of Discrete Flow Units 2.5.1 Visual and Bathymetric Flow Discrimination Three different, recently erupted, mappable units have been identified on the CoAxial Segment. Assessment of differences in depth-data from repeat sea beam surveys of the segment clearly shows bathymetric anomalies corresponding to three different regions of very fresh lava mapped by submersible, ROV, deep camera tow, and AMS 60 deep tow sidescan sonar [Chadwick et al., 1995, Embley et al., in press]. Figure 2-2 shows AMS60 images for the Floe (2-2a) and Flow sites (2-2b) with outlines depicting the interpreted boundaries of the recent eruptives (after Embley et al. [in press], Figures 8 and 1 1). Superimposed on the sidescan images are the locations of basalt samples discussed below. Timing of the emplacement of these units is constrained to two intervals by the acquisition dates of the bathymetric data, the first occurring between 1981/82 and 1991 and the second occurring between 1991 and 1993. The lava associated with the SOSUS monitored T-phase seismicity [Embley et al., 1995] spatially corresponds with a sole bathymetric feature created between the 1991 and 1993 surveys, supporting this lava flow's association with the 1993 seismic event. This lava flow,

PAGE 54

hereafter referred to as the 1993 flow, is approximately 3.8 km long, and up to ~500m wide and ~30m thick. Morphology of the 1993 flow is dominated by lobate to small pillowed forms. While most of the volume resides in the main body of the flow, several small segments of 1993 flow crop out north of the main body in a small graben that extends north and south of the eruption. This graben presumably served as the eruptive vent for the 1993 flow [Chadwick and Embley, 1998; Chadwick et al., 1995; Embley et al., in press]. Detailed descriptions of the individual flows and eruptive sites (Figures 2-1 and 2-2) are presented in Embley et al. [in press]. Comparisons between the 1981 and 1991 bathymetric data (Flow Site) and the 1982 and 1991 bathymetric data (Floe Site) reveal two additional positive bathymetric anomalies generated prior to the 1993 volcanic event. Seafloor observations confirm predictions of recent volcanic activity based on bathymetric differences between surveys, and identify two additional very fresh-looking volcanic constructional features. The first, hereafter referred to as 1980s Flow, is at the Flow Site approximately 700 meters east of the 1993 lava flow. It is of a similar morphology and size to the 1993 Flow [Chadwick et al., 1995]. The second, hereafter referred to as 1980s Floe, is at the Floe Site about 500 meters west of the zone of active venting and interpreted path of the 1993 dike emplacement [Embley et al., in press]. This flow (and associated bathymetric anomalies) consists of a linear array of several elongated mounds aligned with their long axis parallel to ridge strike. These fresh lavas occur over a distance of at least 7 km, and field relations indicate that these lavas were erupted along a pre-existing topographic ridge. While bathymetric anomalies were identified coincident with the thickest part of the recent flow, absence of bathymetric anomalies where glassy lavas were observed implies

PAGE 55

that this flow is, in places, thinner than the 5-15 meter resolution of the bathymetric differencing technique [Embley et al., in press.]. The nearly complete overlap in the temporal ranges defined by the three surveys prior to 1993, 1981-1991 for the Flow Site and 1982-1991 for the Floe Site, limits the ability to discriminate whether these two older (relative to the 1993 flow), but still recent, lava flows were emplaced during one or two separate crustal accretion events. Although field observations of sediment cover suggest that the 1980s Floe eruption may be slightly older than the 1980s Flow eruption, these data also cannot preclude the possibility that both eruptive units are related to a single crustal accretion event. An examination of the interand intraflow geochemical variation further constrains the genesis of these lavas and seafloor constructional features. 2.5.2 Geochemical Flow Discrimination Examination of the data for the three recent CoAxial Segment lava flows, as shown in Figure 2-4a, reveals several important features. Firstly, for each of the eruptive units there is no statistically significant intra-flow variation in **'Sr/^'^Sr (or in '''-''Nd/''*^Nd, not shown), although lavas from the segment as a whole do exhibit significant variation. Secondly, the 1980s Floe and Flow site eruptives cannot be distinguished from one another using ^^Sr/^^Sr, although the 1993 Flow (number of samples analyzed (n) = 4) is distinctly less radiogenic than both the 1980s Floe (n=4) and 1980s Flow (n=5) eruptive units. There are no obvious crustal components or processes that could lower the Sr isotopic ratio of the 1980s magmas to produce the 1993 lava. Accordingly, the 1980s CoAxial lavas and the 1993 CoAxial lava cannot be easily related to a single parent, and are, therefore, neither comagmatic nor cogenetic. Further

PAGE 56

discrimination of the three identified recent CoAxial eruptive units requires an examination of elemental abundances. Electron microprobe data of natural glass samples recovered from each of the three recent eruptive units, along with average compositions and standard deviations for each unit, are shown in Table 21 . Averaged data and analytical precisions for two MORE standards measured during the analytical runs are shown in Table 2-2. 2.5.2.1 The 1980s Floe Site Eruptive Glasses from the three recent CoAxial flows exhibit both inter-flow and intra-flow major element chemical variability (Figure 2-6). The 1980s Floe Site eruptive, which is an ~7 km long linear array of elongate mounds, exhibits the greatest overall variation in major element concentrations. Data from samples collected in situ range from 7.5 to 7 wt.% MgO. The inclusion of wax core sample 94RC19, which targeted the southernmost mound of this flow (Figure 2-2a), extends the data to 8wt% MgO. This intra-flow variation, while not great (-12% relative to concentration), is well outside of analytical error, which is less than 2%. Perhaps more significant than the overall extent of intraflow MgO variability is the co-variation between the major element oxides occurring in samples from this unit. Figure 2-6a shows liquid lines of descent (LLDs) calculated for fractional crystallization at Ikbar pressure. A key to symbols is given in the figure caption. The models were generated using both the original program of Weaver and Langmuir (dashed line) [Weaver and Langmuir, 1990], and a more recently modified version of the program that includes the oxides K2O and P2O5 in the calculations (solid line) [Reynolds, 1995]. Each of the algorithms of Weaver and Langmuir utilize a combination of experimental

PAGE 57

50 15 14 13 T3 ^ 10 2.75 2.25 .75 .25 — 1 — ' — 1 — ' — 1 — ' — r— ' T A rtr^^^ ^^^Al203 CaO 0 <^ &9-^ 1 . 1 . r . 1 . 1 . FeO 'i' — ' I.I.I Na20 — 1 — . — 1 — . 1 , \ . 1 Ti02 I.I.I 15 14 13 12 11 0^ l kC> j = 5 1 10 2.75 2.25 1.75 .25 6.8 7 7.2 7.4 7.6 7.8 8 8.2 MgO 6.7 6.i 6.9 7 MgO 7.1 7.2 7.3 Figure 2-6: MgO concentrations plotted against concentration of AI2O3 (squares), FeO (triangles), CaO (circles), Na20 (diamonds) and Ti02 (crosses) for the 1980's Flow Site eruptive (open symbols), 1980's Floe Site eruptive (black filled symbols) and 1993 Flow Site eruptive (gray filled symbols). Error bars are shown for the highest MgO content sample in each of the three flows, and represent ± 2 standard deviations (2-sigma) of our standard data. In general, 2-sigma precisions (relative to concentration) are equal to or better than: 0.5% for SiOa, 1.5% for FeO* (total Fe as FeO), MgO, CaO and AI2O3, 3% for Ti02, 5% for Na20 and K2O, and 10% for P2O5 and MnO. (A) Comparison of glass data for the two flows erupted between 1981 and 1991. Trend lines show theoretical 1 kilobar liquid line of decent (LLD) calculations using the original program of Weaver and Langmuir, 1990 (dashed line) and a recently updated version of the same program designed to include K2O and P2O5 (solid line) (See text). (B) Comparison of the two flows recently erupted at the Flow Site. Trend lines (generated using the updated LLD program) show the effects of pressure on modeled LLD paths with trend lines calculated at both 1 (solid) and 2.5 (dashed) kilobars pressure.

PAGE 58

51 and thermodynamic data to predict the chemical evolutionary path, or liquid line of descent (LLD), of MORB liquids as they cool and crystallize mineral assemblages in chemical equilibrium with their host liquids. The algorithms are designed to model liquid chemical evolution during both closed system (equilibrium crystallization) and open system (fractional crystallization) crystallization processes. However, over the limited range of crystallization relevant to the intra-flow chemical variation observed, there are no significant differences between the LLDs generated using either the equilibrium or fractional crystallization models. The parent lava composition used for the calculations shown in Figure 2-6a is 94RC19, the southernmost and most mafic (MgO-rich) sample recovered from the 1980s Floe site eruptive. Five oxides plotted against MgO (AI2O3, CaO, FeO*, NaaO and Ti02) show covariation trajectories parallel to and bounded by the two model LLDs. The primary difference between the two models occurs because plagioclase's arrival onto the liquidus is suppressed in the model that considers effects of K2O and P2O5 in calculating the LLD path. The model shown in Figure 2-6a calculated using the original Weaver and Langmuir program (dashed line) has crystallization of clinopyroxene (cpx) as the sole liquidus phase beginning at a temperature of 1 192°C. After only 4° of cooling (1 188°C) both olivine (ol) and plagioclase (pi) begin crystallizing as the liquid becomes triply saturated. Using the same data to run the newer model (solid line) produces a liquid that once again begins with cpx as the sole liquidus phase, but crystallization now begins at a temperature 10 degrees higher (1202°C) than in the older (no K2O or P2O5) model. Cpx remains the sole liquidus phase through 10 degrees of cooling at which point olivine joins the liquidus. Plagioclase begins crystallizing at the same temperature as in the older

PAGE 59

model (1 188°C), but by that time approximately 5% crystallization has already occurred. The effects of delayed plagioclase crystallization during the first 5% of crystallization are most dramatically seen in the LLDs for FeO (FeO enrichment is initially suppressed) and AI2O3 (AI2O3 shows greater initial enrichment). In general, the newer model does a better job of reproducing the natural variability observed in the 1980s Floe glasses, although our AI2O3 data suggest that initial increases in AI2O3 concentration are not as great as are predicted from the model. TiO^ is the exception to the other major element oxides, showing better agreement with the older model. The models' suggestion of clinopyroxene as a sole liquidus phase at low pressures in such relatively differentiated MORE is to be questioned. There is no experimental evidence that supports cpx occurring as a sole liquidus phase in such evolved compositions. Though small amounts of water may have the capability to suppress plagioclase crystallization [Michael and Chase, 1987], the algorithms of Weaver and Langmuir do not incorporate H2O concentrations as a variable. Moreover, the 10 degree maximum interval of sole cpx crystallization is near the limit of experimental reproducibility [Grove et al., 1992]. It is, therefore, likely that sample 94RC19 was multiply saturated at its liquidus temperature. The close agreement of the data shown in Figure 2-6a with the predicted LLD trends suggests that elemental variations in glasses from the 1980s Floe eruption are consistent, within the limitations of current experimental models, with evolution along a liquid line of descent resulting from progressive cooling and crystallization. Thus, the chemistry of the sample recovered from the southernmost of the 1980s Floe mounds (94RC19) is consistent with it being cogenetic, comagmatic, and parental to the rest of

PAGE 60

53 the samples from the 1980s Floe eruptive. Mineral phases observed in the lavas are consistent with the LLD models, and the total range of MgO compositions observed in the 1980's Floe Site eruptive corresponds to approximately 24° of cooling and 20% crystallization, as predicted by the LLD model (more recent version). Trace element data are not from pure glasses, but instead are of "whole rock glasses" (which include a small amount of crystals which may not be a truly equilibrium assemblage). Accordingly, it is difficult to use our trace element data to rigorously constrain the crystallization conditions modeled, but it is nonetheless clear that some degree of liquid evolution has occurred within this flow unit. Whether this variation truly represents the process of fractional crystallization or is simply reflecting differences in the evolution of the liquid during equilibrium cooling and crystallization cannot be adequately assessed utilizing the glass data alone. Trace element data gathered by XRF on whole rock glasses can be used to help constrain to what extent crystallization processes responsible for liquid differentiation within the 1980s Floe unit were closed (equilibrium crystallization) or open (fractional crystallization) system processes. If liquid differentiation observed within the glass data results from fractional crystallization processes, then composition of the whole rock glasses (crystals+liquid) should evolve with the liquid composition. Comparison of covariation of glass MgO and "whole rock glass" Zr concentration with latitude (Figure 2-7) reveals some interesting relationships. In this case, because the trend of the long axis of the 1980s Floe lava emplacement is aligned roughly parallel to ridge strike at -022°, latitude is proportional to distance along strike of emplacement. A strong correlation exists between MgO content of glasses and distance along strike of the flow.

PAGE 61

Figure 2-7: Latitudinal variation in MgO (upper) and Zr (lower) concentrations for basalt samples from the 1980s Floe Site eruptive. Symbols are keyed to method of recovery and are as shown in the legend. Error bars (±2-sigma) are shown, as are r" values for linear regression lines. Linear regressions are based on all the data shown. Though latitudinal increase in whole rock glass Zr concentrations (XRF) are similar to trends seen in the BMP data (upper) supporting the hypothesis of geochemical differentiation along the strike of flow emplacement, Zr increase is insufficient to be the result of pure fractional crystallization processes.

PAGE 62

55

PAGE 63

with MgO decreasing (cooling and crystallization increasing) from south to north. A similar, though less well defined, latitudinal covariation seems to exist in the Zr data, though the lack of "whole rock" data for the wax core 94RC19 limit the total variation observed. Nonetheless, if one extrapolates along the Zr trend (which over this limited range approximates a linear trend) to arrive at a Zr value for this sample, one would postulate it to have approximately 97 ppm Zr. Using this value as the parent composition, it becomes clear that the increase in Zr from 97 to 106 ppm (< 10% increase relative to concentration) is not sufficient to invoke pure fractional crystallization as the differentiation process. This small an increase would require a liquid-crystal bulk distribution coefficient (bulk D) of approximately 0.57, a value far too high for an element as incompatible during MORB crystallization as Zr (a "bulk D" equal to zero would yield a 25% increase in concentration after 20% crystallization.) Based on the variations of "whole rock" glass Zr data (and other highly incompatible trace elements not shown) it is likely that some intra-flow magmatic differentiation along the strike of emplacement exists. The differentiation process, however, was inefficient and cannot be accurately modeled via true Rayleigh fractionation [Shaw, 1970]. More likely, liquid differentiation was a process intermediate between closed system equilibrium crystallization and pure fractional crystallization. 2.5.2.2 The 1980s and 1993 Flow Site Eruptives As discussed above, two discrete lava flows erupted between 1982 and 1993 have been identified at the Flow Site. Major element compositions of glasses recovered from the 1980s Flow unit (Figure 2-6 a and b, open symbols) and 1993 Flow unit (Figure 2-6b, gray filled symbols) show compositional variations similar to, but more restricted than,

PAGE 64

57 those from the 1980s Floe unit (shown only in Figure 2-6a). LLD models depicted by the trend lines in Figure 2-6b show the affects of varying crystallization pressure between 1 (solid line) and 2.5 kilobars (dashed line). In both models the starting composition is that of sample 2792-7, the most MgO-rich sample recovered from the 1980s Flow Site eruptive. Observed mineral phases and major element covariation with MgO in both of the Flow site units are consistent with LLD models and small amounts of liquid evolution (~9°C of cooling and 2-5% crystallization). Although the total intra-flow chemical variation (4-5% relative to concentration for MgO) for these two units is close to analytical error, the consistent behavior between the different oxides and the LLD parallel trends are compelling. That being said, there is an apparent decrease in MgO and increase in Zr with latitude in the 1993 Flow (not shown), similar to the recent Floe Site eruptive, although only the northernmost and southernmost samples are completely distinguishable from one another considering the 2-sigma uncertainties. Neither of the two recent Flow Site eruptive units occurs over as great a distance along strike as does the 1980s Floe unit. The limited length of the Flow Site units, combined with lesser overall geochemical variability and restricted sample coverage for the 1980s Flow eruptive, inhibits the ability to define along strike chemical variability for these flows. Examination of Figure 2-6 does, however, clearly suggest that the two 1980s units are not cogenetic to one another, a conclusion that could not be drawn from the bathymetric or isotopic data. Relative to the 1980s Floe unit, the 1980s Flow unit exhibits markedly higher Ti02 and FeO* and lower CaO. In fact, as can be seen in Figure 2-3, Ti02 and CaO values for the 1980s Flow site unit appear to be rather anomalous compared to lavas from

PAGE 65

58 the southern JdFR in general. Additionally, an examination of Table 2-3 reveals that, compared to the other recent CoAxial lavas that are both more and less geochemically differentiated, this basalt flow exhibits relatively high abundances in Zr, Hf, and to a lesser extent Nb. Elevated rare earth element (REE) abundances (relative to similarly differentiated samples from the segment) are also seen in samples from the 1980s Row unit (Figure 2-8b, open circles). For example, REE patterns from the 1980s Flow unit are similar to, but have slightly higher chondrite-normalized abundances than those in either the 1993 Flow unit (Figure 2-8b, gray filled diamonds) or the 1980s Floe unit (Figure 28a). Abundances of the middle REE are about 20X chondrites, on average, in the 1980s Flow Site unit. That is approximately 30% higher than those from the Floe site unit (~15X chondrites) which has an average whole rock glass MgO content similar to that of the 1980s Flow unit. In addition to the elevated REE abundances, 1980s Flow samples have chondrite-normalized La/Sm (La/Smn) values (0.66-0.68) that are high relative to other Flow site lavas (0.48-0.62; n=22), but well within the CoAxial segment range in general (0.48-0.75, n=52). Examination of the major and trace element data for the 1993 flow unit (Figures 2-6 and 2-8, Tables 2-1 and 2-3) reveal it to be the most depleted among the recent flows in incompatible elements, and among the most depleted along the entire CoAxial segment as well. While the 1993 flow is the most geochemically evolved (lowest MgO content) of the three recent CoAxial flows discussed, it has similar NaiO contents to the more primitive 1980's lavas and the lowest overall concentrations of the elements Zr, Nb, Hf, Sr, Rb, K2O, U and REE. Correspondingly, the 1993 flow samples are among the most light-REE depleted MORE recovered from the CoAxial Segment axis, having La/Sm„

PAGE 66

Figure 2-8: Chondrite-normalized (Taylor, 1982) rare earth element (REE) concentrations for MORB glasses from the 1980s Floe Site eruptive (upper figure; 6 samples analyzed), the 1980s Flow Site eruptive (lower fig.-open circles; 4 samples analyzed) and the 1993 Flow Site eruptive (lower fig. -gray triangles; 7 samples analyzed). All data were analyzed by ICP-MS at the Geological Survey of Canada with the exception of two wax core samples from the Floe Site (upper fig. -dashed lines) which was analyzed by ICMMS at the USGS (Denver). Note the y-axis is an expanded exponential scale ranging from 5X to 40X chondrites. La and Ce are anomalously high in two of the 1980s Flow Site samples (lower figure-dashed lines), and may be suspect. Excellent intra-flow consistency is seen in the REE data, but variability occurs between different flows. Samples from the 1993 flow have the lowest overall REE concentrations and La/Smn (0.56-0.59), despite being the most geochemically evolved.

PAGE 67

60 40 1 — I 1 1 1 1 1 1 1 1 1 1 1 1 r 20 5 I— J 1 1 1 1 1 1 I I I I I I r La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Element

PAGE 68

61 ratios between 0.56 and 0.59. In addition, as shown in Figure 2-4a, the 1993 eruptive also has the least radiogenic Sr isotopic signature of the three recent CoAxial lavas sampled. All of these observations suggest that the primary magma parental to the 1 993 lava flow was more depleted in mantle magmatophile elements than other recent CoAxial lavas sampled. 2.6. Distinction of Regional Petrochemical Provinces The comparison of minor and trace element characteristics of CoAxial lavas to those from other geologically identifiable terrains along the central JdFR suggests the CoAxial Segment is a magmatic province distinct from other segments of the JdFR. Further, examination of basalt geochemical data from the different geologically identified terrains allow us to test several hypotheses put forth by Embley et al. [in press] regarding their origin. In addition, these data can be used to investigate what relationship, if any, these central JdFR terrains have to the CoAxial Segment. Firstly, it has been hypothesized that a highly reflective and unfractured area of seafloor south of -46° 1 8'N along Axial Seamount's NRZ delineates the extent of Axial Seamount related volcanism (boundary is shown as a solid black outline in Figure 2-1). Secondly, Embley et al. [in press] propose that the ridge north of -46° 1 8'N, termed the western fault block ridge (WFBR), is not tectonically or magmatically related to Axial Seamount as has been previously suggested [Sohn et al., 1998]. Direct comparison of the geochemistry of historical CoAxial flows to the geochemistry of lavas recently erupted within Axial Seamount caldera, lavas recently erupted along Axial Seamount's north and south rift zones (including the 1998 eruption detailed in Embley et al., 1998; Perfit et al..

PAGE 69

1998 and others) [Smith et al, 1997], and lavas recovered along the WFBR allow for the evaluation of these two hypotheses. 2.6.1 Axial Seamount Province Samples recovered from Axial Seamount NRZ and the WFBR are limited to those recovered by wax core and one video guided grab (Figure 2-1). NRZ samples (all within the region of highly reflective and unfractured sea floor) include 4 samples from a very fresh eruptive unit(s), as well as 9 other apparently older samples (Figure 2-1, Tables 2-2, 2-3, 2-4). With the exception of two samples recovered from small seamounts discussed in section 2.6.3. all of the samples recovered from this region bear strong geochemical affinity to samples recovered from within the caldera (Figures 2-3, 2-4, 2-9 and 2-10). Similarly, only one sample recovered from the SRZ [Perfit et al., 1998] shows any geochemical distinction from lavas recovered from within the caldera, and it too was recovered from a small seamount. Isotopically, all of the NRZ and SRZ samples analyzed to date fall within the field defined by samples from Axial Seamount caldera (Figure 2-4). The geochemical similarities between fresh lavas recovered from within Axial Seamount caldera and those recovered from the north and south rift zones extend beyond the Sr isotopic data. Lavas sampled from Axial Seamount' s rift zones are indistinguishable from caldera lavas with respect to the major elements (Figure 2-3), K20/Ti02 ratios (Figure 29), REE (Figure 2-10) and other trace elements. The ratios of elements with similar distribution coefficients are unaffected by crystallization processes. Accordingly, regional differences in primitive magma composition are well illustrated by plots showing the ratio of two highly incompatible elements against latitude (e.g. Figure 2-9).

PAGE 70

63 ^ 1998 Axial Smt. © Recent NRZ <> Axial N+S RZ ^ On-Axis Volcanoes + CoAxial Segment Q Rogue Volcano O WFBR 15X loH I 2-0 045.5 NRZ+SRZ Volcanoes Axial Seamount Proper Recent Floe and Flow Site lavas o -HEnd of NRZ 46 46.5 Degrees Latitude North Figure 2-9: Regional variation in EMP K^O/TiOi shown plotted against latitude. Symbols are as in the legend, and data fields are as labeled. All data are electron microprobe analyses of natural MORE glasses. Maximum possible variation in the ratio x/y due to analytical precision (2-sigma) governed by the equation [(x+2ax)/(y-2ay)]-[(x2ax)/(y+2ay)] equals 16% (shown for a K2O/TiO2=10). The field for Axial Seamount caldera, data points for its SRZ and the lava erupted in within the caldera and SRZ in early 1998 (Perfit et al., 1998) are all based on 73 unpublished EMP analyses (Perfit et al., in prep). The "Axial Proper" field encompasses data from all but one of the samples reported in Rhodes et al., 1990. Vertical dashed line marks NRZ boundary discussed in text.

PAGE 71

Figure 2-10: (Upper) Chondrite-normalized (Taylor, 1982) REE patterns for regional basaltic lavas. Open diamonds show data for recent lavas from Axial Seamount caldera, and black-filled circles the recent NRZ flow discussed in text. Other symbols are as in Figure 2-9. Gray field marks the range of data shown in Figure 2-8 for the 1993 lava flow. Regional differences in LREE depletion are observed in lavas, with La/Sm ranging from -0.9 in Axial Seamount province samples to -0.3 in samples from Rogue Volcano. Samples from the WFBR (see text) have closer geochemical affinity to CoAxial Segment lavas than to lavas from Axial Seamount Province. (Lower) Covariation between EMP K20/Ti02 and ICP La/Smn in regionally sampled lavas. Data from the Cobb-Eickelberg Seamount chain (Desonie and Duncan, 1990) are shown as open triangles. Strong positive correlation confirms that K20/Ti02 in sampled lavas preserves a mantle derived geochemical signal and it is not dominated by crustal processes.

PAGE 72

65

PAGE 73

66 Due to the limited number wax core samples analyzed for trace elements by either XRF or ICP techniques, the ratio K20/Ti02 (microprobe data) serves as the best available elemental parameter relatively unaffected by crystallization, to discriminate regional geochemical differences. Good regional correlation of K20/Ti02 with La/Smn, Zr/Y and ^^Sr/^^Sr (Figures 2-10, 2-1 1 and 2-12) supports the assumption that the geochemical signal seen in K^OfTiOj variations is related to differences in mantle-derived magmatic composition, and not to subsequent shallow level processes such as fractional crystallization and/or crustal assimilation. K20/Ti02 values for Axial Seamount Province range from -0.1 to 0.17, and samples recovered from the north and south rift zones (right and left pointing triangles respectively) are indistinguishable from samples recovered from Axial Seamount caldera. Chondrite normalized REE patterns are similarly indistinguishable between the rift zone lavas and those from Axial Seamount caldera (Figure 210a). REE patterns from four samples of the youngest north and south rift zone lavas are shown (filled circles) and completely overlap in shape and abundance with data from the three caldera lavas analyzed for REE (open diamonds). As with the NRZ lavas for which data is shown, data for caldera lavas plotted are from the youngest identified (visually) in the caldera and have MgO contents comparable to the rift zone lavas. Seven fresh looking samples were analyzed from the Axial Seamount Province, and have a total range in La/Smn of 0.88-0.92. Consideration of the three older rift zone lavas (not shown) analyzed for REE extends the range of La/Smn down to 0.8. All of these observations, combined with the distinct chemistry of lavas recovered outside of the zone defined by Embley et al. [in press] (discussed below), support the hypothesis that the highly

PAGE 74

67 40 30 _ Crystalline Whole C 20 h Rock Analysis 10 0 \ O o O :! A A +++ O + 1 A Axial Smt. Province 3 Zr/Y Figure 2-11: Regional covariation of MORE K20/Ti02 (EMP) with Zr/Y (XRF/ICP). Strong positive correlation suggests that the assertions made in Figure 2-10 hold true for trace element groups other than the REE. Symbols and fields are as in Figures 2-9 and 210. See text for further discussion.

PAGE 75

Figure 2-12: Correlation of the long lived radiogenic isotope Sr/ Sr with both KoO/TiOT (upper) and Sr concentration (lower). Symbols are as in the legend. Fields for the Cleft Segment (grey field) and Axial Seamount Province (stippled field) are based on our unpublished data. Data for Axial Seamount samples analyzed for both isotopic and elemental abundances are limited, and, as such, the field shown encompases the full range of values for all samples in our Axial Seamount database. The range is similar to other published data, with the exception of **^Sr/^^Sr values reported in Rhodes et al. (1990) and discussed in text. Calculated mixing curves (Langmuir et al., 1978) are shown (lower) for mixes between a CoAxial Segment lava (221-0551) and lavas from Axial Seamount caldera (dashed line) and Warwick Seamount (DH4-4-solid line). Symbols on mixing-curves mark 10% mixing increments.

PAGE 76

69 0.70260 0.70250 ^ 0.70240 00 0.70230 0.70220 I I rCleft Seg 0 O + + i -I 1 1 1 1 1 1 , 1 1 r Axial Smt. Province f T I T A A + CoAxial Segment + A C-E Smts. O Rogue Volcano Axial Province ' I 10 15 20 25 30 K2O/TiO2*100 Sr ppm

PAGE 77

70 reflective and unfractured region of seafloor north of Axial Seamount caldera is blanketed with lavas that emanated from a magma source similar to the source that feeds Axial Seamount caldera. 2.6.2 The Western Fault Block Ridge The topographic ridge extending north from NRZ (west of the CoAxial segment axial valley) becomes distinctly less acoustically reflective and more fissured north of approximately 46° 18' N latitude [Embley et al., in press]. This has been interpreted to reflect the greater age and sediment cover of the seafloor surface in this region, which apparently lacks recent volcanic overprinting of seafloor structures. Embley et al. [in press] conclude that the terrain between Axial caldera and 46°18'N is related to recent constructional volcanism associated with NRZ volcanic activity, but north of -46° 18' they propose the topographic ridge is unrelated to Axial Seamount magmatism and refer to it as the western fault block ridge (WFBR). Five lava samples recovered by wax core from the WFBR are, on average, more MgO-rich than Axial Seamount Province lavas. These WFBR lavas have MgO concentrations from 10.1-7.4 (ave=8.8). These samples (shown as open circles on Figure 2-9) have KaO/TiOo between 0.3 and 0.7 and fall outside of the field defined by Axial Seamount Province (K20/Ti02 = 0. 1-0. 17). The low KiO/TiO. values are more characteristic of the range defined by CoAxial Segment lavas (shown as crosses on Figure 2-9), for which 90% of 98 samples analyzed have a KjOATiOs < 0.9. Two of the WFBR lavas (94JdFRC22 and 95JdFRC35) have been analyzed for trace elements, and these data confirm their more incompatible element depleted nature relative to Axial Seamount Province lavas (Table 2-4). The two WFBR lavas have La/Sm„ values of 0.38

PAGE 78

and 0.48 respectively compared to recent Axial Seamount lavas that range from 0.880.92. Zr/Y values of less than 2.6 in these samples are much more characteristic of CoAxial Segment lavas which average 2.7 and have 90% of samples less than 3.0 (101 out of 1 1 1 analyzed). By comparison, Axial Seamount lavas have Zr/Y greater than 3 (Figure 2-11). While the sample base for the WFBR is small, and the two samples analyzed for trace elements do not uniformly exhibit the Sr elemental depletions characteristic of CoAxial lavas, these lavas nonetheless seem to bear a much stronger geochemical affinity to those from the CoAxial Segment than to those from Axial Seamount Province. WFBR lavas' elementally depleted nature relative to Axial Seamount lavas supports the hypotheses that the WFBR is a feature unrelated to Axial Seamount magmatism. Samples recovered from the ridge that bounds the eastern side of the CoAxial Segment over a similar latitude range as the WFBR, are limited to six basalt samples from one dredge haul (95DR2). Though the data are limited, this topographic feature also seems to have geochemical signatures similar to both the CoAxial Segment and WFBR. These samples have XRF measured whole rock glass ZrA' values between 2.4 and 2.8 (average = 2.6) compared with an average value of 2.7 for the CoAxial segment lavas. XRF K^O/TiO: values are also similar to CoAxial, ranging between 0.054 and 0.096 (average = 0.077) compared with the average CoAxial value of 0.065. 2.6.3 Small NRZ Volcanoes Two small seamounts within the NRZ terrain were sampled by rock (Figure 2-1, Tables 2-2 and 2-4, samples 94RC32 and 33) and both have very mafic compositions (greater than 9 wt.% MgO) and exhibit KjO/TiO. values well below those measured in

PAGE 79

72 lavas from Axial Seamount Province (Figure 2-9). A similar volcanic feamre sampled from the Axial Seamount SRZ (Figure 2-9) also has similar mafic and depleted characteristics. Incompatible element depletions in these features appear not to be limited to the major elements. Sample 94RC32 was analyzed for trace elements (Table 2-4), and has both La/Sm„ and Zr/Y that are significantly lower than in lavas from Axial Seamount Province and comparable or lower then WRBF and CoAxial lavas (Figures 210 and 2-11). Two opposing hypotheses arise from consideration of the above observations and the data in Figures 2-9, 2-10 and 2-11. Firstly, these two small volcanoes may represent topographically high eruptive centers associated with the WFBR, and have survived subsequent burial by more recent volcanism along Axial Seamount NRZ. Alternatively, these two volcanic cones might have distinct magmatic plumbing systems similar to the many other small to medium sized near-ridge volcanoes that exist throughout the region (Figure 2-1, dark gray fill). On the one hand, if the WFBR terrain is distinct from Axial Seamount Province as is suggested above and by Embley et al. [in press], then the presence of a volcanic feature geochemically similar to the NRZ volcanoes within the SRZ is difficult to explain via the first hypothesis. If, on the other hand, the alternate hypothesis is true and these small volcanoes have magmatic sources distinct from either Axial Seamount Province or the WFBR, then one might expect to see similar geochemical depletions in lavas from other near axis volcanoes in the region (Figure 2-1). Although there are numerous small near axis volcanoes in the CoAxial Segment region, particularly on the Pacific plate (Figure 2-1), only Rogue Volcano (on the JdF plate) has been sampled (Tables 2-2 and 23). Basalts analyzed from this volcano (6 for major elements, 2 for trace elements and Sr

PAGE 80

isotopes) are all from one dredge (Figure 2-1). Similar to the small volcanoes sampled from the NRZ, lavas recovered from Rogue Volcano tend to be relatively MgO-rich and incompatible element depleted having average values for MgO (wt.%), Zr/Y, La/Smn and K20/Ti02 equal to 9.1, 1.5, 0.33 and .058 respectively. The association of near-axis volcanoes with unusually mafic and depleted basalt compositions has been previously recognized on the JdFR [Leybourne and Van Wagoner, 1991; Finney, 1989] and other regions of the MOR system [e.g., Fomari et al., 1988; Allan et al.]. In particular, lavas sampled from the Heck and Heckle Seamount chains (located west of the Endeavor Offset -225 km north of the Rogue and NRZ volcanoes) have very similar geochemical characteristics to the lavas recovered from Rogue and the NRZ volcanoes. Similar to the samples discussed above, basaltic glasses recovered from the Heck and Heckle seamount chains [Leybourne and Van Wagoner, 1991] tend to be mafic, averaging greater than 8.4wt% MgO and reaching concentrations as high as 9.3 wt.%. Heck and Heckle lavas are also depleted in incompatible elements, having Zr/Y less than 2.4, K20/TiO2 less than 6 and La/Smn less than 0.42. The above elemental considerations support a hypothesis whereby the small NRZ seamounts result from magmatic activity unrelated to either Axial Seamount or the WFBR, possibly having origins similar to other near-axis volcanoes in the central JdFR region. 2.6.4 The CoAxial Segment CoAxial Segment lavas are perhaps most uniquely characterized by their depleted strontium abundances and isotopic ratios relative to other axial JdFR basalts (Figures 2-4 and 2-5). CoAxial segment lavas also tend to be quite incompatible element-depleted having Zr/Y, K20n'i02 and La/Sm ratios ranging from 2.4-3.2, 0.03-0.1, and 0.48-0.75

PAGE 81

and averaging 2.7, 0.07 and 0.6 respectively. Lavas from the 1980s Floe unit and the Source site (though not well represented in our sample suite) tend to be the least LREE depleted of any CoAxial lavas analyzed, consistently having the highest La/Sm,, (greater than 0.68). These samples also tend to have among the most radiogenic Sr isotopic signatures observed for the segment. With the exception of the 1980s Flow unit discussed in detail in section 5.2.2, samples from the flow site tend to be the most radiogenically and elementally depleted along the segment. This effect is exemplified by a trend of generally decreasing K20/Ti02 with increasing latitude along the segment (Figure 2-9). Although CoAxial lavas are elemental-depleted, their compositions do not display the extreme of elemental depletions evident in the "near-axis" volcanoes discussed above (Figures 2-9, 2-10 and 2-11). 2.7. Discussion 2.7.1 Recent CoAxial Segment Volcanism The volcanic eruption that occurred on the CoAxial Segment in 1993 at the Flow Site was a phenomenon unique in the study of mid-ocean ridges because it was an actively monitored crustal accretion event [Fox et al., 1995]. Real time seismic monitoring of the JdFR recorded the diking event that fed the 1993 eruption. The locations and migration of the epicenters of T-phase seismicity recorded during the event led to the initial conclusion that the dike intruded along strike and likely originated from the north rift zone of Axial Volcano [Dziak et al., 1995]. Although it was recognized that topographic effects could have caused a westward bias in the locations of epicenters, Axial Seamount could not be ruled out as a source for the eruption. Subsequent studies of post-emplacement seismicity on the CoAxial Segment concluded that the dike path

PAGE 82

was along the CoAxial spreading center axis, and suggested that T-phase epicenters associated with the 1993 seismic event were mislocated [Sohn et al., 1998]. The documented geochemical differences between Axial Seamount Province lavas and CoAxial Segment lavas in general, and the 1993 flow specifically, make it highly unlikely that Axial Seamount served as the source for the 1993 dike. The geochemical distinction of the two regions is consistently seen in the major element, trace element and isotopic characteristics of lavas from each region. A "CoAxial" origin for the 1993 eruption is further supported by geologic and hydrologic observations [Embley et al., in press], making a compelling argument against Axial Seamount as a source for the 1993 event. Two other eruptive units have been emplaced along the CoAxial Segment since 1981 [Embley et al., in press]. Differences in the major and trace element geochemistry between these two flows strongly suggest that lavas from these two units are neither comagmatic nor are they cogenetic. This leads to the conclusion that at least three separate eruptive events have occurred along this ridge segment during a span of twelve years or less. This is the first such area of the MOR system where such detailed constraints on eruptive frequency and distribution have been established, allowing for a more detailed investigation of the temporal and spatial aspects of crustal accretion at this medium rate spreading center than would otherwise be possible. Relative location of the three lava flows establishes the "neovolcanic zone" for this ridge segment to be greater than 1 km wide at the decadal time scale and over lengths of less than 25 km along strike. The eruption of the 1980s Floe unit along a preexisting constructional ridge [Embley et al., in press] and the in situ sampling of relatively fresh and geochemically

PAGE 83

76 distinct lava from a kipulca in the 1980s Floe unit (sample 2948-1 OR) is intriguing. The presence of relatively fresh, non comagmatic lavas comprising the pre-existing ridge suggests that at least two eruptions have occurred at this location in the recent past (less than lOO's of years), some 500m west of the interpreted ridge-axis and locus of current geologic and hydrothermal activity. Accordingly, although CoAxial eruptive activity might not be as well focussed as in some other regions of the Juan de Fuca Ridge and northern East Pacific Rise studied in comparable detail [Embley and Chadwick, 1994; Fornari et el., 1998], there is likely some control of preexisting structure on the location of dike emplacement along this segment. This is consistent with the observation that the 1 993 dike was emplaced along the trend of a preexisting fissure system that was apparently reactivated during the intrusive event. Perhaps as intriguing as the temporal/spatial variation in lava emplacement is the variation observed in lava chemistry. None of the flows which have been emplaced within the CoAxial Segment axis over a twelve year period appear to be cogenetic and each has a genetic origin uniquely distinguished from the other by elemental concentrations and ratios and/or Sr isotopic ratios. In addition, it has been noted that samples recovered from the CoAxial Segment axis show a greater dispersion in elemental concentrations at a given MgO value than do other Segments of the southern JdFR. These observations suggest that there is not a well-mixed, long-lived magma chamber serving as the source for CoAxial Segment volcanism. Inter-flow geochemical variability is more suggestive that the magmatic plumbing underneath the CoAxial Segment is poorly developed, with individual melt bodies being relatively small and short lived. In fact, differences in mantle derived geochemical signals observed between the two recent

PAGE 84

77 units at the Flow site (Sr and Nd isotopes and the ratios of elements highly incompatible with mantle mineral assemblages) show that mantle geochemical heterogeneity can be preserved by seafloor crustal accretion here on temporal and spatial scales of a decade and 500m. The hypothesis that CoAxial Segment volcanic activity (at least that part north of the poorly geochemically characterized Source site) is not related to a large, long-lived, controlling magma chamber is consistent with the observed geology and geophysics. Seismic data for the upper three km of the crust show no evidence of a low velocity zone associated with an axial magma chamber reflector [Sohn et al., 1997], and seafloor compliance measurements conducted on the segment failed to detect any magma bodies in the crust that comprises the CoAxial Segment axial valley [Crawford, 1994]. Additionally, Sohn et al. [1997] have interpreted the crustal seismic structure of the northern portion of the CoAxial Segment in terms of a dominantly amagmatic extensional terrain. The CoAxial Segment's morphology and structure is not like that of some magmatically robust segments such as the Cleft segment to its south [Embley et el., 1991] or the N. East Pacific Rise (EPR) [Haymon et al., 1993]. Along much of the Cleft Segment and the 9°-10°N segment of the EPR, axial volcanism appears to be dominantly focussed within a relatively narrow region. These areas, presumably underlain by well developed magma chambers [Smith et al., 1994; Fornari et al., 1998], are associated with phenomenon such as high temperature hydrothermal venting, formation of lava lakes and large regions of volcanic collapse [Embley et al., 1991; Haymon et al., 1993; Fornari et al., 1998; Kent et al., 1993]. None of these features have been observed in the northern

PAGE 85

78 portion (Floe and Flow Sites) of the CoAxial Segment. CoAxial Segment volcanism occurs over a zone greater than 1 km wide, and significant portions of the CoAxial terrain north of the Source Site are highly fissured giving the impression that amagmatic extension has played an important role in the development of axial structure, despite the evidence for recent volcanic activity. The spatial patterns of recent volcanism and abundant fissures within the neovolcanic zone of the northern CoAxial segment are consistent with seismic data that suggest no shallow magma chamber underlies this portion of the segment. These observations are consistent with a southerly source for recent volcanism, and support the hypothesized lateral propagation of the 1993 dike, thought to have originated south of the Floe Site. 2.7.2 Intra-flow Geochemical Variability. Although no direct evidence for crustal magma bodies exists, it is clear that the recently erupted CoAxial lava flows have undergone substantial differentiation prior to eruption. One situation where chemical fractionation could have taken place is during the lateral emplacement of the dike through the crust. In the case of the 1993 eruption, the dike is thought to have migrated some 40 km along axis [Dziak et al., 1995] over a period of ~2 days. While evidence of intra-flow fractional crystallization in the 1980s Floe unit along the strike of its emplacement (Figure 2-7) is intriguing, a closer examination of major element data shows this effect to be insufficient to generate the observed fractionated compositions of the 1993 lava flow. Basaltic glasses show a nearly 1 wt.% variation in MgO over the 7 km long axis of the flow. This change in MgO corresponds to -20% fractional crystallization of the most mafic portion of that flow based on LLD

PAGE 86

79 calculations. On the surface then, it would seem that a fractional crystallization rate of -3% per km would be sufficient, over a lateral emplacement distance of ~40km, to account for the approximately 75% fractional crystallization required to differentiate a moderately primitive MORB magma (~9.5wt% MgO) to the composition of the 1993 Flow (~6.7wt% MgO). Two observations argue that chemical differentiation of the 1993 flow could not have entirely occurred during dike intrusion and eruption. Firstly, comparison of the intra-flow whole rock glass data to the EMP data for the 1980s Floe unit shows that the liquid differentiation process was no where near as efficient (by -50%) as pure fractional crystallization predicts. In addition, subaerial analogues tot he 1993 diking event do not exhibit this degree of magmatic differentiation along the path of the dike injection. This is not surprising, as it is difficult to envision efficient separation of crystals from liquid in the dynamic environment of a dike emplacement that occurs over the course of only a few days. Nonetheless, some bulk chemical differentiation does appear to have occurred along the 7 km strike of the 1980s Floe Site flow. Secondly, the petrography of samples from the 1993 flow suggests that the magma likely had some residence in a crustal magma body prior to dike intrusion. Glassy rims of the 1993 lava samples tend to be quite crystalline (visual estimates of up to 10-20 modal percent). Much of the crystal assemblage consists of small, quenched microphenocrysts of plagioclase and clinopyroxene consistent with a phase of rapid cooling and crystallization prior to eruption. There are, however, a second population of larger phenocrysts (up to 2 mm) of olivine, clinopyroxene and plagioclase, some of which show signs of resorption and chemical disequilibrium. It is unlikely that this

PAGE 87

80 second population of larger grains could have formed during the time frame of the intrusion event, and it is more likely that they crystallized from a cooling crustal magma body. This conclusion in no way, however, detracts from the importance of the intraflow chemical variation that does exist within the 1980s Floe unit, but does infer the presence of a CoAxial crustal magma body that has yet to be imaged. Recently a positive correlation between intra-flow chemical variability and eruption volume has been suggested [Rubin et al., 1998; Perfit and Chadwick, 1998]. Additionally, others have suggested an inverse correlation between spreading rate and eruptive flow volume [Sinton, 1997; Perfit and Chadwick, 1998]. The strong correlation of glass geochemistry along the strike of emplacement for the 1980s Floe unit, and to a lesser extent the 1993 Flow unit, may imply a causal mechanism relating these two observations. Intra-flow geochemical variation observed in the three recent CoAxial lava flows are all consistent with small amounts of magmatic differentiation along low pressure LLDs. This is most easily explained via small degrees of cooling and crystallization occurring during lava emplacement. This hypothesis is supported by the petrography of the 1993 CoAxial flow which has a large proportion of microlites and microphenocrysts in the outer most glassy selvage of the lava flow, a characteristic that is likely related to a rapid phase of cooling prior to eruption. The elongate lineated nature of the 1980s Floe unit and 1993 Flow unit, their systematic intra-flow geochemical variation along ridge strike, and the observation of lateral dike propagation during the 1993 event all suggest that the 1980s Floe unit was emplaced via a similar south to north propagating dike. Does the process of lateral propagation of intruding (and eventually erupting) dikes in the shallow ocean crust aid in the generation of intra-flow geochemical

PAGE 88

81 heterogeneity? These data would support that conclusion. Is a dike more likely to propagate long distances laterally in a slower spreading environment? Intuitively it would seem that an increasingly tensile stress regime would promote such propagation, but factors such as ridge-axis structure, magma density, eruptive pressure and topographic considerations are all factors to be considered. It is intriguing that several investigations of JdFR volcanism have led to conclude that lateral dike propagation served as the mechanism of lava emplacement [Embley et al., 1991 and in press]. Conversely, investigations of the northern East Pacific Rise have concluded that recent eruptions were emplaced through mostly vertical dike intrusion [Wright et al, 1995]. 2.7.3 Regional Petrochemical Characteristics This study recognizes four petrochemically distinct magmatic provinces in the Axial Seamount-CoAxial Segment region. Axial Seamount Province and its south and north rift zones comprise one. Another is comprised of the western fault block ridge. The third is the Coaxial Segment, and the last is comprised of several small axial and near axis volcanoes. 2.7.3.1 Axial Seamount Samples recovered from Axial Seamount north and south rift zones are generally geochemically indistinguishable from lavas recovered from within the caldera. Data presented here support the boundaries of NRZ volcanism proposed in Embley et al. [in press]. Axial Seamount Province lavas are the least geochemically depleted of the four provinces discussed. They are enriched in potassium and sodium relative to other southern and central JdFR lavas (Figure 2-3), and compared to the other three aforementioned magmatic provinces, Axial Seamount lavas are less depleted in the light

PAGE 89

82 REE (La/Smn of 0.88-0.92), Sr (Figure 2-5), ^^Sr/^^Sr, and elemental ratios such as Zr/Y and KiO/TiOi. Each of these observations is indicative of geochemical enrichment in mantle-incompatible elements in Axial Seamount lavas relative to CoAxial or WFBR lavas. Lavas from Axial Seamount are not, however, obviously enriched in isotopic and incompatible elements relative to all JdFR lavas south of the Cobb offset [Rhodes et al., 1990; Smith et al., 1994]. Nonetheless, they do generally have higher values of Zr/Y, La/Sm and Sr/Zr, and, as such, the Axial data only partly overlap the geochemically enriched end of the JdFR data-field for each of these parameters [Rhodes et al., 1990]. Previous authors have modeled enrichments in Sr, Na and Ca observed in Axial Seamount lavas to be the result of a greater depth of initial melting of a source similar to that of the ridge-axis MORB [Rhodes et al., 1990]. The more depleted nature of the CoAxial Segment and WFBR provides further constraints on the degree of regional mantle geochemical heterogeneity and the depleted mantle end-member. 2.7.3.2 The Coaxial Segment While a detailed petrogenetic analysis is beyond the intended scope of this paper, and will be dealt with elsewhere [Smith et al., in prep], a few general observations are relevant to the discussion presented here. Qualitatively, depletions in AI2O3 and Na20 in the CoAxial MORB suite relative to the Axial Seamount lavas, coupled with enrichments in FeO, could be explained via increased degree of melting of an "Axial Seamount" melt source [Klein and Langmuir, 1987 and 1989, Langmuir et al., 1992]. This hypothesis is not, however, consistent with the depletion of CaO in CoAxial MORB relative to Axial Seamount lavas. In fact, depletion in CaO of CoAxial lavas relative to Axial Seamount lavas is supportive of smaller degrees of melting, unless melting proceeds well past the

PAGE 90

83 elimination of CPX as a residual phase in the source mantle [Klein and Langmuir, 1987]. Additionally, the inverse correlation between Na20 and FeO contents between the two regions corresponds to the trend associated with the "global" variability vector of Klein and Langmuir [1987, 1989], while the inverse correlation between Na20 and SiO: would be more easily explained via the "local trend vector" and sampling of instantaneous melts from different depths in a melting column. The inability to consistently account for differences in the suite of major element oxides observed between Axial Seamount and CoAxial Segment lavas via differences in the extent and depth of melting (i.e. via either the local or global trends of Klein and Langmuir, 1989) argues against the hypothesis that differences in major element characteristics between the MORB from the two regions are due to different conditions of melting of an "Axial Seamount" source in the generation CoAxial lavas. As in the above arguments, comparison of CoAxial MORB data to the Cleft Segment MORB data also suggests that differences in major element lava chemistry between the two regions cannot be simply explained by different melting conditions of a similar and homogenous source. Lower relative Na20 abundances in Cleft segment lavas would require a higher extent of melting relative to CoAxial lavas, but higher AI2O3 contents suggest a lower extent of melting. Clearly these two observations are incompatible with the above hypothesis. In summary, regional differences in major element chemistry of CoAxial lavas relative to other southern Juan de Fuca lavas cannot be generated by melting of a homogeneous mantle, but must instead be related to mantle chemical heterogeneity and or differences in crystallization and differentiation processes.

PAGE 91

84 Additionally, trace element and isotopic constraints prohibit the generation of CoAxial lavas through melting of a homogeneous "Axial Seamount" source mantle. An alternate hypothesis to that discussed above (differences in primitive melt composition relating to differences in the extent and depth of melting of similar mantle compositions) is that the source mantle sampled by Coaxial lavas is different from the other two regions either by large scale mantle heterogeneity or finer scale heterogeneity sampled differentially through variable conditions of melting. CoAxial Segment lavas exhibit characteristics of long term incompatible element depletion (radiogenically depleted ^^Sr/^^Sr and enriched '"^Nd/^^Nd). Regional correlation between incompatible element ratios (Figures 2-9, 2-10 and 2-11), and between long-lived isotopes of strontium and incompatible element ratios (Figure 2-12), indicate that the range of mantle source characteristics observed can be accounted by mixing of very depleted and slightly enriched components. Figures 2-10, 2-1 1 and 2-12 show regional correlations between samples from the CoAxial Segment, Axial Seamount Province, Rogue and the NRZ volcanoes, the WFBR and the Cobb-Eickelberg Seamount Chain, of which Axial Seamount is the youngest member [Desonie and Duncan, 1990]. All three diagrams show generally hyperbolic trends characteristic of mixing curves on ratio-ratio diagrams [Langmuir et al., 1978]. Figure 2-12 shows mixing of ^^Sv/^^Sr with both KsOATiOs (upper) and Sr (lower). Calculated trends are shown for mixes between a depleted CoAxial lava and lavas from Axial Seamount caldera (dashed curve) and Warwick Seamount, a member of the CobbEikelberg chain (solid line). Trend symbols mark 10% mixing-increments. Examination of Figure 2-12 shows the CoAxial data trend to be elongated in the direction of the

PAGE 92

85 mixing curves, with the most enriched CoAxial samples showing up to a 40% mix of "Axial Seamount melt component" or a 20% mix of a "Warwick Seamount melt component" with the depleted CoAxial end-member. Axial Seamount lavas themselves fall in the range of 40% mixing of the enriched "Warwick Seamount" end member with the depleted CoAxial end-member. While it is not the intention or within the scope of this paper to model in detail exact nature of heterogeneities occurring in the upper mantle, some interesting observations and hypothesis arise from the examination of Figures 2-10,2-11 and 2-12. 2.7.4 Regional Mantle Heterogeneity and the Origins of CoAxial Segment Geochemical Depletion. Regional correlation of MORE Sr isotopic values with incompatible trace element abundances implies some antiquity to mantle geochemical heterogeneity in this region. While some aspect of this heterogeneity must be related to long-term depletion events, significant variability of incompatible elements/ratios about the mixing curves at given values of ^'Sr/^^Sr implies there has also been more recent enrichments/depletions (Figure 2-12). This regional mantle heterogeneity is likely the result of recent (relative to the half-life of *^Rb) upper mantle melting, but the cause(s) of the more ancient heterogeneity is more difficult to assess [Smith et al., in prep]. Distinct Isotopic depletions observed in CoAxial lavas relative to all other Juan de Fuca MORE (Figure 24) imply a depletion of ancient mantle geochemical enrichments that is unique on the Juan de Fuca Ridge. Two hypotheses are recognized to explain these depletions. The first hypothesis is that melting around the outer portions of the Cobb melting anomaly has left "CoAxial Segment" upper mantle depleted in enriched source component. Evidence suggests that much of the northern CoAxial Segment appears

PAGE 93

86 relatively amagmatic, and that the sources of recent eruptions probably have an origin south of the Floe Site. If this were in fact the case, and CoAxial lavas were being depleted by previous source melting due to the Cobb melt anomaly, then one might expect source depletions to show a radial pattern about the melt anomaly. Little evidence exists for comparably depleted lavas south of the Cobb anomaly, although sampling in this region is currently insufficient to truly test this hypothesis. An alternate hypothesis is that recent rapid north westward migration of the ridgeaxis in this area (> 35km in the past 0.5 Ma and up to 100 km in the past 7 Ma) as the ridge has approached the Cobb melting anomaly [Karsten and Delaney, 1989] has placed the current ridge-axis above a region of upper mantle already depleted through the process of MORE magma genesis. This hypothesis of rapid ridge-axis migration was called upon by Karsten and Delaney [1989] to explain the asymmetric distribution of near-axis seamounts occurring on the Pacific plate (Figure 2-1). If both the asymmetric seamount distribution and CoAxial Segment source depletions are due to the rapid northwestward migration of the ridge in this area, then one might expect two conditions to be true. Firstly, one might expect similar depletions in Vance segment just south of the melt anomaly, providing it was equally affected by ridge migration. Although existing data is sparse, this does not appear to be true. Secondly, one might expect that seamounts on the Pacific plate outside of the current ridge-axis zone of melting to have, on average, geochemical signatures more enriched than either the ridge-axis MORE or the near-axis volcanoes to their south east. Whether the depleted character of CoAxial lavas is the result of effects associated with the Cobb melting anomaly, ridge-axis migration, or some other phenomenon, it is

PAGE 94

87 likely that CoAxial lavas were produce by a mantle previously "conditioned" by at least small amounts of previous melt extraction. It is hypothesized that this previous melting preferentially sampled whatever enriched mantle component existed in the JdFR MORB source giving CoAxial lavas a distinct geochemical character relative to other JdFR lavas. Broad scale regional correlation of source characteristics and finer scale variation within individual magmatic provinces suggests that the spatial scale of mantle heterogeneity in this region is smaller than the individual magmatic provinces discussed. 2.8. Conclusions An integrated approach to the study of MORs involving the use of geophysical, geologic and geochemical mapping methods is necessary to address, in detail, questions of spatial and temporal variability in MOR crustal accretion. At least three different eruptive events have occurred on the CoAxial Segment over a span of less than 12 years. Data presented above leads to several important conclusions regarding the process of crustal accretion at the CoAxial Segment and the spatial association of distinct melting regimes on the central JdFR in general. These conclusions include the following: (1) Spatial relations of these units establish a neovolcanic zone greater than 1 km wide over a decadal time scale. (2) Recent CoAxial Lavas are geochemically distinct from recent Axial Seamount lavas and are believed to have originated from within the CoAxial Segment, not from Axial Seamount as has been previously suggested. (3) Inter-flow chemical variability in CoAxial lavas suggests that the CoAxial Segment has a poorly developed magmatic plumbing system, in accord with the lack of any geophysically identified axial magma chamber within the upper 3 km of the crust.

PAGE 95

88 (4) Intra-flow geochemical variability of up to 1 wt.% MgO is observed in CoAxial lavas, and is best explained by small amounts of crystallization and differentiation along low pressure liquid lines of decent. (5) Four distinct magmatic provinces have been identified in the Axial SeamountCoAxial Segment region. The range of mantle source characteristics is consistent with mixing trends established using "source" characteristics of variably depleted/enriched lavas recovered from within the region. (6) Correlation between long lived Sr isotopes and incompatible element ratios imply that some aspect of local mantle heterogeneity has antiquity, but scatter of incompatible element ratios at a given isotopic value further suggest that a more recent component of elemental enrichment/depletion contributes to the geochemical heterogeneity of the upper mantle in this region.

PAGE 96

CHAPTER 3 SUBMARINE INVESTIGATIONS OF A THIRD-ORDER OSC AT 9° 37' N: ESTABLISHING A CAUSE AND EFFECT RELATIONSHIP BETWEEN OSC PROROGATION AND MAGMATIC ACTIVITY 3.1. Introduction The origins, development, and causal mechanisms of disruptions in the continuity and linearity of the East Pacific Rise (EPR) ridge-axis have been the subject of much investigation, discussion, and speculation in recent years [e.g. Macdonald and Fox, 1983 and 1988; Macdonald et al., 1984, 1986, 1987, 1988; Lonsdale, 1983 and 1986; Langmuir et al., 1986; Toomey et al., 1990; Perram and Macdonald, 1990; Haymon et al., 1991; Carbotte and Macdonald, 1992; Kent et al, 1993a, 1993b, Carbotte et al., in press, etc.]. Four different "orders" of discontinuity have been recognized on the basis of their spatial dimensions [Macdonald et al., 1988b], subdividing the northern EPR into discrete segments of varying length and temporal continuity [Macdonald and Fox, 1983; Macdonald et al., 1984; Langmuir et al., 1986; Perram and Macdonald, 1990; Toomey et al., 1990; Haymon et al., 1991; Carbotte and Macdonald 1992; Macdonald et al., 1992; etc.]. Several criteria have been used to define first through fourth-order ridge segments and associated first through fourth-order discontinuities that bound and define them [Macdonald et al., 1988b]. First-order spreading segments are tectonically defined and bounded by large transform offsets or propagating rifts whose offsets are such that plate boundaries behave rigidly, juxtaposing crust of greater than approximately 0.5-1 Myr age 89

PAGE 97

90 difference across the offset boundary. Offsets are typically greater than 50 km and partition the ridge-axis at intervals of 100-1000 km. Second-order segments have a length scale of 50-300 km and are bounded by smaller, non-rigid discontinuities such as large overlapping spreading centers (OSC) with an offset greater than 3-5 km or small offset (less than 20 km) non-rigid transform faults. Third-order segmentation divides the ridge-axis into characteristic lengths of 30-100 km, and the segments are bounded by smaller (0.5 to 3-5 km) OSCs characterized by small (lO's of meters) increases in axial depths. Lastly, fourth-order segments have typical length scales of 10-50 km. The fourth-order discontinuities are not typically associated with axial depth anomalies, and the segment boundaries are defined by small non-overlapping offsets of the ridge-axis and/or small (1-5°) bends in the ridge-axis, referred to as deviations in axial linearity (DEVALS) [Langmuir et al., 1986]. Although first-order segments and discontinuities are thought to be persistent on the order of millions of years, smaller ridge offsets may not be as temporally stable [Macdonald et al., 1988b]. The tendency of smaller ridgeaxis discontinuities to migrate and evolve has led to much speculation regarding their causative mechanism(s). While in some instances far-field plate stresses and/or changes in the direction of spreading have been identified as the driving mechanism for migration of second and third-order discontinuities [Hey et al., 1980, 1982, 1988; Carbotte and Macdonald, 1992], magmatic activity (or lack there oO has also been postulated as the active mechanism for the formation and evolution of similar features [Lonsdale, 1983; Sempere et al., 1984; Macdonald et al., 1984, Langmuir, 1986; Mutter et al., 1988; Toomey et al., 1990 and 1994; Smith et al., 1998; etc.]. The association between magmatic activity and fine-scale

PAGE 98

91 ridge segmentation has been largely based on correlation between the location of discontinuities and factors such as axial depth, continuity and depth of the seismically imaged axial magma chamber (AMC), magnetic, seismic and structural characteristics of the crust, and the geochemistry of mid-ocean ridge basalt (MORB) dredged from either side of offsets. While many such discontinuities in the eastern Pacific spreading axis have been imaged bathymetrically [e.g. Macdonald et al., 1984], seismically [Toomey et al., 1990 and 1994, Kent et al., 1993a, Kent et al, 1993b; Carbotte et al, in press], and magnetically [Sempere et al., 1984], relatively few have been optically imaged or petrologically sampled [Langmuir et al., 1986; Sinton et al., 1991; Batiza and Niu, 1992]. Those that have been sampled for basalt geochemistry have been only sparsely sampled by dredge, without having the sample density or positional accuracy to reconstruct in detail the geometry of geochemical variation across the ridge-axis discontinuity. Here the first detailed report of in situ observational and geochemical data recovered by submersible at a small third-order discontinuity is presented. Submersible data is supplemented by dense wax core sampling of basalt glass from the crestal region surrounding the OSC. 3.2. Previous Studies and Regional Geology 3.2.1 Previous Studies The study area lies between -9° 35' and 9° 37' N latitude, within a second-order spreading segment of the fast spreading EPR bounded to the south at -9° 03' N by an OSC and to the north at -10° 10' N by the intersection of the ridge-axis with the Clipperton transform fault (Figure 3-1) [Haymon et al., 1991]. This second-order segment spreads at approximately 1 Icm/yr full-rate [Klitgord and Mammerickx, 1982],

PAGE 99

Figure 3-1: Regional bathymetric map of the second-order EPR segment between -9° 03' N and 10°N latitude. Regional map-inset and scale are shown at left. Entire figure is after Haymon et al. [1991]. Contour interval for rise-crest bathymetry is 20 m and columns to right to show fourth-order segment boundaries and geologic, hydrothermal and biologic observations based on ARGO deep tow data gathered in 1989.

PAGE 100

93 104»30'W 104"'20' C&po*non Tr«ntlorm -10' Adventure Oive Area ~ IIO'W I 9*21.0' 9*1 7.0" ** From Haymon et ai.. 1991 Axial Summit Caidera (ASC) Dive Areas 1991 Eruption Area 0 5 10 kms

PAGE 101

and has been subdivided into 10 fourth-order segments based on the occurrence of numerous small ridge-axis discontinuities identified through bathymetric, morphologic, and structural observations [Haymon et al., 1991]. No third-order axial discontinuities have been recognized along this second-order segment. The EPR ridge-axis between 9° 09' N and 9° 54' N has been optically and acoustically imaged using the ARGO I near-bottom imaging system, with -90% coverage over an 800 m wide swath centered on the ridge-axis [Wright et al., 1995; Haymon et al., 1991]. Subsequently, over 300 Alvin dives have been conducted in the region between 9° 17' N and 10° N, though greater than 80% investigated the area of 9° 46-5 1 ' N, which experienced a volcanic eruption in 1991 [Haymon et al., 1993; Gregg et al., 1996]. Since 1991, hydrothermal, geologic and biologic activity have been monitored [Fomari et al., 1998a; Von Damm, in press], and study areas have been investigated and sampled annually by submersible. Between the latitudes of 9° 17' and 10° 05' N nearly 850 MORB glass samples have been recovered by submersible and wax core, and geochemically analyzed (Figures 3-2 and 3-3) [Perfit et al., 1994 and in prep]. In addition, this segment has been well characterized through a variety of geophysical techniques including seismic reflection, refraction, and tomography [Detrick et al., 1987; Kent et al., 1993a and 1993b; Harding et al., 1993; Christeson et al., 1996 and 1997; Toomey et al., 1990 and 1994; Dunn et al., 1997; Mutter et al, 1988], seafloor compliance measurements [Crawford et al., 1999], magnetic surveys [Sempere et al., 1984, Shouten et al., 1999], and near bottom gravity surveys [Cochran et al., 1999]. Clearly, this second-order segment is among the most thoroughly studied mid-ocean ridge (MOR) environments. This study addresses both the overall character of this

PAGE 102

Figure 3-2: Map showing regional distribution of Alvin and wax core recovered MORB. Submersible recovered samples are shown as gray filled diamonds and wax core samples as open circles. Black outline denotes the boundaries of the axial summit collapse trough (ASCT). Note the map is not a true projection, and has been adjusted to fit the page. For reference, 0.01 ° latitude or longitude is approximately equal to 1 km at this latitude.

PAGE 104

?^ £ =H P S o c 2 C ns C ~ Cu o c c I — ^ 1^ lu y < o ^ o O _ ^ c ^ 3 m o U ^

PAGE 105

98

PAGE 106

99 second-order segment, but particularly the region between -9° 35 '-37' N. A comprehensive review of existing data is well beyond the scope of this paper, and, consequently, only a brief summary of some previous findings relevant to discussions here is given below. 3.2.2 General Second-Order Scale Observations This second-order MOR segment has a well-developed axial summit collapse trough (ASCT) [Fomari et al., 1998b] over 80% of the axis between 9° 24' and 9° 54' N [Haymon et al., 1991]. Previous investigations by submersible and deep tow imaging have shown hydrothermal venting to be limited to the axial zone, and to within the ASCT when present. ASCTs are generally interpreted to form as a primary volcanic feature, but they can subsequently be modified by tectonic processes [Fomari et al., 1998b]. ASCT positions often strongly correlates with the presence of an axial magma chamber (AMC) detectable in seismic reflection surveys [Haymon et al., 1991; Carbotte et al., in press]. Ten fourth-order segments have been recognized along this portion of ridge [Haymon et al., 1991], with boundaries established by DEVALs identified at 9° 17', 9° 51.5', 9° 28', and 9° 35' N [Langmuir et al., 1986; Toomey et al., 1990; Haymon et al., 1991] or changes in the character of the ASCT. In general the youngest axial volcanic terrain and lava flows are found north of 9° 37' N, the shallowest portion of the segment. An exception to this are young lavas that crop out near the region of the 9° 17' N DEVAL [Haymon et al., 1991: Wright et al., 1995]. Relative ages of lavas are based on data recovered within the axial zone during the 1991 ARGO survey. These data show that the youngest lavas (termed Age 1 and estimated to be on the order of less than -50 years old) only crop out over three limited

PAGE 107

100 regions, with the area from 9° 14-19' being the only age 1 lavas identified south of 9° 37' N [Wright et al., 1995]. The 9° 14-19' area is thought to have experienced an eruption sometime between 1987 and 1989 based on geologic and hydrothermal data [Fornari et al., 1998b]. North of 9° 37' N, age 1 lavas have been noted to crop out over -90% of the length of the axis, occurring in two zones located from 9° 37.1 '-42.6' N and 9° 44.3'-52' N [Wright et al., 1995]. Apparent lava ages increase southward and are generally oldest in the region between 9° 35' and 20' N. Lavas within the ASCT along this segment of ridge all have relative ages greater than Age 1 and are mostly Age 1.5-Age 2 relative to the classification scheme of Haymon et al. [1991], which is largely based on relative degrees of sediment cover and vitreous luster of lavas [Wright et al., 1995]. Although one must view the absolute ages assigned to this classification scheme with caution, lavas of age "1.5" are considered to be on the order of less than 500 years old. By comparison age "1" lavas are considered to be less than 50 years old, and age "2" lavas are assigned an age range of 1000-5000 years old. Rates of sediment accumulation and development of surficial coatings on basalt glass (thereby reducing vitreous luster) are poorly calibrated, and this type of lava dating is best suited to establishing the relative age relationships of lavas from a given region. Nonetheless, the above time scales are in general accordance with the time scales of ASCT development proposed by Fornari et al. [1998b], which were calculated based on spreading rate and ASCT widths. Additionally, they are in general accord with isotopic models of crustal residence ages for lavas from the 9°-10° N segment (Sims et al., 1997a)

PAGE 108

101 Detailed examination of crustal fissuring also supports a change in the age of the ridge-axis at 9° 37' N. The 1989 ARGO survey not only revealed a general southward aging of axial lavas south of -9° 52' N, but marked 9° 35' N as a boundary of distinct change in fissure density and average fissure width. South of 9° 35' N, fissure density increases and average fissure width decreases, both attributes associated with a predominance of tectonic activity rather than magmatic activity [Wright et al., 1995]. The southward increase in fissure density continues to the recently volcanically active region around 9° 18', where it drops again sharply. These observations are also consistent with the inferred age of the ASCT, based upon its morphological characteristics and a recently posed model for ASCT evolution [Fomari et al., 1998b]. The model of Fomari et al. [1998b] describes an evolutionary development of ASCT morphology over a time frame of approximately 5000-10,000 years whereby widening of the ASC and straightening of its bounding walls are associated with aging of the system. The ASCT is narrowest north of 9° 37' N, being only 40-80 m wide in the region from -9° 40'-5r N. This area is also coincident with the shallowest portion of this second-order segment (-2500 m depth), and it was the site of an eruption in 1991. The eruption emanated from an 8.5 km long eruptive fissure located between 9° 46' and 51 ' N [Haymon et al., 1993; Rubin et al., 1994; Gregg et al., 1996]. The ASCT north of 9° 37' N are highly scalloped in plan view, and this region is considered to be in "stage 2" of the Fomari et al. (1998b) developmental cycle; perhaps 50-100 years removed from initiation of collapse trough development. The youthfulness of the crestal plateau volcanic terrain north of 9° 37' N, relative to the region further south, is supported by near bottom gravity surveys at [Cochran et al., 1999]. Cochran et al. (1999) have

PAGE 109

102 proposed that because of tectonic activity and off-axis volcanism, crustal density increases with age due to collapse and infilling of voids within the extrusive layer of the crust. The surveys at 9° 50' N and 9° 30' N determined the average crustal density near 9° 50' N to be 280 kg/xrr' less than that at 9° 30' N. Average crustal densities measured at 9° 30' N are typical of those determined elsewhere on the EPR and Juan de Fuca Ridge [Luyendyke, 1984; Holmes and Johnson, 1993; Stevenson 1994; Cochran et al., 1999], confirming the anomalous nature of the crustal densities around 9° 50' N. In contrast to the region north of 9° 37' N, the area from 9° 30'-35' N contains the widest, straightest walled ASCT observed along the segment. In places, ASCT widths exceed 250 m, and this region of the ASCT has been interpreted to be in stage 4 of the conceptual ASCT developmental model. Stage 4 of ASCT development is associated with relative magmatic quiescence and widening of the ASCT due to tectonic extension [Fornari et al., 1998b]. The time scale for stage 4 ASCT development is considered to be on the order of thousands of years (but less than 10,000 years) [Fornari et al., 1998b]. ASCT morphology in this area is consistent with the observed large number of narrow, tectonically generated fissures, large cumulate fissure width and older appearance of ASCT flooring lavas [Wright et al., 1995]. 3.2.3 Previous Geologic Observations of the EPR Axis Between 9° 34.9' N and 9° 37.1' N The portion of the EPR ridge-axis between 9° 34.9' N and 9° 37. 1' N has been characterized as a fourth-order segment (Segment "D" of Haymon et al., 1991) bounded to the north by an approximately 400 m right-stepping offset in the ASCT and to the south by a slight bend in the bathymetric contours of the ridge and smaller right stepping offset of the ASCT (Figure 3-U [Haymon et al., 1991]. Closer examination of the ASCT

PAGE 110

103 configuration as imaged in 1989 using the ARGO system shows there to be an overlapping relationship between the ASCTs south (western) and north (eastern) of the northern fourth-order boundary at 9° 37. 1 ' N (Figure 3-3c). ARGO data show the eastern ASCT to be offset by approximately 0.4 km in a right-lateral sense relative to the western ASCT, overlapping it by nearly 0.75 km. Though geometrically this configuration does not correspond to the 3:1 overlap to offset ratio typical of OSCs [Macdonald et al., 1984], this feature has been referred to as a "mini OSC" [Wright et al., 1995]. This fourth-order segment (Segment "D") was identified as hydrothermal gap by Haymon et al. (1991), lacking signs of hydrothermal venting or vent biota in the 1989 ARGO survey (Figure 3\). A more detailed analysis of the 1989 survey [Wright et al., 1995] shows there to be relatively steep gradients within Segment D in profiles of fissure width, density, and lava age versus latitude. All data indicate a transition to a more tectonically dominated regime at the southern end of this segment, when compared to the northern portion. Despite these observations, it is noteworthy that a local minima in cumulate fissure width (one of three such minima along the 9-10°N segment) occurs just north of 9° 35' N. Each of the other two minima on the 9-10°N segment are located in regions of very recent magmatic activity (-9° 50' and 9° 18'), and it has been proposed that local minima in the cumulate fissure width mark the areas of most recent magmatic activity along the ridge-axis. In summary, the region of the fourth-order segment between 9° 34.9' N and 9° 37. 1 ' N latitude marks a significant transition in geologic character of the sea floor. Differences north and south of this region are apparent in the nature and abundance of fissuring, age of axial lavas, and the morphology and developmental stage of the ASCT. Additionally, gradients in each of these characteristics can be seen within the fourth-order

PAGE 111

104 segment itself, supporting the hypothesis that this portion of the ridge not only marks a boundary in the geologic character of the seafloor, but may be an area of transition between regimes of different magmatic activity. All geologic observations of the seafloor structure are consistent with the hypothesis that the ridge immediately south of Segment D is magmatically less active than the region to the north, and is instead more strongly affected by tectonic processes and amagmatic extension. As will be discussed below, petrologic and hydrologic samples recovered by submersible further support these hypotheses. 3.3. Data Acquisition and Alvin Observations Detailed visual and photographic data from Segment D were obtained during Alvin dives in 1991 and 1994. Optical data, geologic samples and hydrothermal fluid samples were recovered primarily from two dives to this area (Alvin dives 2371 and 2750), and a limited amount of data was recovered during a third dive (2753) aborted early due to mechanical difficulties. Dive tracks for the completed dives and sample localities are shown in Figures 3-3c and 3-4. Dives were navigated using an array of bottom-moored acoustic transponders that provide an accuracy to within 5 m in both absolute geodetic and relative navigational frameworks [Haymon et al., 1991]. This navigational accuracy is supported by the observed ability to navigate to hydrothermal and geologic features mapped in the 1989 Argo survey to within -10 m of their observed location [Fornari et al., 1998b]. 3.3.1 In situ Geologic Observations. Direct submersible observations of "Segment D" extend as far south as 9° 36' N (9.60° N). They include investigation of both the eastern and western ASCTs within the

PAGE 112

Figure 3-4: Map showing Alvin dive tracks, sample locations and geologic interpretation for the region of the 9° 37' OSC. Dive track 2371 is shown as a gray line and 2750 as a black line. ASCT boundaries as mapped in 1989 are indicated by a discontinuous black outline. Basalt sample localities (gray filled diamonds), diffuse hydrothermal fluid samples (black filled circles), and higher temperature focussed hydrothermal fluids from "R" vent (black filled triangles) are shown. Stippled region represents the interpreted extent of the eastern ASCT, which both lengthened and widened between 1989 and 1991. Southern limit of ASCT advance is unconstrained by submersible observation, but based on the observation of flocculated microbial material in the water column, may extend at least as far south as the dashed line. Presence of pillow lavas is shown by graphic. Horizontal dashed line marks the segment C-D boundary after Haymon et al. [1991], and arrows mark the location of numerous extinct sulfide edifices.

PAGE 113

106 •i ' • ' t t I • • : : • • i 9.62 C/3
PAGE 114

107 zone of overlap, the region of seafloor between the two overlapping ASCTs, and the region south and east of the eastern ASCT (Figure 3-4). Dive coverage south of the overlap zone (as mapped in 1989) extends to -9° 36.5' N (9.608° N) along the extension of the eastern ASCT trend. Dive coverage continues southward from there to -9° 36' N (9.6° N), but is -300 m east of that trend and 500-600 meters east of the western ASCT (Figure 3-4). Submersible coverage of Segment D's western ASCT terrain is limited to the region north of -9° 36.7' N (9.612° N). Visual and photographic data consistently show the volcanic terrain of the eastern ASCT and the southward extension of its trend to be younger in appearance than the region of the western ASCT investigated. Glassy lava surfaces of the freshest lavas in the eastern ASCT terrain (which is considered here to include a region at least 300-400 m south of its boundaries as mapped in 1989) are consistently more vitreous in appearance than the freshest lavas seen in the western ASCT. Lavas comprising the floor of the western ASCT are largely broken up sheet and lobate lava forms, and were observed to have significant sediment cover and dull "fuzzy" appearances due to Mn and Fe oxide deposits on glass surfaces. Basalt talus piles stained bright orange by surficial coatings on the basalt were observed in numerous localities. Several extinct or weakly venting sulfide edifices were identified in this area; many of considerable size (greater than 5-8 meters in height and several meters in diameter) based on in-sub visual observations. The western ASCT extends north to approximately 9° 37' N, at which point it becomes a series of discontinuous ridge parallel elongate collapse pits extending to 9° 37. 1 ' N., the northern terminus of Segment D.

PAGE 115

108 In contrast to the western ASCT, the eastern ASCT, which extends from Segment "C" southward across the Segment C-D boundary has a volcanic terrain that is much younger appearance than its western counterpart. The youngest appearing lavas observed along the eastern ASCT terrain display less sediment cover than the youngest lavas of the western ASCT, and glassy lava crusts have a more vitreous appearance. The lava flows flooring the eastern ASCT are low relief lobate flows, lineated sheet flows and jumbled sheet flows. Flow contact relationships indicate that the lineated and jumbled sheet flows are stratigraphically the youngest, although the underlying lobate flows are also fresh in appearance and the contact relationship, while clear in places, is not always so. Extensive areas of collapsed volcanic terrain indicative of ASCT development, including lava pillars and archways [Fornari et al., 1998b], extend well beyond the boundaries that were mapped in 1989 using the ARGO system [Haymon et al, 1991 ; Wright et al., 1995]. ASCT-like collapse extends at least as far south as 9° 36.5' N, the southern most locale where the submersible survey crosses the southward extension of the eastern ASCT trend. The possibility that areas of extensive volcanogenic collapse extend even further southward along this trend cannot be precluded with the current data. The area of seafloor in between the western and eastern ASCTs is crossed by the dive tracks in three locations, roughly in the southern, central and northern portion of submersible coverage (Figure 3-4). Old looking (Age 2) lavas and relatively heavy sediment cover generally characterize this region of seafloor. Numerous discontinuous and roughly ridge-parallel fissures in this region, and lavas forms include lobate and channelized sheet flows and several outcroppings of constructional pillow mounds or ridges (limited submersible coverage limit the ability to accurately characterize the true

PAGE 116

109 dimensions of these features). The occurrence of pillow lavas in this region is surprising, as throughout the ridge between 9° 17-52' N pillow ridges have only been observed well outside of the "neovolcanic zone" as expressed by the ASCT location. Small pillow basalt ridges/mounds were also observed eastern ridge flank some 500-600 m east of the western ASCT, and south of the southern extension of the eastern ASCT (Figure 3-4). 3.3.2 Hydrothermal and Biologic Activity In 1991, i4/vj>z dive 2371 documented extensive areas of diffuse venting of hydrothermal fluids directly out of sea floor basalt and enhanced biologic activity along the southern extension of the Segment C ASCT, as it was mapped in 1989. Diffuse venting and biologic activity extended well south of the 9° 37.1' N southern boundary of Segment C, and was confirmed as far south as 9° 36.5' N (9.61° N). No sulfide structures or focused high temperature venting was observed in this region, but diffusely venting fluids with measured temperatures of up to 39° C were observed at several sites. Biologic activity observed in 1991 included extensive areas of seafloor blanketed by white microbial mats several cm thick. As with the area from 9° 46-52' N affected by the 1991 eruption, this microbial biomass was pervasive and associated with abundant diffuse venting of warm hydrothermal fluids directly out of seafloor basaltic flows. In places the microbial covering of the seafloor was thick enough as to cause "white out" conditions when trying to land the submersible. In addition to the bacteria, other types of vent associated fauna were observed in the region of diffuse venting between 9° 37.1' and 36.5' N, including the tube worm species Alvinella and galatheid crabs. Diffuse venting of hydrothermal fluids was not only observed emanating from the seafloor, but also from the tops of lava pillars. Venting lava pillars have only been documented in two other

PAGE 117

locations (the 9° 46-52' area and at Axial Seamount on the Juan de Fuca Ridge [Haymon et al., 1993; Perfit, personal comm.]), and in each of these cases, the area had been subject to an eruptive event in the very recent past (less than 2 years). Diffusely venting hydrothermal fluid samples were collected in several locations along the eastern ASCT in 1991, and from a focussed, higher temperature venting sulfide structure within the western ASCT in 1991 and 1994. Data for fluids sampled from the western ASCT "black smoker" chimney, referred to as "R" vent have been reported elsewhere as part of a broader second-order segment scale investigation of hydrothermal vent fluid chemistry for this segment [Von Damm, in press]. Typically, because of the mechanics involved in the sampling of submarine hydrothermal fluids, some seawater is entrained during the sampling process because of both the "dead space" present in the sampling bottles and entrainment of ambient seawater during sampling. Not surprisingly, this effect is even more pronounced when sampling diffusely venting fluids as opposed to a more focussed flow. For this reason, fluid chemistries are typically reported as "end member" fluid chemistry by extrapolating measured values along a linear regression of their correlation with Mg, which is quantitatively removed during hydrothermal fluidbasalt interaction in the shallow ocean crust. Because pure end member hydrothermal fluids are devoid of Mg [Bischoff and Dickson, 1975], end member fluid compositions are calculated by regressing measured "impure" fluid compositions along a line that is weighted to pass through the seawater Mg value and extrapolated to Mg = 0 mmol/kg. Chloride (CI) is the primary anionic species seawater and in hydrothermal fluids, largely because the two other major seawater anionic species, sulfate and alkalinity, are largely removed during heating of the fluid and interaction with crustal rocks. Chloride,

PAGE 118

Ill however, behaves in a largely conservative fashion during hydrothermal alteration of oceanic crust, and is primarily only greatly affected (in particular reduced) by the process of phase separation [Butterfield et al., 1990 and 1994; Von Damm, 1995]. As the major anion, CI plays a large role in the flux of cationic species, which largely mobilize as CI complexes. For that reason and because it serves as an excellent indicator of the physical process of phase separation, it is one of the most commonly measured chemical characteristics of hydrothermal fluids. Samples of 39°C diffusely venting fluids recovered in 1991 were found to contain 519 mmol/kg of CI (actual measured concentration) [Von Damm, unpublished data]. Extrapolation to a Mg=0 mmol/kg yields and end member hydrothermal fluid with a CI composition of 206±1 mmol/kg. This value (both the measured and extrapolated end member) is significantly less than the 540 mmol/kg of CI present in ambient seawater indicating that these fluids represent the "vapor phase" of fluids that underwent subcritical phase separation [Von Damm, 1995 and in press]. In contrast to this, high temperature fluids (< 123°C) recovered from "R" vent in 1991 and 1994 had end member CI = 805±20 mmol/kg, one and a half times that of seawater. Von Damm [in press] concluded that the chemistry of "R" vent fluids are consistent with them being the "brine phase" of phase separated hydrothermal fluids. 3,4. Basalt Geochemistry 3.4.1 Basalt Sample Recovery and Analytical Procedures. Basalt geochemical data from Segment D include analyses of 21 samples recovered by submersible within the general axial region and 89 samples recovered by wax core from 78 sites within the broader region of the crestal plateau. Wax core

PAGE 119

112 samples were recovered from an approximately 2x5 km rectangular array of sample targets centered on the OSC at 9° 37' N. Sample spacing is less than 400 m in a N-S dimension and less than 200 m in the E-W direction, and precision of sample locations is generally better than the spacing interval. The overall coverage of the sample grid is from -9° 36.5-37.5' N latitude and 104° 14-16.8' W longitude. Basalt glasses were analyzed for major element oxide concentration by electron microprobe (EMP) using facilities at Lamont Doherty Earth Observatory (LDEO) and the United States Geologic Survey (USGS) Branch of Geochemistry in Denver, CO. Operating conditions for EMP analyses can be found in Reynolds [1995] and Arbogast [1990] for the LDEO and USGS analyses respectively. All EMP data were beta corrected to minimize analytically generated variance due to differences in focus, polish and carbon coating. LDEO data were corrected after Reynolds [1995]. Strong correlation between sum of all oxides (Total) and an individual oxide (an indicator that the use of beta corrections are appropriate) were only observed for Si02 in the USGS EMP data, and therefore only Si02 was corrected in this data set. Measured SiOa values were adjusted along a linear regression between Oxide Total and SiOi to a Total value of 99.7, that of the MORB standard JdF. All corrected data was then normalized to the MORB standard JdF [Reynolds, 1995] which was analyzed multiple times during all analytical runs in order to facilitate direct comparison of this data to other published MORB data. Analytical precision of MORB standards run throughout the length of the analytical program are given by Rubin et al. [1998], and are similar to those presented in chapter 2. Major element data and locations for the Alvin and wax core samples recovered from Segment D are shown in Tables 3-1 and 3-2 respectively.

PAGE 120

113 o O Oh o O z o u o 60 o c O b O < o o T3 D 'Sib c o -J u Co q o CO q r~ m CN 00 00 m 00 o rOS CO d so CO so d so so CO so so m so so so od in CN SO (N so CN SO d so 00 in On in CO so CN so r 00 so WN CO Os f — 00 >n so 00 ON CO 00 od so so ro 00 — ON so 00 t~so so r~r~ ON On ON ON On ON ON ON On ON ON ON ON ON On d o On On OS On ON On ON On ON On ON ON ON ON ON On ON Os ON Os •d -q SO m OO CO in 'O CO SO in t CO d d c O o c O O O o C O O O O o o o O ro CN o o ON o OS o — CO CN CN CO CO o d d d d d d d d d d d d d d d d d d d OS so o r~ so so so so 00 OO CN SO CN SO OO OO so o CO r~ON in CO 00 in n OO CN CN OO 00 in 00 r00 00 00 00 r-^ 00 OO 00 00 OO o -6 o CN -6 o CN -d a o CN CN CN CM o CN so o CN m d c d c d d d d d d d d d d d d d d d 10.0 10.3 ON 10.2 On ON ON 11.0 CO ON CN ON 10.1 10.7 so Os rON ON OS OO ON 10.0 10.6 10.5 so ON q so q q CO CO so in so m m CO vri in in in in in OO in >ri CN »o 00 CM CO SO CM o CO so Os ON ^. so o in Os SO Os CM so 00 m in SO m O d d d d d UO d d in d in d m d in d >n ON ON d in d in d m d m d in d r-minininsoinininxin-'^in'^cocoTtvn'^ o r-~ r-O o CN CM 00 o o o O so so so SO On ON ON o^ in r-~ — — r~ 00 00 o o so CM CN CM CM ON OS O O 00 00 rm m in in m in in in CN CN CN CM CN CN CN CM O O o o o o o o so so r~rCN r-rOO CM CO CO o o o o O so so so so SO so so so ON ON ON C3N O^ o^ ON o in C<-| OO in in o o CO so in CN CN in o in m in in m in in in in CM CM CN CM CN CN CM CN o o o o o o O o o so ON Os in in ON On ON sOONCOSOOscOincNTtVDSDO inom — CM — •<:) — ocNcoTjOCMCNcOCO-rl-TtmOOO CN so — n in in m in rrrrCN CN CM CN CN CN (N CN o o .2? ^ Q < V o c

PAGE 121

114 ca Z o U o O c 00 o OS 00 00 o m in 00 NO o in NO ON m 00 NO ON in ON m d vo NO NO m ON IT) d NO NO in ON in NO NO in NO NO in NO o^ in (N NO m ON m ON in (N NO (N NO NO NO 00 in NO m o^ in NO o (N ON o 00 p NO in 00 in On cs On CO rn NO rn (N (N NO rNO 00 n >o (N On NO o6 r-^ 00 o6 in NO NO 00 00 ON NO NO 00 NO 00 rr00 NO \o NO NO NO m NO NO in NO NO NO NO NO in rr-; On t-; NO On 00 On On Qs On On On ON On ON On On On ON On On On o^ On On On On On On On ON ON ON On On ON WN On WN ON ON On On ON On ON ON On ON On a\ On XJ -a T3 •n a •a T3 -a T3 T3 •o a •o T3 T3 •a -a CN m O in r~m C c C c c c C c C c C c c c C C c c s o o o o O o o o o o CM o m NO CO CM o 00 o o o m o On O CN o o m d d d d d d d d d d d d d d d d d d d d d d d d d d d d d \o rm o 0\ ON >ri On 00 00 in 00 o o NO in (N On NO in ON NO m in ON in 00 NO NO m NO NO rin NO NO On NO NO NO >n NO (N CN (N (N (N r4 04 (N (N ri NO 00 in r~ 00 Nqt~~;r<^>ninroooNqinNOinroinr-'— roNONOcN — oor-ooNoo — r-^t~~r~^r~^t~-r-^r--odo6r-^odKodr~-^odt^r-^r^ ON 00 00 (N 00 00 On o On OO ON On o On NO ON 00 o 00 in o NO ON On n iri in in mi in in in mi in mi ml NO mi ^' mi o m 00 00 oo 00 NO m ON o oo NO in NO in On NO o m in CO NO m r-NO NO in (N NO m rNO NO in fN NO NO NO m m in NO in NO m m rin ONl in NO NO m rd d d d d d d d d d d d d d d d d d d d d d d On d d d d d m m in in in m m in in in m in in m m in m in in in in m in m m in in m NO in rNO NO NO NO NO m in 00 NO NO in in m in NO in in NO in NO in 00 roo in in o (N 00 t-~00 in m r<-i OO o r<-i NO NO NO NO NO NO m 00 oo m NO NO m in NO rm m o O ON NO NO NO o o O (N (N (N CN fN NO NO NO NO NO NO NO NO NO NO NO NO NO NO NO NO NO NO NO NO NO NO NO NO NO NO ON On On ON On On On ON ON ON ON On ON ON ON On On ON ON ON ON On ON ON O^ ON ON o O c75 T3 3 c O O a. Q a. E ON On rn nm rj cnI o o rNO NO m NO NO O O o o o NO NO o o NO O m NO NO in CN| (N ^ ^ o o o m 04 Tio fN cn 00 in m NO rrin NO ro m ON CM NO o m m in in m m 00
PAGE 122

115 CO z O U o o o O < O H q O Ov as VO ON 00 m ri NO d NO m VO VO NO rj NO VO NO d VO d VO NO NO On in in NO VO 00 00 VD 00 r— CO o 00 q 00 00 q VO 24.9 24.7 q r~00 00 00 00 00 00 00 00 00 r-00 rr-rr~ ON ON On ON ON o\ ON On o o ON ON On On ON ON as ON On ON Ov On ON ON ON On ON On Ov On as Ov Ov ON ON Ov Ov On ON Ov Ov Ov ON ON o^ UN Ov tJN ON o< ON Ov IJN Ov ON O ON On ON in ON rrVO 00 ON (N NO NO O ONl m m o in d d d d d d d d d d d d d d d d d d d d d d d d d d d d d 00 o o IT) o o ON o ON o On O o d d d d d d d d d d d d d d d d d d d d d d d d d d d d d rVO VO 00 VO 00 00
PAGE 123

116 -a c o o I — X) CO H C3 z o u o 60 O ri NO oi NO 00 >n d NO On in ON IT) NO U-) NO in o NO 00 in 00 in in NO in NO in NO 00 in NO m NO d NO On m NO [ Lr, ON m 00 00 q NO (S q NO On in ri rNO On O in NO rO d d d d d d d d d d d d d d d d d d d d d d d d d d d d o in o O CN m CM m NO •n en 00 O o 00 o 00 o 00 00 r-m o ro CM o in rrNO O r<-) On NO ON NO o NO o NO On in ON ON NO o m m CN o On NO NO NO in 'stm Tjm rr-NO NO NO NO NO in NO NO NO CM (N CN CN CN CN CN CN CN CN CN CN CN CN (N CN CN CM CN CN CM CN CN CN CN ^' •*
PAGE 124

117 Table 3-3. Trace element concentration in representative samples analyzed by laser ablation inductively coupled mass spectrometry (LA ICP-MS) P Mantle 2495-5 2372-1 ARC 93 ARC 95 ARC 97 ARC 99 0 032 0.08 0.012 0.101 0.038 0.013 0.007 Rb 0 635 6.1 0.65 4.1 2.4 0.84 0.46 Ra 6.989 36 8.5 47 28 9.2 6.3 Th i. 11 11. u. n a 11. cl. n.a. n.a. n.a. n.a. n.a. TT yj 11. £1. n a 11. £1. n.a. n.a. n.a. n.a. n.a. Nh 0 713 V/. / 1 J 9 8 2.1 8.8 5.7 2.5 1.6 Ta 0 041 <0.2 0.19 0.7 0.46 0.2 0.13 0.0301 0.36 0.09 0.42 0.26 0.13 0.10 La 0.687 9.2 3.7 8.2 6.0 4.3 3.0 Ce 1.775 22.0 10.8 18.9 14.8 12.6 8.8 Pr 0.276 3.3 1.9 2.9 2.3 2.3 1.5 Sr 21 165 126 183 147 117 114 Nd 1.4 17.0 10.2 13.9 11.9 12.1 8.2 Zr 11.2 151 111 139 128 134 89 Hf 0.309 3.8 3.0 3.6 3.3 3.7 2.5 Sm 0.444 4.6 3.6 4.2 3.8 4.3 2.9 Eu 0.168 1.6 1.4 1.6 1.4 1.5 1.1 Ti02 0.21685 1.8 1.6 1.6 1.7 1.8 1.3 Tb 0.108 1 0.93 0.97 0.96 1.11 0.76 Dy 0.737 6.4 6.4 6.6 6.5 7.4 5.4 Y 4.55 39 35 35 35 41 29 Ho 0.164 1.4 1.3 1.3 1.3 1.6 1.1 Er 0.48 3.9 3.9 3.8 3.8 4.6 3.2 Yb 0.493 3.7 3.7 3.6 3.7 4.2 3.1 notes: All data reported as parts per million (ppm) except K20 and Ti02 which are reported as weight % oxide. Primitive mantle concentrations after Sun and McDonough [1989]

PAGE 125

118 A limited subset of wax core samples was analyzed for trace element concentrations by laser ablation inductively coupled plasma source mass spectrometry (LA ICP-MS) at the USGS in Denver. Details of operational procedures, instrumental conditions, and analytical detection limits and precision are given by Ridley and Lichte [1998]. Concentrations of select trace elements for samples analyzed by LA ICP-MS are shown in Table 3-3. Examination of MORB major element data from the entire region between -9° 17' N and 10° N (Perfit et al., 1994; our unpublished data) and comparison of that data to the data from Segment D basalt samples reveals several notable relationships. 3.4.2 Second-Order Segment Scale Geochemical Observations Nearly 850 well located basalt samples from the EPR axis and crestal plateau between -9° 17' and 10° N have been analyzed for element concentrations [Perfit et al., in prep.]. The majority of these samples have elemental compositions characteristic of normal incompatible element depleted mid-ocean ridge basalt (N-MORB). A small, but significant, percentage (less than -15%) of the samples recovered outside of the ridgeaxis on the crestal plateau have elemental signatures that are transitional to incompatible element enriched basalt. These elementally enriched MORB have been referred to as TMORB or E-MORB (referring to their transitional and enriched character), depending on the level of incompatible element enrichments observed [Sinton et al., 1991; Batiza and Niu, 1992; Reynolds et al., 1992; Perfit et al., 1994]. The boundary at which a T-MORB becomes an E-MORB has not been well defined, and this group of basalts has generally been treated as a continuum of compositions resulting from the mixing of enriched and depleted melts or melt sources. As such, this paper will simply refer to the group of

PAGE 126

119 elementally enriched MORS recovered on the 9-10° N segment as E-MORB, noting that this designation may not correspond to similar levels of elemental enrichment reserved for the E-MORB classification in other existing literature. Differences in the ratios of highly incompatible elements [Perfit et al.l994; Batiza and Niu, 1992] and isotopic signatures [Sims et al., 1997b] have led investigators to conclude that E-MORB from the 9-10° N segment of the EPR (and E-MORB in general) do not have a direct genetic link to the more voluminous N-MORB. It has been further hypothesized that these E-MORB have retained their geochemically enriched signatures by avoiding significant mixing in an axial magma chamber [Reynolds et al., 1992; Perfit et al., 1994; Perfit and Chadwick, 1998]. This conclusion is consistent with the observation that no E-MORB have been recovered within or proximal to the ASCT between 9° 17' and 10° N where an AMC has been imaged [Perfit and Chadwick, 1998]. In order to accurately assess recent axial magmatic conditions and processes, the geochemical database for the 9-10° N segment requires filtering to remove data from EMORB, which are considered to have origins distinct from typical AMC melts. Because many of the samples in question were recovered by wax core and have only been analyzed for major element concentrations, the ratio K20/Ti02 serves as the best available discriminator. Examination of the relationship between K20/Ti02 and MgO measured in MORB glasses (Figure 3-5) reveals several important observations allowing for the discrimination of E-MORB. Firstly, no significant correlation exists between K20/Ti02 and MgO. Because MgO is strongly controlled by crystallization and assimilation processes, a lack of correlation with K20/Ti02 implies that K20/Ti02 values are

PAGE 127

120 Figure 3-5: K^O/TiO: plotted against MgO (weight %) for all lavas recovered from the 9°-10° N second order segment. Maximum possible variation in K20/Ti02 is less than 1 6% (c.f. Figure 2-9) of the value. No significant correlation between MgO and K20/Ti02 is observed. E-MORB fall above the horizontal line at a K20/Ti02 value of 1 1 or greater (see text).

PAGE 128

121 established during magma genesis in the mantle, and that subsequent crustal level modification would most likely be the result of mixing processes. That KiO/TiOi is a genetic trait inherited from parental magmas confirms its use as a discriminator of EPR magmas with different genetic origins [Langmuir et al., 1986; Sinton et al.. 1991; Reynolds et al., 1992; Perfit et al., 1994]. Additionally, investigations of other ridge segments have also shown KiO/TiO: to be an accurate measure of mantle derived geochemical enrichment by showing strong correlation between KiO/TiOi and other mantle derived geochemical characteristics including Sr isotopic ratio, Zr/Y and La/Sm [Smith et al., in prep]. Secondly, examination of Figure 3-5 reveals a break in the majority of the K20/Ti02 data at a value of K20/Ti02 =1 1 , with the vast majority of samples having a value of < 10 (rounded to the nearest whole number which is appropriate to the analytical precision of this parameter). That this K20/Ti02 level serves as a reasonable dividing line between the Nand E-MORB is supported by several observations. Firstly, if one evaluates the distribution of KiO/TiO^ values observed in MORB recovered from the wax core grids that sampled the crestal plateau regions around 9° 30' N and 9° 37' N (Figure 3-2), one finds that there is a bimodal distribution of values with K20/Ti02 being either < 10 or > 15. No values between 10 and 15 are in evidence in what is otherwise a relative continuum of values between ~5 and 25. This relationship, however, is not as clear in data from the wax core grid centered around 9° 50' N where a limited number of samples recovered outside of the axial region have K20/Ti02 intermediate between 10 and 15 (-15% of 189 samples recovered from the sampling grid). Secondly, only 1 of 294 analyzed samples recovered within or proximal to (within 300-500 m) the ASCT (i.e.

PAGE 129

directly associated with the axial melt lens) has a K20/Ti02 value greater than 10, and that one sample is an age 2 lava from the 9° 50' N region, recovered outside of the ASCT. This observation implies that lavas along the 9-10° N segment, lavas emanating from the axial melt lens characteristically have K20/Ti02< 10. Lastly, an examination of the "off-axis" crestal plateau lavas in the 9° 50' region, for which there is greater submersible coverage and a greater amount of trace element data that can be correlated to the major element data, shows there to be mantle derived trace element distinctions between lavas with K20/Ti02 < 10 and those with values greater than 10. These distinctions include enrichments in both Sr abundance and the ratio Zr/Y in samples with K20/Ti02 greater than 10 [Perfit et al., in prep.]. For all of the above reasons, samples w/ a K20/Ti02 > 1 1 have been filtered out of the N-MORB database, being considered E-MORB. This value is roughly similar to the value of 9 used to discriminate N-MORB from T-MORB on the EPR at -12° N [Reynolds et al., 1992]. After exclusion of the aforementioned E-MORB, 695 analyzed basalt samples recovered onand off-axis by submersible (378) or wax core (317) are included in the N-MORB database. In order to discriminate the most recent axial magmatic activity from possible longer temporal signals, further subdivision of the 9°-10° N N-MORB database is required, distinguishing those samples recovered very proximal to (less than -300 m) or within the ASCT from those recovered further off-axis. Although evidence suggests that eruptive activity (including both Nand E-MORB) can occur on this segment as far as 4 km off-axis [Perfit et al., 1994], the inability to unambiguously evaluate the entire database as to which of the off-axis samples recovered were erupted within the axis then

PAGE 130

123 moved off-axis by subsequent seafloor spreading and which were more recently erupted off-axis requires that those samples considered to be most recent include only samples recovered either within or proximal to the ASCT (within a ~300m). After applying this second filter to the submersible recovered database (very few wax cores were recovered from within the axis) 272 axial N-MORB analyses remain (192 north of Segment D, 58 south of Segment D and 22 within Segment D). Axial N-MORB data for the oxides MgO and TiOa are shown in Figure 3-6. A field that encompasses the total range of all off-axis N-MORB sampled between 9° 17' and 10° N is shown for reference. Axial MORB exhibit an overall range of MgO content from -9.3 to 7.1 wt.%. Consideration of off-axis N-MORB extends the range to more evolved (relative to primitive, high MgO MORB) compositions with MgO as low as 5.9 weight percent (wt.%). The overall trend of data for axial N-MORB (and off-axis NMORB as well) corresponds well to a calculated liquid line of descent (LLD) that assumes low-pressure (Ikbar) fractional crystallization as the magmatic differentiation process [Weaver and Langmuir, 1990]. This implies that the dominant control of geochemical heterogeneity among the N-MORB suite in the 9°-10°N region is cooling, crystallization and subsequent differentiation of magma bodies within the shallow oceanic lithosphere. This conclusion is in accord with many other investigations of medium to fast spreading ridge environments [e.g. Batiza and Niu, 1992; Smith et al., 1994]. While significant heterogeneity in MgO content within this lava suite exists, compositional variation at a given MgO is quite limited. For N-MORB recovered from within the axis, variation in Ti02 at a given MgO content is representative of the variation observed in the other major oxides at a given MgO and is quite limited, being

PAGE 131

C3 c« (N ~" O Ml c: ° ?^ y n >+c o T3 N > O o O 00 •m o ^ ^ (N S 2i +' u 15 c _o o 03 ^ u i_ ^ 2 c c c ^ o 03 ri "O O' 2i £ I— O (U O O "r; U « « (U ^ On rC .E u 5 II c -E u m ^ I m 3 O 9J = 2 03 OJ Do o t — '-^

PAGE 133

126 barely outside of (and less than twice that of) analytical uncertainty. This leads to the conclusion that the entire range of observed compositions in axial N-MORB can be generated by similar crystallization conditions of very similar parent magma compositions. That axial basalt in this area can all be generated from common or very similar parental compositions was also noted by Batiza et al. [1992], who, unlike Langmuir et al. [1986] found no evidence of major changes on MORB parental chemistry associated with fourth-order ridge-axis discontinuities in this region. Inclusion of offaxis samples does, however, as much as double the range of Ti02 observed at a given value of MgO. In order to facilitate the evaluation that Segment D marks a significant change in the character of the ridge crest along the 9°-10°N segment, axial MORB samples plotted in Figure 3-6 have been divided into two subsets, those south of Segment D (open circles) and those north of Segment D (gray filled diamonds). Note that samples from the axis of Segment D itself are not shown on this diagram, but fall within the overall field for axial N-MORB. Samples from each of these subsets fall into distinctly different groups based upon their MgO content (any consequently many other elements displaying strong correlation to MgO), which is the best available discriminator since all axial N-MORB along this second-order segment appear to be derived from similar parents and crustal differentiation processes. North of Segment D, 95% of all (192) axial samples have a MgO content greater than 7.9 wt.%, with none having compositions more evolved than 7.6 wt.% MgO. All of the samples north of Segment D that are more evolved than 7.9 wt.% MgO were recovered on Segment C, the fourth-order segment adjacent to, and immediately north of

PAGE 134

127 Segment D. South of Segment D, the majority of axial lavas fall within a more limited compositional range of 7.1-7.9 wt.% MgO, with only 5 of 58 samples falling outside of (to higher MgO contents) that range. Liquid line of descent calculations [Weaver and Langmuir, 1990] give melt temperature estimates that ranging from 1210-1 170°C for the axial N-MORB glass data shown in Figure 3-6. With the exception of the five aforementioned samples, data south of Segment D have calculated melt temperatures less than 1185°C whereas those north of Segment D generally have calculated melt temperatures greater than 1 185°C. Figure 3-7 illustrates chemical variation along the second-order segment by examining the relationship between latitude and Magnesium number (Mg#). Mg# represents the ratio of atomic Mg to the sum of atomic Mg and Fe""^, and it is strongly correlated to both MgO abundance and magma temperature. If one excludes the Segment D data which exhibits a broader range of Mg# values than at any other latitude on the 9°10° N segment, a strong positive correlation exists between latitude and Mg# for the portion of ridge south of Segment A. An exception to this are the five aforementioned samples south of Segment D that have compositions anomalously mafic (MgO rich) for their latitude. These 5 samples are indicated in Figure 3-3b by circles filled by crosses, and are consistently among the oldest (all age 2) samples identified in the region of their recovery. Linear regression of the remaining data yields a t~ value of 0.807. Smooth latitudinal correlation of Mg# at the second-order segment scale have previously been noted for this segment and portions of the southern EPR [Batiza and Niu, 1992; Sinton et al., 1991].

PAGE 135

(11
  • ° -a • c — O C 03 .E :5 O <*. -a — > o U u w)x: a: _^ 3 -a ^ ^ ^ 03 + U + 00 o U t/3 t/3 5 z — J? 3 O ^ — U C3 E u O t/3 > > O «3 •— O o OX) 3 2 « _ « = 00 ON ON T3 s: U a c 03 1/3 o .2 c flU C "z: 3 o > o (U > C3

    PAGE 136

    129 T — ' — W o On ^^^^ C C < CL _ 1— CD CD W CO oo CM CO < CO ^
    PAGE 137

    130 3.4.3 Fourth-Order Segment Scale Geochemical Observations of Segment D Some limited, but significant, systematic relationships between lava chemistry and ridge-axis (ASCT) geometry are observed in the region of the 9° 37' N OSC. The youngest lava flows flooring the eastern limb ASCT (which extends southward from Segment C) are the most mafic recovered along the segment having MgO concentrations of 8.4-8.5 wt.%, and they fall well above the regional trend in Mg# established in Figure 3-7. These young sheet flows (samples 2371-3, 1 1 and 13) were sampled in three localities along the eastern ASCT over an approximately 1 km length of ridge, and all have major element compositions indistinguishable from one another. Additionally, these lavas have Mg#s ranging from 63.7 to 64.5 (indistinguishable given analytical uncertainty) identical to the average Mg# of 64 measured in the 1991 eruptive to the north suggesting they may be related to the same pulse of magmatic/volcanic activity. Other very fresh (age 1), but slightly older lavas from the eastern ASTC and slightly east of it (2371-4, 5, 14 and 2750-3a, 3b and 4) tend to have lobate and sheet forms, and have slightly more evolved compositions ranging between 7.8 and 8.1 wt.% MgO. MORB recovered from the ASCT associated with the western OSC limb are limited to two samples, but the few sampled are more chemically evolved than the most recent flows on the eastern ASCT. The most evolved (lowest MgO content) lavas are consistently found in the region of overlap between the two ASCTs. In this region pillow lavas are common and the four samples recovered from this region (2371-10 and 15 and 2750-6 and 7) all have MgO contents less than 7.8 wt.% (as low as 7.3) and Mg#s below the regional trend.

    PAGE 138

    131 3.5. Discussion 3.5.1 Southward Aging of Volcanic Activity Numerous geologic observations support the hypothesis that there is an overall southward decrease in magmatic input and volcanic activity between -9° 52' and 9° 17' N. Differences in axial depth, density and character of crustal fissuring, size and shape of the ASCT, apparent age of axial lavas, crustal density, and lava eruption temperature (Mg#) all are suggestive that southward along the ridge-axis conditions become less magmatically/volcanically robust, the axis has a more mature volcanic terrain, and it is more affected by tectonic stretching and amagmatic extension [Haymon et al., 1991; Wright et al., 1995; Fomari et al., 1998b; Cochran et al., 1999]. The southward decrease in melt temperature (Mg#) illustrated in Figure 3-7 has also been previously recognized by Batiza and Niu [1992], although it is not as well constrained in their data as it is by data shown in Figure 3-7. The poorer resolution and greater scatter about the trend presented by Batiza and Niu [1992] is most likely due to the lower sampling density (dredges every 1.8 km), lesser overall number of samples involved in the study, lack of sample recovery north of 9° 40.5' N, and the poorer spatial control in dredging as compared to recovery by submersible or wax core. Poorer precision in the location of sample recovery limits the ability to efficiently filter out true axial lavas from those that were near, but outside of the zone dominated by recent volcanics erupted from the axial melt lens. Rather than calling for differences in magmatic input to the AMC to account for observed differences in melt temperature, Batiza and Niu [1992] explain the observed southward decrease in melt temperature by postulating the presence of a continuous,

    PAGE 139

    132 chemically zoned magma chamber over the entire second-order segment. According to their model, melt is centrally supplied and cooler, denser, Fe-rich melts gravitationally migrate to the distal, deeper ends of the axial magma chamber. This model makes the assumption that the numerous fourth-order discontinuities present over this second-order segment do not correspond to discontinuities within the AMC that would inhibit such along axis crustal magma migration. This assumption will be further addressed below. Inclusion of N-MORB data from the three wax core grids (Figure 3-2) with the axial data presented in Figure 3-7 greatly reduces, if not completely eliminates the latitudinal correlation with Mg#. This observation may have important ramifications for the time scale of variability in magmatic input to the ridge. The wax core grids at -9° 30', 9° 37', and 9° 50' N extend up to 4 km on either side of the ridge-axis. At a full spreading rate of 1 1 cm/yr, this area represents in excess of 70,000 years of seafloor spreading. On the one hand, correlation of lava geochemistry and latitude seen in ridgeaxis lavas, but not observed over the broader crestal region may reflect a sudden and recent change in the behavior of this portion of the ridge-axis. This hypothesis seems unjustified and contrary to the principle of uniformitarianism. If, on the other hand, ridge-axis magmatism waxes and wanes through time as suggested by other segment-scale studies of intermediateand fast-spreading ridges [Reynolds et al., 1992; Smith et al., 1994; Batiza et al, 1996], and the current observations of axial zone geology and MORB geochemistry are therefore a snapshot in the temporal evolution of this waxing and waning, then this cycle must occur on a time scale finer than the 70,000 years of crustal accretion represented in the area of wax core survey. Thus, the spatial signal would be muted by the inclusion of several such

    PAGE 140

    133 temporal cycles within the areal/temporal extent of the wax core surveys. This time scale of temporal magmatic evolution (on the order of 10^-10'* years) is generally consistent with estimations of the range of lava ages observed within the neovolcanic zone of the segment and consistent with time scales of formation for observed variations in ASCT morphology. It is, however, important to note that volcanism and crustal accretion are not limited to occurring within the axis. Though its volumetric significance is as of yet to be established, Perfit et al. [1994] established that recent volcanic activity has occurred up to 4 km outside the ASCT near 9° 31.5' N, greatly confusing the relationship between distance from the ridge-axis and lava age. Several authors have recognized that on the EPR the seismic layer 2a (thought to represent the extrusive portion of the crust) increases in thickness by up to a factor of two within a few km of the axis [Harding et al., 1993; Christeson et al., 1994; Vera and Diebold, 1994]. Some have proposed periodic volcanic flooding of the crestal plateau as a mechanism to accomplish the thickening [Hooft et al., 1996]. Given these observations, it is likely that off-axis volcanism will at least dampen any temporal geochemical signal established in the ridge-axis, and it possibly may dominate the broader crestal plateau data. Exactly what the magnitude of off-axis volcanism' s effect on the overall geochemical character of the crestal plateau cannot be adequately addressed until a more precise means of distinguishing axial from off-axis volcanics is achieved. 3.5.2 The 9° 37' OSC as a Magmatic and Hydrothermal Divide Several observations support that the 9° 37' N OSC marks a boundary between different magmatic and hydrothermal systems. Conditions of crustal structure such as

    PAGE 141

    134 lava age, fissure density and fissure widtli make an abrupt rather than continuous change in character at Segment D [Wright et al, 1995]. The large difference in the morphology and apparent age between the eastern and western ASCTs of the OSC, as well as the apparent difference in the age of lavas flooring them, strongly imply that the effects of discrete eruptive events do not extend across the OSC. This contention is further supported by significant differences in hydrothermal venting across the OSC. Numerous large sulfide edifices, both venting and extinct, have been identified within the ASCT of the western limb of the OSC, while no sulfide structures or high temperature focussed hydrothermal venting have been observed within the ASCT of the eastern limb. Vent fluid chemistries north of the OSC have chemistries very different from those south of the OSC. Von Damm [in press] noted that in 1991 hydrothermal vents north of the 9° 37' N OSC (Segments B and C) were commonly venting fluids in excess of 350°C, and that all vents north of that boundary vented hydrothermal fluids with end member chlorinity less than that of seawater. Von Damm [in press] also noted that those hydrothermal vents south of the OSC and north of 9° 17' N (including "R" vent which is in the ASCT of the western OSC limb) all vent hydrothermal fluids with temperatures generally below 300° and end member chlorinity greater than that of seawater. End member hydrothermal fluids with chlorinity less than that of seawater represent the vapor phase of fluids that have undergone the process of subcritical phase separation. Low end-member fluid chlorinities, previously unprecedented rapid changes in fluid chemistry in individual vents, shallow depth of fluid reaction (as evidenced by low end member Si contents), and short apparent crustal residence times of fluids (as evidenced by low Li/Cl and KJC\

    PAGE 142

    135 ratios) led to the conclusion that vents north of the 9° 37' OSC had been affected by the 1991 volcanic eruption between 9° 46 and 52' N [Von Damm, in press]. Conversely, no vents south of the OSC appear to have been effected by the 1991 event. In fact, vents south of the OSC (but north of the recently volcanically active 9° 17' N area) have end member fluid chemistry consistent with being the brine phase of phase separated fluids. If phase separation of hydrothermal fluids is a common process associated with seafloor volcanism, and vapor phase fluids are vented after eruption, venting of the conjugate brine phase may indicate that it has been some time since the hydrothermal system was volcanically perturbed [Von Damm, personal comm.]. All of the above geologic, petrologic and hydrologic arguments above strongly suggest that the 9° 37' N OSC marks a barrier to communication between the magmatic and hydrothermal systems north and south of it. This hypothesis is further bolstered by seismic studies of this region. Seismic tomography and 2-d seismic reflection surveys of the region around the 9° 37' N OSC provide information independent of petrologic and hydrothermal data that is relevant to magmatic activity. Kent et al. [1993b] used seismic reflection common depth point profiles to image the AMC and distribution of magma along this portion of the segment. Although they found the AMC to be continuous across the 9° 17' and 35' DEVALS, Kent et al. [1993b] concluded that the DEVALS each mark a place where the melt lens exhibits a right lateral offset and melt connectivity may be disrupted. Additionally, the 9° 35' DEVAL is associated with the narrowest AMC (250 m) observed on the 9-10° segment. These conclusions are consistent with tomographic data indicating that DEVALS at 9° 28' and 35' mark perturbations in crustal the crustal velocity

    PAGE 143

    136 structure of the ridge-axis [Toomey et al., 1990 and 1994]. Maximum low-velocity anomalies at the center of the fourth-order segments that taper off towards the segment ends (at the DEVALS) led Toomey et al. [1990 and 1994] and Dunn and Toomey [1997] to conclude that along-axis segmentation resulted from melt injection at fourth-order segment midpoints (at -10 km intervals). DEVALS were hypothesized to mark boundaries between different melt injection regimes. The above geophysical arguments support the hypothesis that the 9° 37' OSC marks a divide between different crustal magmatic plumbing systems. 3.5.3 Evidence for a magmatic perturbation at 9° 37' 3.5.3.1 Hydrothermal and biologic evidence for recent volcanic activity Many observations support the hypothesis that the eastern limb of the 9° 37' OSC was affected by a magmatic perturbation between 1989 and 1991. The abundant microbial material covering the seafloor in this area in 1991 was not previously observed in the 1989 ARGO surveys, nor was it present afterward in 1994. An abundance of microbial material covering the seafloor has been only previously observed associated with very recent volcanic activity [Haymon et al., 1993; Embley et al, in press]. Based on time-series investigations of recent eruptive sites on the Juan de Fuca Ridge (Cleft and CoAxial Segments) and the East Pacific Rise (-9° 46-52' N) these bacterial blooms appear to be a relatively short-lived phenomenon. Abundant diffuse venting of hydrothermal fluids directly out of the seafloor as observed on the eastern OSC limb, is also thought to be a relatively short lived condition occurring in areas that have been recently volcanically active [Haymon et al., 1993; Embley et al., in press]. More than just the occurrence of abundant diffuse venting, the chemistry of the vent fluids

    PAGE 144

    137 themselves strongly indicate a very recent magmatic perturbation. Low end member chlorinity of diffuse fluids sampled along the eastern OSC limb indicate venting of vapor phase fluids similar to vents on segments B and C affected by the 1991 eruption at 9° 50'. Additionally, these diffuse fluids have anomalously high-end member ratio of ^He/heat [Lupton, unpublished data]. Anomalously high "^He/heat is also associated with venting shortly after a volcanic eruption [Baker and Lupton, 1990]. Unlike the phase-separated fluids, high ''He/heat requires not just a thermal input to the system, but a magmatic input to supply the excess ^He. 3.5.3.2 Southward propagation of the eastern OSC Umb In addition to biological and hydrothermal evidence, changes in the areal extent of the eastern limb ASCT between 1989 and 1991 also suggest a volcanic event occurred during this time frame. Figure 3-4 illustrates changes in the ASCT configuration between 1989 and 1991. Interpretations are based on submersible observations and indicate that the ASCT extended southward by more than 400 m and widened by several lO's of m between 1989 and 1991. This extension of the ASCT changes the overlap to offset ratio to ~3, the dimensional ratio typical for overlapping spreading centers along the EPR [Macdonald et al., 1984]. It is also intriguing that very fresh lavas, with a similar chemistry to the 1991 eruption confirmed to the north, floor the ASCT at 9° 37'. All of the above observations are consistent with the hypothesis that the eastern limb of the OSC experienced a volcanic event in the time frame between the 1989 ARGO survey and the Alvin dives in 199 1 . Further, it is proposed that the eastern limb of the OSC is actively propagating southward at the expense of the dying western limb, and that this propagation is due to more robust magmatic supply north of the OSC than exists

    PAGE 145

    138 immediately south of it. Migration of this ridge-axis discontinuity is not unprecedented. Northward migration of the 9° 37' discontinuity can be traced south to -9° 20' and fault traces extend nearly thirty km off-axis suggesting that this discontinuity may have been in existence for nearly 500 thousand years [Macdonald, unpublished data]. That the 9° 37' OSC now may be propagating in a southerly direction, opposite to its previous propagation direction is not unprecedented as a similar northward to southward change in propagation direction has been noted for the 11° 45' N OSC [Perram and Macdonald, 1990]. Considering the 400m offset distance of the 9° 37' OSC, its apparent longevity and history of propagation, it is suggested that this discontinuity be classified as thirdorder. 3.5.4 Association of E-type MORB Examination of geochemical data from MORB recovered in the wax core sample grid at 9° 37' N shows that the distribution of E-MORB is neither even nor symmetrical about the ridge-axis in this area (Figure 3-8). Perfit et al. [1994] similariy noted that distribution of E-MORB about the crestal plateau in the vicinity of 9° 3 1 ' N was not symmetrical about the ridge-axis, E-MORB being more abundant on the Cocos plate. Reynolds et al. [1992] also noted a similar asymmetric distribution of enriched MORB compositions in the 12°N region of the EPR. Geochemical data of samples recovered from the crestal plateau around the 9° 37' OSC, however, show a very different distribution of E-MORB than noted previously by Perfit et al. [1994] or Reynolds et al. [1992]. Four times as many E-MORB were recovered on the Pacific plate (8) as were on recovered on the Cocos plate (2). Perhaps more intriguing is the observation that EMORB on the Pacific plate are mostly concentrated within the latitudinal range of the

    PAGE 146

    o 1) t« s: o ^ 3 .a > o u o T3 C C • 03 O 0 t/3 X) C E C/3 § C Z .2 ^ =? o J= 3 C O X) 1 A °S ^ o o ^ ai O 3 re ^ 1(U CJ bO O .E c S .2 « .-a S "5 o 00 I m 3 ili u O

    PAGE 147

    140 en vo as O o o o CP o o o o o o o o o o o o o o o o o in , o o o o

    PAGE 148

    141 OSC, and occur closer to the ridge-axis (ASCT) than E-MORB sampled anywhere else on this second-order segment. Previous studies have noted the association of enriched MORB geochemistry with ridge-axis discontinuities on fast spreading MORs [Langmuir et al., 1986; Sinton et al., 1991], although it has been recently observed that this may not be the case in slower spreading environments [Batiza, 1996; and references therein]. Relative enrichments in incompatible elements between of-axis N-MORB and EMORB recovered around the region of the 9° 37' OSC by wax core (Figure 3-9) suggest that they do not share a common parental magma. Instead, the E-MORB' s greater enrichment of trace elements that are highly incompatible during mantle melting is more likely the result of N-MORB and E-MORB having distinct petrogenetic origins. The occurrence of dispersed, areally limited E-MORB has been frequently observed on the EPR [Langmuir et al., 1986; Sinton et al., 1991; Reynolds et al., 1992; Perfit et al., 1994]. E-MORB incompatible element abundances have been generally modeled to be the result of variable amounts of melt induced mixing of a heterogeneous, possibly veined, upper mantle source [Langmuir et al., 1986; Sinton et al., 1991, Reynolds et al., 1992; Perfit et al., 1994; Niu et al., 1996 and 1999]. Generally, E-MORB have been considered to result from lower extents of melting which, compared to greater extents of melting, preferentially samples lower melting temperature incompatible element enriched mantle heterogeneities. EPR E-MORB are highly subordinate to N-MORB in terms of overall crustal volume, and have been observed to occur off-axis in areas where axial lavas are exclusively N-MORB [Reynolds et al., 1992; Perfit et al., 1994]. This observation led the authors to argue that preservation of E-MORB characteristics in erupted lavas

    PAGE 149

    ^ W) ^ 3 E c CD O X) s = E c O D .S 1B o V ^ u .S C3 — C •a c 3 X) o ^ U -E U E 3 .E o u o c oa O 3 C3 3 a. t3 -o 3 CQ O 1) ca t/5 3 1) E (U 3

    PAGE 150

    143 O X Q 3 N -a c 00 u H J3 H X) u

    PAGE 151

    144 requires that the enriched melt avoid significant mixing with the N-MORB dominated AMC. To satisfy the above observations and the spatial association of E-MORB with axial discontinuities, Reynolds et al. [1992] proposed that E-MORB are always present in small volumes, but that they may only be erupted onto the seafloor under special conditions. These conditions include volcanic eruption either at structural offsets where magmatism is likely to be less robust, during the very early or late stages of a magmatic cycle when magmatic plumbing is poorly developed, or off-axis away from the influences of the AMC. Off-axis E-MORB have been identified around 9° 30' N [Perfit et al., 1994] and around 9° 50' N [Perfit et al., in prep], but E-MORB around the 9° 37' OSC occur much closer to the ridge-axis than in either of these two localities. In general, E-MORB distribution around the 9° 37' OSC are consistent with the conclusions of Reynolds et al. [1992] regarding E-MORB occurrence. It is proposed that E-MORB identified on the Pacific plate near the 9° 37' OSC relate to the northward propagation of the western OSC limb and subsequent magmatic waning of activity along that limb as the eastern limb advanced southward. 3.6. Conclusions Submersible studies and wax core sampling in the region of a ridge-axis discontinuity at 9° 37. 1' N on the EPR has led to the identification of a third-order OSC that marks a divide between axial magmatic and hydrothermal systems. These investigations make, for the first time, a direct link between magmatic activity and OSC propagation. Off-axis fault traces from northward propagation of this axial discontinuity suggest that it has been in existence for on the order of lO*^ years, further supporting its designation as a third-order discontinuity. Changes in lava and hydrothermal fluid

    PAGE 152

    145 chemistry, ridge structure, volcanic morphology and apparent lava age all suggest that the eastern OSC limb and the ridge-axis north of it (Segments B and C of Haymon et al. [1991]) reflect a magmatic system that is currently more robust the that of the western OSC limb and ridge south of it (Segments D, E, and part of F). Geologic, chemical, biologic, and hydrologic data all suggest that a shallow intrusive event between 1989 and 1991, and likely a volcanic eruption, affected the eastern limb of the OSC. Chemical similarities between fresh sheet flows in the floor of the eastern limb's ASCT and the 1991 lava flow documented at 9° 46-51' further suggest that recent volcanic activity at 9° 37' may be related to the same magmatic pulse as the 9° 50' eruption. A widening and southward extension of the ASCT on the eastern OSC limb by more than 400 m between 1989 and 1991 indicates that the eastern limb is actively propagating southward in response to this recent magmatic activity. The presence of extinct and brine venting sulfide structures in the ASCT of the western limb, and its generally older appearance further suggest that southward advancement of the eastern OSC limb may be at the expense of the western limb magmatic activity. Southward advancement of the eastern OSC limb further suggests that the fourth-order Segment "D" of Haymon et al. [1991] is likely to disappear as Segment "C" advances southward. It is further suggested that the DEVAL at 9° 35' which defines the southern terminus of segment D may be an artifact of its magmatic waning. The narrowness and overall character of the ASCT on the southern portion of Segment D suggests that despite the old appearance of lavas flooring it, the ASCT is not in the late stages of the proposed ASCT developmental cycle [Fornari et al., 1998b]. This suggests that magmatic plumbing along the western limb of the OSC may not have had the opportunity to fully develop into a

    PAGE 153

    146 magmatically robust and mature system. Thus, Segment D may represent a transient fourth-order segment formed in response to propagation of a longer-lived axial discontinuity. Regional wax core sampling about the crestal plateau in the vicinity of the OSC has identified a concentration of E-MORB on the Pacific plate within ~ 1 km of the ridge-axis. The asymmetrical distribution of E-MORB in preference of the Pacific plate, their proximity to the OSC, and their unusual proximity to the ridge-axis (relative to other E-MORB recovered from the 9°-10° N segment) all suggest that E-MORB in this region are related to the western OSC limb. It is proposed that E-MORB are erupted in this region because the less robust magmatic system underneath the western OSC limb discourages efficient mixing and melt homogenization within the melt lens relative to regions with a robust, well developed and frequently replenished AMC.

    PAGE 154

    CHAPTER 4 CONCLUSIONS The detailed multi-disciplinary investigation of two recently active mid-ocean ridge segments in the eastern Pacific has resulted in a more detailed understanding of the nature of crustal accretion along these two ridge segments. The use of submersible and other deep water near-bottom systems, in conjunction with rock sampling by wax core, which provides better spatial control than traditional rock dredging, has resulted in a database of precisely located in situ geologic observations and samples. The combination of dense sampling, fine-scale spatial resolution of data, and temporal control introduced by the recent eruptive activity and longer term monitoring efforts make these investigations unique in MOR research. In particular, the addition of temporal control to the interpretation of MOR volcanic and hydrothermal activity on these two ridge segments has allowed for a more detailed account of the spatial and temporal associations between ridge structure, magmatic activity and geologic processes associated with oceanic crustal accretion within the neovolcanic zone of medium to fast spreading ridges. Investigation of the CoAxial Segment on the Juan de Fuca Ridge, which is known to have erupted in 1993, showed that ridge segment has been volcanically active on at least three occasions between 1981 and 1993. Inter-flow chemical variations indicate that the three eruptive units do not have a direct genetic link with one another, and suggest the CoAxial Segment does not have a well-developed and connected magmatic plumbing system. Seismic monitoring of the 1993 eruptive event suggested that the lava 147

    PAGE 155

    148 flow was emplaced after the intrusive dike had propagated laterally through the oceanic crust for lO's of kilometers, and inter-flow variations in lava chemistry suggest that lateral dike propagation may be a common emplacement mechanism on this segment. The structure of the segment is characteristic of a terrain where amagmatic extension plays a significant role in development of the ridge-axis. Volcanism within the neovolcanic zone of this segment does not appear to be particularly spatially focussed, with the three recent eruptions being emplaced over a kilometer-wide zone. Nonetheless, the eruption of some recent lava flows over preexisting, relatively fresh looking constructional volcanism suggests there is some structural control on the emplacement of lavas within the neovolcanic zone. The fissured character and poor focussing of axial volcanism are suggestive of a magmatically starved environment, yet there have been three eruptions in the last twelve years. This apparent contradiction may be indicative that magmatic activity on this medium spreading rate ridge segment is episodic, having short periods of relatively intense magmatic activity followed by longer periods of magmatic quiescence. Extreme isotopic and elemental depletion of CoAxial segment lavas relative to other lavas from the Juan de Fuca ridge may indicate an origin for this ridge segment that is unique to the ridge. A regional survey of different morphological/structural/volcanic provinces within the central Juan de Fuca Ridge shows each to be fed by distinct mantle melt regimes. Generally, geochemical variation observed between the different melt regimes can be modeled by mixing of mantle source characteristics identified in regionally identified basaltic lavas. Correlation between long lived isotopes and trace element characteristics in central Juan de Fuca lavas suggest that at least some

    PAGE 156

    149 component of regional upper mantle heterogeneity has antiquity to it, although trace element variability not correlated to isotopic ratios requires at least a third component of mantle heterogeneity. Close spatial association of ridge-axis lavas with distinct mantle geochemical signatures suggests that the spatial scale of heterogeneity may be quite small relative to the melt regimes identified. Close spatial association of differently geochemically enriched/depleted lavas also occurs on the East Pacific Rise, but the relationship between ridge-axis structure and magmatic activity are very different that what is observed at the CoAxial Segment. Relative the medium spreading, apparently magmatically starved CoAxial Segment, the fast spreading 9-10° N segment is magmatically robust, although spatial variability in melt supply to the ridge-axis does appear to exist. Data from a third-order overlapping spreading center (OSC) located at 9° 37' N suggest that there is a close association between magmatic activity and axial discontinuities, and further suggest that this OSC is a magmatic and hydrothermal divide between adjoining third-order ridge segments. The ridge-axis north of the OSC is currently in a robust phase of magmatic activity, while the ridge south of the OSC seems to be magmatically waning. North of the OSC, ridge-axis morphology is indicative of overall magmatic control, and south of the OSC structures related to amagmatic extension become more prominent. Temporal constraints that correlate a recent magmatic event to southward propagation of the eastern OSC limb establish, for the first time, a direct causal relationship between magmatic activity and evolution of ridge-axis discontinuities. The fourth-order segment associated with the OSC is likely a short lived, transitional feature that has not established a well-developed magmatic plumbing system. Support for this hypothesis includes a relatively poorly

    PAGE 157

    150 developed ASCT floored with old lavas and an abundance of enriched MORE associated with the apparently dying western limb of the OSC. This data confirms that segmentation of melt delivery to the ridge-axis can be a significant factor in the evolution of ridge-axis structure along the fast spreading EPR.

    PAGE 158

    APPENDIX FREQUENTLY USED ACRONYMS AND DESCRIPTIONS OF SAMPLING TECHNIQUES Frequently Used Acronyms and abbreviations JVHJKmiQ-occan nuge MORBmid-ocean ridge basalt JdFRJuan de Fuca Ridge EPREast Pacific Rise OSCoyerlapping spreading center ASCTaxial summit collapse trough ROVremotely operated vehicle DEVALdeviation from axial linearity REErare earth elements Sampling Techniques Methods used to recover lava samples from the seafloor include in situ sampling by the manned submersible Alvin, remotely operated vehicle (ROV) and autonomous underwater vehicle (AUV). Precise location of sample acquisition in each of these cases is accomplished by means of a long-baseline transponder net which provides a precisional accuracy to approximately 10 meters. In situ sampling by means of a deep submergence vehicle (submersible, ROV or AUV) provides the most accurate and informative means of sampling the deep seafloor. Other methods of lava sample recovery include a variety of surface deployed sampling techniques. These include wax coring (sometimes referred to as rock coring), rock dredging, and video guided grabs. Wax coring: This technique utilizes a lead weight with a barrel affixed on its underside. This barrel is equipped with a "head" coated with wax. The apparatus is lowered to the seafloor at between 70 and 100 meters per minute, and upon contact with unsedimented 151

    PAGE 159

    152 volcanic terrain pieces of volcanic glass become embedded in the wax. Typical recovery is between a few hundred milligrams and a few grams of volcanic glass. Precisional accuracy of sample location is dependent on depth, but typically is better than a few hundred meters. Rock Dredging : Rock dredges lowered to the seafloor and dragged for variable distances thereby recovering rock fragments from lava flows on the seafloor. Sample recovery and precisional accuracy of sample recovery are highly variable and determined by surface ship location and setback calculations determined by the water depth and amount of cable spooled out. Video guided grabs : This technique utilizes a hydraulic scoop with a real time video signal observed from a surface ship. This technique gives the precisional benefits of wax coring and excellent sample recovery. Additionally, the live video signal allows for in situ observations of the geologic associations of the sample recovered. 151

    PAGE 160

    REFERENCES Allan, J. F., R. Batiza, M. R. Perfit, D. J. Fomari, and R.O. Sack, Petrology of lavas from the Lamont seamount chain and adjacent East Pacific Rise, 10°N, J. Petrol, 30, 1245-1298, 1989. Arbogast, B. F., Quality assurance manual for the Branch of Geochemistry, U.S. Geological Survey, United States Department of the Interior Geological Survey, Reston, VA, 1990. Baker, E. T., G. J. Massoth, R. A. Feely, R. W. Embley, R. E. Thomson, and B. J. Burd, Hydrothermal event plumes from the CoAxial seafloor eruption site, Juan de Fuca Ridge, Geophys. Res. Lett., 22(2), 147-150, 1995. Baker, E. T. and J. E. Lupton, Changes in submarine hydrothermal 3He/heat ratios as an indicator of magmatic/tectonic activity. Nature, 346, 556-558, 1990. Batiza, R.Magmatic segmentation of mid-ocean ridges: a review, in Tectonic, Magmatic, Hydrothermal and Biological Segmentation of Mid-Ocean Ridges, edited by C. J. MacLeod, P. A. Tyler, and C. L. Walker, pp. 28, The Geological Society, London, 1996. Batiza, R., and Y. Niu, Petrology and magma chamber processes at the East Pacific Rise 9°30'N, J. Geophys. Res., 97, 6779-6797, 1992. Batiza, R., Y. Niu, J. L. Kartsten, W. Boger, E. Potts, L. Norby, and R. Butler, Steady and non-steady state magma chambers below the East Pacific Rise, Geophys. Res. Lett., 23, 221-224, 1996. Bischoff, J. L. and F. W. Dickson. Seawater-basalt interaction at 200°C and 500 bars: Implications for origin of seafloor heavy-metal deposits and regulation of seawater chemistry, Earth Planet. Sci. Lett., 25, 385-397, 1975 Bryan, W. B., G. Thompson, F. A. Frey, and J. S. Dickey, Inferred settings and differentiation in basalts from the Deep Sea Drilling Project, J. Geophys. Res., 8], 4285-4304, 1976. Butterfield, D. A. and G. J. Massoth, Geochemistry of North Cleft segment vent fluids: temporal changes in chlorinity and their possible relation to recent volcanism, J. Geophys. Res., 99, 4951-4968, 1994. 153

    PAGE 161

    154 Butterfield, D. A., G. J. Massoth, R. E. McDuff, J. E. Lupton and M. D. Lilley, Geochemistry of hydrothermal fluids from Axial Seamount Hydrothermal Emissions Study Vent Fluid, Juan de Fuca Ridge: Subseafloor boiling and subsequent fluid-rock interaction, y. Geophys. Res., 95, 12895-12921, 1990. Carbotte, S. M., and K. C. Macdonald, East Pacific Rise 8°-10°30'N: Evolution of ridge segments and discontinuities from SeaMARC II and three-dimensional magnetic studies, J. Geophys. Res., 97(3569), 6959-6982, 1992. Carbotte, S. M., A. Solomon, and G. Ponce-Correa, Evaluation of morphological indicators of magma supply and segmentation from a seismic reflection study of the EPR 15°30'-17°N, J. Geophys. Res., in press. Chadwick, W. W., Jr., and R. W. Embley, Graben formation associated with recent dike intrusions and volcanic eruptions on the mid-ocean ridge, J. Geophys. Res., 70i(B5), 9807-9825, 1998. Chadwick, W. W., Jr., R. W. Embley, and C. G. Fox, SeaBeam depth changes associated with recent lava flows, CoAxial segment, Juan de Fuca Ridge: Evidence for multiple eruptions between 1981-1993, Geophys. Res. Lett., 22(2), 167-170, 1995. Christeson, G. L., G. M. Kent, G. M. Purdy, and R. S. Detrick, Extrusive thickness variability at the East Pacific Rise, 9°-10°N: Constraints from seismic techniques, J. Geophys. Res., 707(82), 2859-2873, 1996. Christeson, G. L., G. M. Purdy, and G. J. Fryer, Seismic constraints on shallow crustal emplacement processes at the fast spreading East Pacific Rise, J. Geophys. Res., 99, 17957-17973, 1994. Christeson, G. L., P. R. Shaw, and J. D. Garmany, Shear and compressional wave structure of the East Pacific Rise, 9°-10°N, J. Geophys. Res., 702(B4), 7821-7835, 1997. Cochran, R. J., D. J. Fomari, B. J. Coakley, R. Herr, and M. A. Tivey, Continuous nearbottom gravity measurements made with a BGM-3 gravitmeter in DSV Alivn on the East Pacific Rise crest near 9°31'N and 9°50'N, J. Geophys. Res., ]04{B5), 1,841-10,861, 1999. Crawford, W. C, Determination of oceanic crustal shear velocity structure from seafloor compliance measurements, PhD thesis, Univ. of Calif., San Diego, 1994. Crawford, W. C, S. C. Webb, and J. A. Hildebrand, Constraints on melt in the lower crust and Moho at the East Pacific Rise, 9°48'N, using seafloor compliance measurements, J. Geophys. Res., 104(32), 2923-2939, 1999.

    PAGE 162

    155 Delaney, J. R., H. P. Johnson, and J. L. Karsten, The Juan de Fuca Ridge--Hot SpotPropagating Rift System: New Tectonic, Geochemical, and Magnetic Data. J. Geophys. Res.,86{B\2), 11,747-11,750, 1981. Desonie, D. L., and R. A. Duncan, The Cobb-Eickelberg Seamount chain: Hotspot volcanism with mid-ocean ridge basalt affinity, J. Geophys. Res., 95(B8), 12,69712,711, 1990. Detrick, R. S., P. Buhl, E. Vera, J. Mutter, J. Orcutt, J. Madsen, and T. Brocher, Multichannel seismic imaging of a crustal magma chamber along the East Pacific Rise, Nature, 326, 35-41, 1987. Dunn, R. A., and D. R. Toomey, Seismological evidence for three-dimensional melt migration beneath the East Pacific Rise, Nature, 388, 259-262, 1997. Dziak, R. P., C. G. Fox, and A. E. Schreiner, The June-July 1993 seismo-acoustic event at CoAxial segment, Juan de Fuca Ridge: Evidence for a lateral dike injection, Geophys. Res. Lett., 22(2), 135-138, 1995. Eaby, J., D. A. Clague, and J. R. Delaney, Sr isotope variations along the Juan de Fuca Ridge, J. Geophys. Res., 89 (B9), 7,883-7,890, 1984. Embley, R. W., D. Butterfield, K. Roe, M. Stapp, W. W. Chadwick, L. Evans, S. Merle, J. Baross, J. Kaye, J. Huber, G. Massoth, J. Gendron, S. Maenner, V. Tunnicliffe, M. Tsurimi, J. Marcus, K. Juniper, D. Grelon, C. Levasque, C. Moyer, K. Pelletreau, G. Wheat, S. Scott, J. Chadwick, M. Perfit, E. Guenther, B. McLaughlin-West, J. Getsiv, and E. Williamson, Time series investigations of an active volcano: The NeMO project (abstract), Eos Trans AGU, 79(45), 921, 1998. Embley, R. W., and W. W. Chadwick, Jr., Volcanic and hydrothermal processes associated with a recent phase of seafloor spreading at the northern Cleft segment: Juan de Fuca Ridge, J. Geophys. Res., 99, 4741-4760, 1994. Embley, R. W., W. W. Chadwick, Jr., I. R. Jonasson, D. A. Butterfield, and E. T. Baker, Initial results of the rapid response to the 1993 CoAxial event: Relationships between hydrothermal and volcanic processes, Geophys. Res. Lett., 22(2), 143-146, 1995. Embley, R. W., W. W. Chadwick, Jr., M. R. Perfit, and E. T. Baker, Geology of the northern Cleft segment, Juan de Fuca Ridge: Recent lava flows, sea-floor spreading, and the formation of megaplumes. Geology, 19, 11\-115, 1991. Embley, R. W., W. W. Chadwick, Jr., M. R. Perfit, M. C. Smith, and J. R. Delaney, Recent eruptions on the coaxial segment of the Juan de Fuca Ridge: Implications for mid-ocean ridge accretion processes, J. Geophys. Res., in press.

    PAGE 163

    156 Finney, A. R., Petrology and geochemistry of the Vance samounts: Near-ridge seamounts along the Juan de Fuca Ridge, Undergraduate Honors thesis, 115 p., Univ. of Fla., Gainesville, Fl, 1989. Fomari, D. J., R. M. Haymon, M. R. Perfit, T. K. P. Gregg, and M. H. Edwards, Axial summit trough of the East Pacific Rise 9°N to 10°N: Geological characteristics and evolution of the axial zone on fast-spreading mid-ocean ridges, J. Geophys. Res., ;0i(B5), 9827-9855, 1998b. Fomari, D. J., M. R. Perfit, J. F. Allan, R. Batiza, R. Haymon, A. Barone, W. B. F. Ryan. T. Smith, T. Simkin, and M. A. Luckman, Geochemical and structural studies of the Lamont Seamounts: Seamounts as indicators of mantle processes. Earth Planet. Sci. Lett., 89, 63-83, 1988. Fornari, D. J., T. Shank, K. L. Van Damm, T. K. P. Gregg, M. Lilley, G. Levai, A. Bray, R. M. Haymon, M. R. Perfit, and R. Lutz, Time-series temperature measurements at high-temperature hydrothermal vents. East Pacific Rise 9°49'-5rN: evidence for monitoring a crustal cracking event. Earth Planet. Sci. Lett., 160, 419-431, 1998a. Fox, C. G., W. E. Radford, R. P. Dziak, T.K. Lau, H. Matsumoto, and A. E. Schreiner, Acoustic detection of a seafloor spreading episode on the Juan de Fuca Ridge using military hydrophone arrays, Geophys. Res. Lett., 22(2), 131-134, 1995. Gregg, T. K. P., D. J. Fomari, M. R. Perfit, R. M. Haymon, and J. H. Fink, Rapid emplacement of a mid-ocean ridge lava flow on the East Pacific Rise at 9° 46'Sl'N, Earth Planet. Sci. Lett., 144, E1-E7, 1996. Grove, T. L., R. J. Kinzler, and W. B. Bryan, Fractionation of Mid-Ocean Ridge Basalt (MORB), in Mantle Flow and Melt Generation at Mid-Ocean Ridges, edited by J. Phipps Morgan, D.K. Blackman, and J.M. Sinton, pp. 281-310, AGU, Washington, 1992. Harding, A. J., G. M. Kent, and J. A. Orcutt, A multichannel seismic investigation of upper crastal stmcture at 9°N on the East Pacific Rise: Implications for cmstal accretion, / Geophys. Res., 9
    PAGE 164

    157 submersible observations of seafloor phenomena associated with an eruption event in April, \99\, Earth Planet. Sci. Lett., 119, 85-101, 1993. Hegner, E., and M. Tatsumoto, Pb, Sr and Nd isotopes in basalts and sulfides from the Juan de Fuca Ridge, J. Geophys. Res., 92, 1 1380-11386, 1987. Hey, R. N., F. K. Duennebier, and W. J. Morgan, Propagating rifts on mid-ocean ridges, J. Geophys. Res., 55(1371), 3647-3658, 1980. Hey, R. H., H. W. Menard, T. M. Atwater, and D. W. Caress, Changes in direction of seafloor spreading revisited, J. Geophys. Res., 95(3388), 2803-281 1, 1988. Hey, R. N., and D. S. Wilson, Propagating rift explanation for the tectonic evolution of the northeast Pacific the pseudomovie. Earth Planet. Sci. Lett., 58, 167-188, 1982. Holmes, M. L., and H. P. Johnson, Upper crustal densities dervived from bottom gravity measurements: Northern Juan de Fuca Ridge, Geophys. Res. Lett., 20, 1871-1874, 1993. Hooft, E. E. E., H. Schouten, and R. S. Detrick, Constraining crustal emplacement processes from the variation of seismic layer 2A thickness at thee East Pacific Rise, Earth Planet. Sci. Lett., 142, 289-310, 1996. Ito, E., W. M. White, and C. Gopel, The O, Sr, Nd and Pb isotope geochemistry of MORE, Chemical Geology, 62, 157-176, 1987. Johnson, P. H., and M. L. Holmes, Evolution in plate tectonics; The Juan de Fuca Ridge, in The Geology of North America, edited by E. L. Winterer, D. M. Hussong, and R. W. Decker, pp. 19, Geological Society of America, Boulder, Colo., 1989. Kappel, E. S., and W. R. Normark, Morphometric variability within the axial zone of the southern Juan de Fuca Ridge: Interpretation from SeaMARC II, SeaMARK I, and deep sea photography, J. Geophys. Res., 92(B1 1), 1 1,291-1 1,302, 1987. Kappel, E. S., and W. B. F. Ryan, Volcanic episodicity and a non-steady state rift valley along northeast Pacific spreading centers: Evidence from SeaMARC I, J. Geophys. Res., 97(B14), 13,925-13,940, 1986. Karsten, J. L., and J. R. Delaney, Hot spot-ridge crest convergence in the northeast Pacific, J. Geophys. Res., 94(B\), 700-712, 1989. Karsten, J. L., J. R. Delaney, J. M. Rhodes, and R. A. Liias, Spatial and temporal evolution of magmatic systems beneath the Endeavour Segment, Juan de Fuca Ridge; tectonic and petrologic constraints, J. Geophys. Res., 95(B12), 19,23519,256, 1990.

    PAGE 165

    158 Kent, G. M., A. J. Harding, and J. A. Orcutt, Distribution of magma beneath the East Pacific Rise near the 9°03'N overlapping spreading center from forward modeling of common depth point data, J. Geophys. Res., 95(B8), 13,971-13,995, 1993a. Kent, G. M., A. J. Harding, and J. A. Orcutt, Distribution of magma beneath the East Pacific Rise between the Clipperton Transform and the 9° 1 7'N deval from forward modeling of common depth point data, J. Geophys. Res., 9S(B8), 13945-13969, 1993b. Klein, E. M., and C. H. Langmuir, Global correlations of ocean ridge basalt chemistry with axial depth and crystal thickness, J. Geophys. Res., 92, 8089-81 15, 1987. Klein, E. M., and C. H. Langmuir, Local versus global variations in ocean ridge basalt composition: A reply, J. Geophys. Res., 94{BA), 4241-4252, 1989. Klitgord, K., and J. Mammerickx, Northern East Pacific Rise: Magnetic anomaly and bathymetric framework, J. Geophys. Res., 87(1352), 6725-6750, 1982. Langmuir, C. H., Geochemical consequences of in situ crystallization. Nature, 340, 199205, 1989. Langmuir, C. H., J. F. Bender, and R. Batiza, Petrologic and tectonic segmentation of the East Pacific Rise, 5° 30-14° 30' N, Nature, 322, 422-429, 1986. Langmuir, C. H., E. M. Klein, and T. Plank, Petrological systematics of mid-ocean ridge basalts: Constraints on melt generation beneath ocean ridges, in Mantle flow and melt generation at mid-ocean ridges, Geophysical Monograph Series, edited by J. Phipps-Morgan, D. K. Blackman, J. M. Sinton, pp. 183-280, AGU, Washington, D.C., 1992. Langmuir, C. H., R. D. J. Vocke, and G. N. Hanson, A general mixing equation with applications to Icelandic basalts. Earth Planet. Sci. Lett., 37, 380-392, 1978. Leyboume, M. I., and N. A. Van Wagoner, Heck and Heckle seamounts, northeast Pacific Ocean: High extrusion rates of primitive and highly depleted mid-ocean ridge basalt on off-ridge seamounts, J. Geophys. Res., 96(B10), 16,275-16,293, 1991. Lonsdale, P., Overlapping rift zones at the 5.5°S offset of the East Pacific Rise, J. Geophys. Res., 88, 9393-9406, 1983. Lonsdale, P., Comments on "East Pacific Rise from Siqueirios to Orozco Fracture Zones: Along-strike Continuity of Axial Neovolcanic Zone and Structure and Evolution of Overlapping Spreading Center" by K. Macdonald, J. Sempere and P. J. Fox., J. Geophys. Res., 97(B10), 10493-10499, 1986.

    PAGE 166

    159 Lupton, J. E., Water column hydrothermal plumes on the Juan de Fuca Ridge. J. Geophys. Res., 95(B8), 12,829-12,842, 1990. Luyendyk, B. P., On-bottom gravity profile across the East Pacific rise crest at 2\°N, Geophysics, 49(2428), 2166-2177, 1984. Macdonald, K. C, and P. J. Fox, Overlapping spreading centers: New accretion geometry on the East Pacific rise, Nature, 307(5903), 55-58, 1983. Macdonald, K. C, and P. J. Fox, The axial summit graben and cross-sectional shape of the East Pacific Rise and indicators of axial magma chambers and recent volcanic eruptions. Earth Planet. Sci. Lett., 88, 1 19-131, 1988. Macdonald, K. C, and J. C. Sempere, Reply: The debate concerning overlapping spreading centers and mid-ocean ridge processes, J. Geophys. Res., 97(B10), 10501-10511, 1986. Macdonald, K. C, P. J. Fox, S. Carbotte, M. Eisen, S. Miller, L. Perram, D. Scheirer, S. Tighe, and C. Weiland, The East Pacific Rise and its flanks 8 18°N: History of segmentation, propagation, and spreading direction based on SeaMarc 11 and Sea Beam studies. Mar. Geophys. Res., 74(1255), 299-344, 1992. Macdonald, K. C, P. J. Fox, L. J. Perram, M. F. Erson, R. M. Miller, S. P. Carbotte, M. H. Cormeer, and A. N. Shor, A new view of the mid ocean ridge from the behavior of ridge axis discontinuities. Nature, 335(6187), 217-225, 1988. Macdonald, K., J. C. Sempere, and P. J. Fox, East Pacific rise from Siqueiros to Orozco fracture zones: Along-strike continuity of axial neovolcanic zone and structure and evolution of overlapping spreading centers, J. Geophys. Res., S9(B7), 6049-6069, 1984. Macdonald, K. C, J. C. Sempere, P. J. Fox, and R. Tyle, Tectonic evolution of ridge-axis discontinuities by the meeting, linking, or self-decapitation of neighboring ridge segments. Geology, 75(November), 993-997, 1987. Michael, P. J., and R. L. Chase, The influence of primary magma composition, H2O and pressure on mid-ocean ridge basalt differentiation, Contrib. Mineral. Petrol, 96, 245-263, 1987. Mutter, J. C, G. A. Barth, P. Buhl, R. S. Detrick. J. A. Orcutt, and A. J. Harding, Magma distribution across ridge-axis discontinuities on the East Pacific Rise from multichannel seismic images, Nature, 336, 156-158, 1988. Niu, Y., K. D. Collerson, R. Batiza, J. I. Wendt. and M. Regelous, Origin of enrichedtyoe mid-ocean ridge basalt at riges far from mantle plumes: The East Pacific Rise at ir20'N,y. Geophys. Res., ]04(B4), 7067-7087, 1999.

    PAGE 167

    160 Niu, Y., D. G. Waggoner, J. M. Sinton, and J. J. Mahoney, Mantle source heterogeneity and melting processes beneath seafloor spreading centers: The East Pacific Rise, 18° 19°S, J. Geophys. Res., 101, 27,711-27,733, 1996. Perfit, M. R., and W. W. Chadwick, Jr., Magmatism at mid-ocean ridges: Constraints from volcanological and geochemical investigations, in Faulting and Magmatism at Mid-Ocean Ridges, edited by W. R. Buck, P. Delaney, J. A. Karson, and Y. Lababrille, pp. 57, AGU, Washington, DC, 1998. Perfit, M. R., D. J. Fomari, M. C. Smith, J. F. Bender, C. H. Langmuir, and R. M. Haymon, Small-scale spatial and temporal variations in mid-ocean ridge crest magmatic processes. Geology, 22, 375-379, 1994. Perfit, M. R., D. J. Fomari, M. C. Smith, C.H. Langmuir, J. F. Bender, and R. M. Haymon, Fine-scale petrological variations along the East Pacific Rise crest 9° 1 7'N to 9°54'N: results from ALVIN diving and rock coring during the Adventure program (abstract), EOS Trans AGU, 72, 491, 1991. Perfit, M. R., M. C. Smith, J. Chadwick, R. W. Embley, W. W. Chadwick, III, I. R. Jonasson, and P. Herzig, Petrologic and geochemical investigations of recent lavas erupted from axial seamount and adjacent rift zones: New insights (abstract), Eos Trans AGU, 79(45), 921, 1998. Perram, L. J., and K. C. Macdonald, A one-million year history of the 11°45'N East Pacific Rise discontinuity, J. Geophys. Res., 95(3536), 21363-21381, 1990. Reynolds, J. R., Segment-scale systematic of mid-ocean ridge magmatism and geochemistry, PhD thesis, 483 p., Columbia University, Palisades, NY, 1995. Reynolds, J. R., C. H. Langmuir, J. F. Bender, K. A. Kastens, and W. B. F. Ryan, Spatial and temporal variability in the geochemistry of basalts from the East Pacific Rise, Nature, 359, 493-499, 1992. Rhodes, J. M., C. Morgan, and R. A. Liias, Geochemistry of axial seamount lavas: Magmatic rlationship btween the Cobb htspot and the Juan de Fuca Ridge, J. Geophys. Res., 9J(B8), 12,713-12,733, 1990. Ridley, W. I. and F. E. Lichte, Major, trace, and ultra trace element analyses by laser ablation ICP-MS, /?ev. Econ. GeoL, 7, 199-215, 1998. Rubin, K. H., J. D. Macdougall, and M. R. Perfit, ^'°Po-^'"Pb dating of recent volcanic eruptions on the seafloor. Nature, 368, 841-844, 1994.

    PAGE 168

    161 Rubin, K. H., M. C. Smith, M. R. Perfit, D. M. Christie, and L. F. Saclcs, Geochronology and petrology of lavas from the 1996 North Gorda Ridge eruption, Deep-Sea Research Part II, 45, 2571-2597, 1998. Schilling, J. G., M. Zajac, R. Evans, T. Johnston, W. White, J. D. Devine, and R. Kingsley, Petrologic and geochemical cariations along the Mid-Atlantic Ridge from 27°N to 73°N, Am. J. Sci., 283, 510-86, 1983. Schouten, H., M. A. Tivey, D. J. Fomari, and J. R. Cochran, Central anomaly magnetization high: Constraints on the volcanic construction and architecture of seismic layer 2A at a fast-spreading mid-ocean ridge, the EPR at 9°30'-50'N, Earth Planet. Sci. Lett., 169, 37-50, 1999. Sempere, J. C, and K. C. Macdonald, Deep-tow studies of the overlapping spreading centers at 9°03'N on the East Pacific Rise, Tectonics, 5(1253), 881-900, 1986. Sempere, J. C, K. C. Macdonald, and S. P. Miller, Overlapping spreading centers: 3-D inversion of the magnetic field at 9°03'N on the East Pacific Rise, Geophys. J. R. Astron. Soc. , 79{ 1 649), 799-8 1 1 , 1 984. Shaw, D. M., Trace element fractionation during anatexis, Geochim. Cosmochim. Acta, 34, 237-243, 1970. Sims, K. W. W., S. J. Goldstein, D. J. Fomari, M. R. Perfit, M. T. Murrell, and S. R. Hart, U-series analyses of young lavas from 9°-10° N East Pacific Rise: Constraints on magma transport and storage times benearth the ridge axis (abstract), Eos Trans AGU, 78, F792, 1997a. Sims, K. W. W., S. J. Goldstein, M. R. Perfit, D. J. Fornari, S. R. Hart, M. T. Murrell, and M. C. Smith, Spatial and temporal variations in mid-ocean ridge basalt geochemistry along the East Pacific Rise: U-series disequilibria, radiogenic isotopes and elemental variability in a closely spaced, two-dimensional grid of samples from 9°48'-52'N., in Seventh Annual Goldschmidt Conf., LPI 921, pp. 194, Lunar and Planetary Institute, Houston, Texas, 1997b. Sinton, J. M., Spreading rate dependence in volcanic eruptions on mid-ocean ridges: Speculations from observations, reservoir constraints and ridge structure (abstract), Eos Trans AGU, 75(46), 663, 1997. Sinton, J. M., S. M. Smaglik, J. J. Mahoney, and K. C. Macdonald, Magmatic processes at superfast spreading mid-ocean ridges: Glass compositional variations along the East Pacific Rise 13° 23° S, J. Geophys. Res., 96(4), 6133-6155, 1991. Sinton, J. M., D. S. Wilson, D. M. Christie, R. N. Hey, and J. R. Delaney, Petrologic consequences of rift propagation on oceanic spreading ridges. Earth Planet. Sci. Lett., 62, 193-207, 1983.

    PAGE 169

    162 Smith, M. C, M. R. Perfit, R. W. Embley, and W. W. Chadwick, Interpretations of magmatic activity and crustal accretion along the CoAxial Segment and Axial Seamount North Rift Zone: Using combined acoustic and geochemical data to map the seafloor at the scale of individual flows (abstract), Eos Trans AGU, 78(46), 677, 1997. Smith, M. C, M. R. Perfit, D. J. Fomari, and K. L. Von Damm, The 9°37'N DEVAL: Evidence for southward propogration of a small OSC driven by recent volcanic activity (abstract), Eos Trans AGU, 79(45), 838, 1998. Smith, M. C, M. R. Perfit, and I. R. Jonasson, Petrology and geochemistry of basalts from the southern Juan de Fuca Ridge: Controls on the spatial and temporal evolution of mid-ocean ridge basalt, J. Geophys. Res., 99, 4787-4812, 1994. Sohn, R. A., J. A. Hildebrand, and S. C. Webb, Postrifting seismicity and a model for the 1993 diking event on the CoAxial segment, Juan de Fuca Ridge, J. Geophys. Res., 103{B5), 9867-9877, 1998. Sohn, R.A., S.C. Webb, J. A. Hildebrand, and Domuelle, Three-dimensional tomographic velocity structure of upper-crust, CoAxial segment, Juan de Fuca Ridge: Implications for on-axis evolution and hydrothermal circulation, J. Geophys. Res., 702(B8), 17,679-17,695, 1997. Stevenson, J. M., J. A. Hildebrand, and M. A. Funberge, An ocean bottom gravity study of the southern Juan de Fuca Ridge, J. Geophys. Res., 99, 4875-4888, 1994. Sun and McDonough, 89 Sun, S. S., R. W. Nesbitt, and A. Y. Sharaskin, Geochemical characteristics of mid-ocean ridge basalts, Earth Planet. Sci. Lett., 44, 1 19-38, 1979. Taylor, S. R., Planetary Science: A Lunar Perspective, Lunar and Planetary Institute, Houston, 1982. Thompson, G., and D. C. Bankston, Sample contamination from grinding and sieving determined by emission spectrometry. Applied Spectroscopy, 24(2), 210-219, 1970. Toomey, D. R., G. M. Purdy, S. C. Solomon, and W. S. D. Wilcock, The threedimensional seismic velocity structure of the East Pacific Rise near latitude 9°30'N, Nature, 347{\2A1), 639-645, 1990. Toomey, D. R., S. C. Solomon, and G. M. Purdy, Tomographic imaging of the shallow crustal structure of the East Pacific Rise at 9°30'N, J. Geophys. Res., 99, 24,13524,157, 1994.

    PAGE 170

    163 Vera, E. E., and J. B. Diebold, Seismic imaging of oceanic layer 2A between 9°30'N and 10°N on the East Pacific Rise from two-ship wide-aperture profiles, J. Geophys. Res., 99, 3031-3041, 1994. Von Damm, K. L., Controls on the chemistry and temporal variability of seafloor hydrothermal fluids, 222-247 pp., AGU, Washington, DC, 1995. Von Damm, K. L., Chemistry of hydrothermal vent fluids from 9-10°N, East Pacific Rise: "Time Zero" the immediate post-eruptive period, J. Geophys. Res., in press. Weaver, J. S., and C. H. Langmuir, Calculations of phase equilibrium in mineral-melt systems. Computers and Geosciences, 16, 1-19, 1990. Wright, D. J., R. M. Haymon, and D. J. Fomari, Crustal fissuring and its relationship to magmatic and hydrothermal processes on the East Pacific Rise crest (9° 12' to 54'N),y. Geophysical Res., 700(B4), 6097-6120, 1995.

    PAGE 171

    164 BIOGRAPHICAL SKETCH Matthew C. Smith attended the University of New Hampshire from the fall of 1984 through the spring of 1989, at which time he earned a bachelor of science degree in geology. He then began graduate studies in geology at the University of Florida in the fall of 1989. He earned a master of science degree in geology at the University of Florida in the summer of 1993.

    PAGE 172

    I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a thesis for the degree of Doctor of Philosophy. Michel R. Perfit, Chairman Proiessor of Geology I certify that I have read this smdy and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a thesis for the degree of Doctor of Philosophy. Paul A. Mueller Professor of Geology I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope ^d quality, as a thesis for the degree of Doctor of Philosophy. Ann HeatHerington Assistant in Geology I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate,^ s^jape and quality, aj^ a thesis^ for the degree of Doctor of Philosophy. NeirOpdyki Professor of Geology I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a thesis for the degree of Doctor of Philosophy. Vaneica Young A Associate Professor of Cher^ strv This dissertation was submitted to the Graduate Faculty of the Department of Geology in the College of Liberal Arts and Sciences and to the Graduate School and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. December 1999 Dean, Graduate School