6 août 2020 · Broad-band ocean-bottom seismometer (OBS) deployments present an opportunity to investi- gate the seafloor sediment thickness, which is
Each colored band represents a time span of 10 million years 1 Use the legend for the Detailed Ocean-Floor Age (Ma) theme to complete Table 1
Part Seven - The governance of the ocean 1 Oceans under control Philippe Cury, Sophie Arnaud-Haond and Françoise Gaill
17 juil 2013 · acoustics is termed the very-low-frequency (VLF) band For frequencies below 100 Hz in shallow water, mode representations are usually
3 mai 2022 · At void fractions below 10?6 5 the bubble size distribu- tions do not collapse to a narrow band when normalised by void fraction It appears
May-July 2017 www un org/africarenewal Global economic gravity pulling towards Africa Plastics pose biggest threat to oceans Life under water
6 juil 2011 · ming pools to the ocean, with playback unfolding above water or beneath Composers of underwater music are especially curious about
model (MICOM) to the ocean cavity beneath the Ross Ice Shelf to investigate the circulation of homogeneity and a series of alternating bands of northward
a Continent and Ocean the accumulation of convective heat beneath the thermally insulating layer of the The emergence of a light color band of a heated
measure the electrical conductivity beneath the mid ocean ridge the middle temperature range is thougiit to be moving holes in the valence band cliarig-
Citation: Sleep N H Small-scale convection beneath oceans and continents Chinese Sci The data in Figure 2 come from the latitude band in the box Most of
PDF document for free
- PDF document for free
![[PDF] ELECTRICAL PROPERTIES OF THE MANTLE UPWELLING ZONE](https://pdfprof.com/EN_PDFV2/Docs/PDF_8/28674_8NQ27967.pdf.jpg "[PDF] ELECTRICAL PROPERTIES OF THE MANTLE UPWELLING ZONE")
28674_8NQ27967.pdf
ELECTRICAL PROPERTIES OF THE MANTLE UPWELLING
ZONE BENEATH A MID OCEAN RIDGE, AN APPLICATION
OF VERTICAL GRADIENT SOUNDING
Marion D. Jegen
Geophysics Lnboratory
Departnent of Physics
University of Toronto
Toronto,
Canada
.A thesis submitted in conformity with the requirements for the degree of
Doctor of Philosophy
at the
University of Toronto
@Copyright by ihn'on D. Jegen. 1997
Nationaf Library
1*1 of Canada
Bibliothèque nationale
du Canada
Acquisitions and Acquisitions et
Bibliographie Services services bibliographiques
395 Wellington Sheet 395. rue Wellington
OttawaON K1AON4 ûttawaON KtAON4
Canada Canada
The author has granted a non-
exclusive licence dowing the
National Library of Canada to
reproduce, loan, distribute or sell copies of this thesis in microfom, paper or electronic formats.
The author retains ownership of the
copyright in this thesis. Neither the thesis nor substantial extracts fiom it may be printed or otherwise reproduced without the author's permission.
L'auteur a accordé une licence non
exclusive permettant a la
Bibliothèque nationale du Canada de
reproduire, prêter, distribuer ou vendre des copies de cette thèse sous la fome de microfiche/film, de reproduction sur papier ou sur format
électronique.
L'auteur
conserve la propriété du droit d'auteur qui protège cette thèse.
Ni la thèse ni des extraits substantiels
de celle-ci ne doivent être imprimés ou autrement reproduits sans son autorisation.
ELECTRICAL PROPERTIES OF THE MANTLE UPWELLING
ZONE BENEATH A MID OCEAN RIDGE? AN -4.PPLICA4T1ON
OF VERTICAL GRADIENT SOUNDING
Doctor of Philosophy
Marion D. Jegen
Department of Physics
University of Toronto. 199'7
Abstract
On mid-ocean ridges, as adjacent plates move apart. the mantle material rises to fil1 the void created. During its ascent the solidus of the material is crossed and melting occurs.
The melt itself
is eventually emplaced at the ridge axis producing new oceariic crust. The understanding of the flow of the solid and molten material is harnpered bu the lack of knowledge of vital mode1 parameters such as the connectivity of the partial melt.
C'onnectivity is
related to the permeability in the uprvelling region. It t herefore cont rois the migration pattern of the buoyant melt. the Rom of the solid phase material. and the mantle upwelling mechanism. Changes in the geometry of the distri bution of meit in t lie solid material have a large impact on the electrical conductivity. I have measurecl the conductivity of the upwelling region to constrain possible partial melt geomet ries.
1 present results of vertical gradient sounding (VGS) experirnents on the Encleavour
and Explorer ridge. which are part of the Juan de Fuca and its nort hern extension. the Explorer ridge, respectivel. The VGS method is a natural source Ehl methocl based ent irely on measurements of the magnet ic fields.
Electrical responses
of the 1D layered normal seafloor combined wi t h a 'LD region repre- senting the made upwelling zone and proposed upwelling mechanisms are derived. A cornparison of the synthetic response of a range of models with data measured on the Endeavour segment shows that the conductivity in the upwelling region is very high (in the order of 1 to 5 ohm m depending on the shape of the upwelling region). The results of this experiment suggest that the pore space containing the conductive melt is well connected. The melt must be able to move freely through the upwelling region. The experiments support so called melt migration models.
The data
measured on the Esplorer segment yielded a different conductivitj- model. The data do not require the presence of a pronounced 7D conductivity anomaly at depth and therefore precliides the existence of large or well connected melt fractions in an upwelling region underneath
Esplorer
ridge.
The data therefore support a stiidy of
magnetic anomaly maps that suggests that the Explorer ridge is degenerate, i.e. that the Explorer plate is slowly deforming internally and its mo~~ement is governed by the movement of the surrounding plates.
Acknowledgment s
My forernost thanks goes to Prof. R.N. Edwards for suggesting the topic and supervis- ing this work. Working with Nigel \vas a great pleasure and most of all. always very interesting. I am grateful for his guidance during this work and in geophysics as a whole. I would further like to thank Prof. R.C. Bailey and Prof. G.F. West who were always open for discussions. encouraging, helped me on various aspects of my work and showed a great deal of patience and care. kIany thanks also to my fellow student and post-docs on the quest for the three Lettres. especially (sorted by height ), Roderick Fisher. Milie
Kendall.
Graeme
Cairns. John Wayne,
Rob Evans and Liming Yu. for sharing the domns
and celebrating the ups. as well as for giving a hand if t hings got too heavp. I ivould also like to thank Paul Ruppert. the head of the Physics Electronics Resource Centre. for his support wit h the magnetometers and thereby success of the experiments and to
XIarianne
Khurana, secretary of the graduate chair. who made dealing with the administrative side of the university a pleasure. I am furthermore grateful to Patricia Edwards for quite a few fabulous dinners and conversations and to the Edwards farnily as a whole who alivays made me feel welcome and brightened up many holidays and weekends.
1 thorough1~-
enjoxecl my time spent at the phjesics department on a professional as well as on a persona1 levei. which
1 rnainly (but not only) owe to the above mentionecl people.
Outsicle
of the great city of Toronto I would like to thank Dr. P. Tarits at the (.'niver- site Occidentale in Brest. France. my external cornmittee member. for encouraging and interesting discussions. Dr. S.C. Webb of the SCIRPPS Institute in La Jolia. Ca. and Dr. L.K. Law at the PGC in Sidney. B.C. for support in the logistics of the experinients and wit h the eqiiipment. I am also indebted to Prof. S. Scott of University of Toronto (often reierred to as -Cap- tain Canada"). chief scientist on many cruises
1 have been on, and the crew and masters
of the CSS J.P. Tully. RV Melville, CSS Endeavour. 1 would furthermore like to mention Bruce Milliman and his family, who provided the space For the land magnetometer and turned the set up of the land station magnetometer into a fun holiday or two.
Finally
1 would like to express rny gratitude to Gabor Iiulcsar for his patience ancl
humour diiring t lie final stressful mont hs of completing this thesis and Cornelia Kappler. Francois Vial, Jutta Luettmer Strathman and Moira Daly for being such good frientls. -4s to my parents and farnily. there is no may I can express my gratitude or acknowledge the part you contributed to this work adequately by 1 want to thank you dearly for al1 the support. understanding and encouragement you gave me.
Contents
1 Mid Ocean Ridges 1
..................................
1.1 Introduction 1
.............................. 1 . 2 Mid Ocean Ridges 3 ...............................
1 . .3 Mantle Upwelling 12
...................... 1.3.1 Passive Llantle Upwelling 12 ......... 1.3.- Discussion of the Passive Mantle Ii~welling Mode1 15 ......... 1.3.3 Focused Melt Migration iblodel: Passive Upwelling 16 ......... 1.3 Focused Mantle Flow Model: Dynamic lipwelling 16 ......................
1.3.5 Summary and Conclusion 1S
2 Electrical Conductivity Mode1 of a Mid Ocean Ridge 21
............................
2 . I Layered Normal Mode1 21
-) . > ................................
1 Ocean Layer
..................................
2.12 Crut 33
..............................
2.1.3 Lithosphere 2-1
.............................
2.1.4 Asthenosphere 26
2.2 Electrical Conductivity of the Tivo Phase Material in the I'pwelling Region 27
......................
1.2.1 I-Iashin-Shtrikman Bounds 29
2.2.2 Films .................................. 29
'2.2.3 Tubes ................................. 30
2.2. .\rchie's Law ............................. 30
2.2.4 Pietwork Analogies .......................... 30
2.3 Electrical Conductivity Representation of a MOR ............. 32
3 The Juan de Fuca Mid Ocean Ridge 36
3.1 Tectonic History of the Juan de Fuca Ridge System ............ 96
............... 3.2 .L lorpliology of the Juan de Fuca Ridge System 39 .......................... 3.. Endeavour Segment 40 ..................... 3 Sout hern Explorer Segment 42
4 Marine Natural Source Electromagnetics 43
..................................
1 Source Field -45
.................... 4 Basic Concepts of the .\'SEM Llethod 50 .............................
4.3 Marine XSEM Fields 5-i
............................... 4.3.1 1D Earth 5-1 ............................... 4.3. 2DEarth 60 - ................. 4.4 Distortion of iVSEhI Fields by Topography (2 -- .......................... 4 .. Summar~ and C!onciusion i i
5 Feasibility of Using the VGS Method on a MOR 79
......
5.1 Sensitivity of YSEM Response to the bIantle Upivelling Mode1 $0
.............. 5.1.1 Expected Signal S trength at the Seafloor SO ...... 5-12 Resolution of the bfantle Upwelling Model Parameters 8-4 vii ......................... 5.2 Data Interpretation Scheme 91 ........................... 5.2.1 The 3 parameter 91 .............................. 5.22 Type Curves 93 ............................... 5.2.3 Summary 9s ........................... 5-3 Proposed Survey Design 953
9.3.1 Measurement of Sea Surface Magnetic Field Variations ...... 90
.................. 5.3.2 Bias of the Data by Coast Effect LOO .................. 5.3.3 Bias of Response by Topography 101 ............................... 34 Summary 103
6 Marine Experiment 104
...................................
6.1 Apparatus 104
...................... 6.1.1 EDA Land Magnetometer 104 .................... 6.12 Ocean Bottom hlagnetorneters 109 .......................... 6.1..3 Bottom.Assembly. 112 ........................... . 6.2 Esperimental Procedure 114 ......................... 6.2.1 Location of Stations Il4 ........... 6.2.2 Measurernent Procedure and Data Description 115 ........................... 6.3 Data Analysis Concepts 117 ......... 6.3.1 Least Square Estimation of the Transfer Function 1 IS .................. 6.3.2 Coherency and Confidence Region 119 ............................
6.3.3 Nul1 Hypothesis 121
................................
6.4 Data Processing 122
...............................
6.4.1 Raw Data 122
6-42 Preparation of Series for Fourier Analyses ............. 124
6.4.3 Orientation of Seafloor Magnetometer ................ 127
6.4.4 The PowerSpectra .......................... L31
6.43 Estimation of the Transfer Function ................. 1:31
7 Mantle Upwelling Underneath the Explorer and Endeavour Ridge 135
7.1 Data Discussion ................................ 1.35
. r 2 ID Interpretation ............................... 137 . i . 2.1 Endeavour Ridge ........................... 137
7.2.2 Explorer Ridge ............................ 139
7.3 1D Interpretation ............................... 139
- r . 3.1 Endeavour Ridge ........................... 142 - r . 3.2 Explorer Ridge ............................ 149
7.4 Summary ................................... 151
8 Conclusions 153
8.1 Sumrnary ................................... 1-53
S . 2 Outlook .................................... 1.5.5
List of Figures
1.1 Sketch of oceanic plates ........................... 4
1.2 Magnetic anomaly map of the Juan de Fuca ridge system ......... 6
1.3 BathymetryrnapoFtheEastPacificRiseat9"N .............. 8
1.4 Axial topography of %[ORS with different spreading rates ......... 9
1 ..5 Effect of a change in the potential mantle temperature on mantle iipwelling 10
1.6 Passive mantie upwelling mode1 ....................... 13
...................... 1 . T Dy namic mant le upwelling mode1 17
1.d Dynamic mantle upwelling model with reduced viscosity in melt regioii . 19
. )*)
2.1 Normal 1 D layered seafloor model ........................
2.2 Iipper crustal resistivities and porosities . ODP hole 504b ......... 23
....................... 2.3 Electrical conductivit~ of dunite 26 ....................... 2.4 Possible partial melt geomet ries 2s
2.5 Effective conductivity for melt inclusions with variable degree of intercon-
..................................... nection 31.
2.6 Electrical conductivity as a function of melt fraction for rarious
................................ melt geometries 32
2 . 7 Electrical models of various made upwelling models ........... 3-1
3.1 The Juan de Fuca ridge system ....................... 117
3.2 BLock diagram of the Juan de Fuca and Explorer plate . . . . . . . . . . 39
3.3 Topography of the Endeavour ridge segment . . . . . . . . . . . . . . . . 40
3.4 Seismic cross section interpretation of the Endeavour segment . . . . . . 4 1
4.1 Mean spectral densi ty of geomagnet ic variations . . . . . . . . . . . . . . -Ki
4.2 Equivalent current system for a polar substorm . . . . . . . . . . . . . . 47
4.3 Horizontal magnetic field variations on the British Isles and \Vestern Europe 49
4.4 Hall-space model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
- -
4.5 Induced image of source field in eartli . . . . . . . . . . . . . . . . . . . . .x
4.6 Two Iayer marine model . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
- -
4.7 Sensitivity of marine NSEM fields to half-space conductivity . . . . . . . a
4.8 Induced image of source field in ocean layer . . . . . . . . . . . . . . . . 58
-1.9 Two Iayer marine model with srnall 2D conductivity perturbation . . . . 61
4.10 Sensitivity of marine TE mode fields to a 2D conductivity perturbation . 65
4.1 1 Reflection of anomalous TE mode fields at ocean lqer . . . . . . . . . . 66
4.12 Sensitivity of NSEhI TAI mode fields to a 2D concl~~ctivit~ perturbation . 69
4.1:3 Reflection of anomalous T'II mode fields nt the ocean layer . . . . . . . . TO
4-14 Topographe- of the Endeavour ridge segment . . . . . . . . . . . . . . . . 72
4.1.5 Simplified topography mode1 . . . . . . . . . . . . . . . . . . . . . . . . . 73
4.16 Response of the TM mode fields to topography . . . . . . . . . . . . . . 74
-w
4.17 Response of the TE mode fields to topography . . . . . . . . . . . . . . . / :I
4.18 Current density profiles in the ocean layer across topography moclel . . . TG
5.1 Espected signal strength of sea surface and seafloor magnetic fields . . . dl
5.3 VGS response to standard electrical mantle upwelling mode1 . . . . . . . 8%
... 5.3 Estimated relative error in VGS response as a function of frequency Si ....... 5 . 4 3 response of the standard electrical mantle upwelling mode1 93
5.5 J' response of the standard mode1 with different normal Ia~er conduct ivities 95
5.6 J response of upwelling regions with different shapes ........... 96
- - ....... 3 . J response for upwelling regions with various melt fractions 97 .................. 5 . S EMSLXB coastal resistivity cross section 101 ............ 5.9 Bias in J parameter caused by seafloor topography 102 6.1
Filter
response and calibration constants of
EDA magnetometer ..... 106
6.2 The ocean bottom magnetometer ...................... 107
6 . :3 OBM schematic ................................ LOS
6.4 OBM software flow chart ........................... 111
6.5 Attenuation of magnetic signal by closed pressure vesse1 .......... 113
6.6 Map of Endeavour ridge segment sites .................... 116
6.7 Map of Explorer ridge segment sites ..................... 116
6.8 Raw data seafloor station .......................... 123
............................ 6.9 Raw data land station 124 ............................. 6- 10 Removal of t idal lines 126
6.1 1 Coherency of seafloor data and rotated land data ............. 127
6.12 Time series aligned with strike direction .................. 129
6.13
Coherency
of geographically and strike aligned data ............ 130
6 . 14 Amplitude spectra of magnetic field variations ............... 132
6.15 Magnitudes of transfer functions at ail stations ............... 133
6.16 Phases of transfer functions at al1 stations ................. L3-I
sii
7.1 TM mode response of Endeavour and Explorer sites ............ 136
.... 7.2 1D 1-red model response At to TM mode data at END2 station 13s .... 7.3 1D layered mode1 response fit to Th1 mode data at EXP3 station 1-40
7.4 Electrical models of a various mantle upwelling models .......... 1-11
- w . ....... i . a Measured and standard mode1 J response Endeavour ridge 1-13 .................... 7.6 Land and seafloor END2 time series 1-14 - - ................. r . r Yight time and original transfer f'iinctions 1-45
7.8 Original and night time 3 ratio ....................... 146
. ....... 7.1) LIeasured and thin sheet mode1 J response Endeavotir ridge 1-17 . ......... 7-10 Measured and prism mode1 ,'J response Endeavour ridge 1-H . ........ 7.1 1 Measured and standard mode1 J' response Es~lorer ridge 150
List of Tables
1 SVD analysis of the TM mode VGS response in the range of IO-.' Hz to
10-'Hz .................................... ss
5.2 SVD analysis of the TE mode VGS response in the range of 10-' Hz to
10-' Hz. .................................... SS
5.3 SVD analysis of the Th: mode VGS response in the range of [O-*' Hz to
[O-' Hz .................................... 90 -5.4 SVD analysis of the TE mode VGS response in the range of 10-qHz to
IO-' Hz ................................... !IO
- - ...................... a..3 SVD analysis of the 3 parameter 92
6.1 List of crtiises, sites and data characteristics. ................ 1 1.5
6.2 Tidal lines removed from ratv data. ..................... 12.7
6.3 Seafloor magnetometer rotation angle .................... 1%
6.4 Determination of strike angle from coherency estirnates .......... 131
Chapter 1
Mid Ocean Ridges
1.1 Introduction
The focits of this tliesis is a vertical gradient sounding experiment carriecl out on the .Juan de Fuca ridge. The goal of the esperiment was to ohtain information about the electrical conductivity of the mantle upwelling region beneath the riclge and to use this information to constrain possible mechanisms of melt migration and mani le flo~ Mid ocean ridges occur at the boundaries of diverging tectonic plates.
As t tiese acljaceiit
plates move apart. the underlying mantle rises to fil1 the voici crcated. During its ascent the mantle crosses the solidus and decornpressive melting occiirs. The nielt forined is event uallu emplaced at the ridge asis and forms new oceanic criist t hrough volcanic pro- cesses. Our understanding of the Aow of both the solid and rnolten niaterial is haniperecl II? a Iack of linowledge of vital mode1 parameters such as the amount of melt present. its connectivity and how it is distributed. Connectivity and melt fraction are relatecl to the permeability in the upwelling region and therefore control the migration pattern of the huoyant melt.
Various
models esist. which. based on the assumption of different iipwelling processes. simiilate mantle upwelling (see
Forsyth
(1991) or Turcotte and Morgan (1991) for a discussion).
Mode1
results fa11 into roughly two categories. predict ing eit her broad or narrow mantle upwelling regions with different partial melt geometries and distri but ions. To evaluate tliese upwelling simulations. additional geophysical information of the rnantle upwelling region is needed. Since the electrical conductivity is very serisitiïe to melt fract ion and partial me1 t geomet ry. we propose to use an elect romagnet ic esperiment measuring the electrical conductivity beneat h a [nid ocean ridge. The proposecl mant le upwelling models are evaluated by convert ing t lie predicted upwlli ng and melt patterns into conductivity distributions, or electrical test rnodels. and comparing the response of these electrical test models wit h rneasured
Eh1 data.
The thesis topic calls for discussion of a number of subjects. Our basic knowledpe of mid ocean ridges and the spreading of tectonic plates is based on geological. geophysical and petrochemical observations. various aspects of which I shall descrihe and disciiss in section
1.2. The aspects of geophysical and petrochemical observations chosen represent
the background in format ion needed to understand the mant le upwelling processes and also to understand the geological framework of the Juan de Fuca ridge. on which the data for this thesis were obtained. .Aiter the general discussiori of mid ocean ridges. rnotlels simiilating the upwelling process are int roduced and discussed. h ciliant it ative anal>-sis of the upwelling process shows that two classes of models exist whose predictions are in agreement with observation. Thcse two classes of rnodels predict an iipwelling region with distinctive shapes and partial melt geometries. The mit1 ocean ridge environnient according to t hese mode1 predictions is t lien converted into conductivity distribut ions which are subse~~uently referred to as test models (ctiapter
2). --1 test niodel consists
of a background ID layered structure upon which the predicted iipwelling regions are stiperiniposecl. Tlie conversion process is based on a disciission of the Iayerecl coridiictivity distribution of the seafloor and the dependency of electrical conductivity ori nielt fraction and melt geometry. The ridge on which the experiment will be performed is cliaracterized in chapter 3. This characterization is based on background information presented in the section
1.2 and is
aimed at understanding the general tectonic structure of the Juan de Fuca ridge.
Ridge
segments, on which data for this thesis have been obtained. are discussecl in niore debail.
1 next introduce, in chapter 4. Yatural Source Ekl (NSEM) methods. whicti are iised to
measure the electrical conductivity beneath the mid ocean ridge. On land and in t lie marine environment. the ratio of orthogonal elect ric and tnagnet ic field \-ariations are. under certain conditions, diagnostic of subsurface conduct ivity variations. This met hoc1 is referred to as the rnagnetotelluric (MT) method. In the marine environment the presence of the ocean layer offers an alternative scheme of electrical conductivity mea- surements.
Electrical
conduct ivity measurements can also be obtained using magiiet ic field measurements on the sea surface and seafloor. This rnethod is referred to as the vertical magnetic gradient sounding (VGS) method. While NSEM has been used ex- tensively on land and is ive11 understood for t hat application. only a few marine NSEM surveys have been conducted and published. One of the reasons for tlie lack of siiccess of marine surveys is. as 1 will show. that the application of NSEM in a marine environment ma); not be fully understood or investigated. The preconception that the presence of the ocean Iayer does not alter the nature of the response of the fields is one exaniple of t his lack of understanding. 1 investigate the response of the electric and magnetic fields to ID and 2D conductivity perturbations beneath the seaffoor and show. t hat the ocean layer has a fundamental impact on the behaviour of marine NSEM fields. Furthermore. the impact of seafioor topography on the fields for both
MT and VCiS measurements
is studied.
Based
on these studies. the VGS method is identified as an optimal marine NSEM method for studying the mantle region beneath ridges. In chapter 5. 1 discuss a feasibility study of the VGS method to resolve the tuantle upwelling region and derive requirements on the seafloor instruments to resolw tliese structures. The feasibili ty study leads also to the identification of an optimal interpreta- tion scheme of the acquired data.
1 propose and discuss the design of a VGS esperiment
on the Juan de Fuca ridge and investigate possible bias of the data arising frorn the particular geological setting of the .Juan de Fuca ridge. The actual esperiment is described in chapter 6. 1 first give a description of the equipmrnt iised and the esperirnental procedure. The chosen measurement sites are discussed: t lie data analysis techniqiie is introduced and its application shown on a sample data set. Finally the data are interpreted. A ID Iayered model uncierneatli two locations ori the Juan de Fuca ridge is derived.
The response of test-models with various shapes
ancl conductivities of the mantle upwelling region are compared witti tlie data and the implications for mantle upwelling processes are discussed.
1.2 Mid Ocean Ridges
The outermost rigid surface of the earth, the lithosphere. consists of a mosaic of plates which move relative to each ot her and to the underlying astlienosphere. Figure 1.1 shows a schematic cross section through an oceanic plate. The lithosphere and ast henosphere Figure 1.1: Sketch of oceanic plates. Modified sketch from Forsyth and Uyeda ( 1975). differ in their rheological property. While the lit hosphere is rigid. the ast tienosphere deforrns plastically to forces acting on a long time scâle. Tlie rlieological property is a function of temperature. rvhich increases mith depth. The houndarj. between the lithosphere and asthenosphere is usually defined by an isotherm of around 1200 "C'. Additional to a classification by rheology. the outermost region of the eartli cm be classified by chemical composition. a classification yielding a crustal ancl a mantle region. The chemical boundary is shaliower than the rheological boiindary so t hat the lit liospliere contains the crust and part of the upper mantle. The boundary regions between the plates can be classified into three Liroad categories. Along subduction zones two plates converge. resulting in the destruction of tlie oceanic plate as one plate subducts beneath another. Transform faults are plate hoiindaries where plates slide parallel to each other, a type of plate houndary neiitral in ternis of production or destruct.ion of plate material. Mid ocean ridges (MOR). rdiich are the focus of this chapter. are boundary regions between diverging tectonic plates. As the plates move apart. the underlying asthenosphere rises to fil1 t,he void created and melts decompressively. An upwelling zone containing partial me1 t is created beneat h t lie ridge axis. The rnelt. which is more buoyant than the solid material. migrates upwards and is emplaced within a narrow zone at the ridge aris where it forms nea oceanic criist by volcanic processes. The oceanic crust. which is typicaily 6 km thick. is therefore characterized by a change in bulk chemical composition compareci with the resiclual mant le. Mid ocean ridges are divided into three classes by their rate of spreading. Fast spread- ing ridges such as the East Pacific Rise have a full spreading rate of 2 80 mm/year: intermediate spreading ridges. such as the .Juan de Fuca ridge. a full spreading rate of
50-S0 mmlyear; and slow spreading ridges. such as the Mid Atlantic ridge a full spreading
rate of 5 50 mm/ÿear. The movement and kinematics of plates are inferred from mag- netic anomaly maps. As the oceanic crust forrns and cools below the Curie temperature. it records the local direction of the earth's magnetic field at the time. The spreading of the ocean floor away from the ridge avis and the reversals of the earth's magnetic field on a geological time scale produce a pattern of magnetizat ion on the oceanic fioor rvhich is symmetric about the ridge aais. Since the dates of magnetic field reversals are known. the movement. the spreading rate and fracturing of the plate can be connected to a geological time scale. Figure 1.2 shows a map of magnetic field anomalies at the Juan de Fuca ridge by Wilson (LSSY). based on measurements by Raff and Mason (1961) who were the first to record magnetic anomalies on a mid ocean ridge. Their magnet ic anomaly rnap in fact led to the discovery and postulation of plate tectonics and plate spreading. White regions in Figure
1.2 correspond to negative magnetic anomalies. a
polarization of the magnetic field opposite to the present direction and black regions to positive anomalies. The labels in Figure
1.2 refer to periods of positive anomalies. wliere
the most present period. labeled 1. reaches back from present tirnes to about I .\la and t,lie oldest period
4A from 7.5 to S Ma.
Forsut
h and Uyeda ( 1975) used the kinematics of the plate mot ion worldwide t,o ini-es- tigate the relative importance of the forces t liat drive the plates. They assiimed tliat the motions of the plates are controlled by gravitational body forces and friction acting bot h. on the bottom of the plates. at the interface of the lit hosphere and ast lienosphere. and on the perimeters of the plates. Using global data on the movement of plates the- concluded. that the dominant force that clrives the plate motion is the weight of the relatively cold plate subducting into the hotter upper mantle. Other theories concerning the driving force of plate movement exist. such as the hypothesis that the movement is correlated or driven by mantle convection. However. the size of mantle convection cells by far exceeds the size of some plates. so that esplaining plate movement by mantle con- vection alone seems difficult, though mantle circulation must have some impact on plate Figure 1.2: Magnetic anomaly map of the Juan de Fuca ridge system. Wilson (1988). motion. A comprehensive and widely accepted model on the balance of forces tlriving plate tectonics does not current ly exist.
Figure
1.3 shows a bathymetry map of the East Pacific Rise ridge axis at 9'N. This
fast spreading ridge axis has, to a first approximation, a linear character. Most melt is emplaced at the centre of the ridge axis in the neovolcanic zone which is a few kilometers in width. Some melt surfaces off-axis in volcanic sea mounts. Sea rnounts often form a chain oriented perpendicular to the ridge axis on one of the plates. Although these chains are a dramatic topographic feat ure on the seafloor, volumet rically t hey const it ute only a small fraction of the total crustal production. Davis and Iiarsten ( 1986) explained the existence of the asymrnetric sea mount cha,ins with early melting of chemical and thermal heterogeneities included in the rising mantle underneath a ridge that migrates with respect to the upper mantle. The existence of sea rnounts is generally associated wi t h large melt supplies underneath the ridge. The cross sectional topography of the ridge axis varies with spreacling rate (see Fig- ure
1.4). Fast spreading ridges such as the East Pacific Rise depicted in Figure 1.3 have
lotv amplitude topography. The ridge axis is a topographic high. The neovolcanic zone. which is marked by the letters &V7' in Figure 1.4. is elevated and called the sumrnit rift. The flanks of the ridge axis are labeled "F". Slow spreading ridges on the other hand have large amplitude variations arid the ridge axis is situated in a deep axial dey. or graben. The most plausible model explaining the axial topography has been $\-en b~- Phipps Morgan et al. (1987) and considers the axial topography to be stress supportecl.
The stresses are provided
by the horizontal extension of the lithospliere which tliickens with increasing distance from the ridge axis. .\t slow spreading ridges the temperature in the lithosphere has a large horizontal gradient. the plate thickens cpickly procliicing a large torque moment which is balanced by a high amplitude, short wavelengtti topogra- phy with a wide central graben and high flanks. At fast spreading ridges. the lithosphere thickens more slowly. resulting in a low amplitude. long wavelength topography. and a characteristic axial high. While the amplitude of the ridge axis topography ~aries wi t li spreading rate. the neovolcanic zone at which the melt is emplacecl in the crut and which defines the centre of the ridge axis stays very narrow? at most a few kilonieters in width for ridges at al1 spreading rates. The narrowness of the neovolcanic zone poses. as ive will see later! an important const raint on mantle upwehg simulations.
Besides
topographic variations across strike. srna11 scale deviat ions in the linearit- of
Sea mounts -
/ off axis volcanisrn Figure 1.3: Bathymetry map of the East Pacific Rise at 9ON. Tighe ( 1996). Figure 1.4: Axial topography of MORS wit h different spreading rates. Macdoriald ( 1982). the ridge axis are observed along strike. Deviations in axial linearity divide the ridge into segments, which are typically .50 to 100 km long. Two ridge segments can be identified in Figure 1.3. These segments show an overlap. which is often observed at fast spreading ridges. but which is less common for intermediate and slow spreading ridges. The across-strike ridge morphology changes along segments. and is usually characterized by an increase in depth and development of a graben towards the end of each segment. The change in topography has been associated with a change in melt supply. It implies a maximal supply in the segment centre and a decreasing supply towards the ends of the segment.
Although
the melt supply. and therefore the melt floiv. ~inderneath the ridge riiight be three dimensional, bathymetry maps of the ridge show that mid ocean ridges on a scale of several hundred kilomet ers are a two dimensional feature. The dominant topograp hic signature of the ridge is an increase in ocean depth wi th increasing age of the lit liosphere. The change in ocean depth from the ridge axis to the lithosphere at an age of 8 SIa is more than
2000 m. The topography is a result of the cooling and sinking of hot
young lithosphere as it migrates away Erom the ridge crest and can be explained by a simple cooling model (Parsons and SLater, 1977). The model assumes a half-space at a given temperature which is held constant at one side of the model. Assurning vertical conductive cooling to the ocean layer, horizontal movement due to plate spreading. and isostatic equilibrium, the model predicts topography due to thermal contraction ivit h an erplicit square root of age dependency, which is in agreement with observations to crustal ages of Y Ma independent of spreading rate.
TEMPERATURE
Y RIDGE X b
SOLI DUS =40kb
Figure 1.5: Effect of a change in the potential mantle temperature on melting (upper panel) and ridge axis depth and vertical extent of upwelling region (lower panel). Klein and Langmuir (1987). While the increase in depth with increasing distance from the ridge asis shows a square root of age depeodency on al1 ridges. the large scale subsidence rate of the ridges varies from
200 ml J!Ma to 450 ml dXZ and is inversely correlated with the mean axial dept h
of the ridge (3.5 km to 2 km). In context of the simple lialf-space cooling model, a possible explanation of this correlation is a change in the initial temperature of the half-space representing the mantle underneath the ridge aris.
The first
order effect of a change in mantle temperature is summarized in a sketch by Klein and Langmuir ( 1987). Figure 1.5. The upper panel shows the adiabat along rvhich the mant le mows during its ascent in the upwelling zone for high (X) and low (Y) initiai temperature of the mantle. If the temperature of the mantle underneath the ridge is high the solidus is crossed at greater depth resulting in a deep melt column with large melt fractions beneath the ridge axis (lower panel Figure 1.5). An isostatic balance of the taller melt column results in shallower axial dep t h and larger su bsidence rates. At low initial mant le temperat cires. the amount of melt produced is less, the mean axial depth is greater and the subsidence rate smaller. The esplanation of a change in asial depth and subsidence rate by a change in man- tle temperature underneath ridges is also supported by petrochemical analyses of mid ocean ridge basalts (hIORB) (Klein and Langmuir, 1987). bIORBs are the end procliict of melting of the mantle. migration of the melt through the mantle and cooling and differentiation of the magma near the surface. They therefore preserve the liistory of the melting process. .An analysis ol the chernical e~oliition of the melt shoived. that the abundance of Sodium and Iron oxides relative to other oxides can be used as an indicator of the extent of melting and the pressure at which melt was formed. Klein and
Langmuir
( 1987) reported a positive correlation of the mean estent of melting and the pressure at wliich melting takes place. thus linking high rnean melt fractions to melting at great depth.
Furthermore.
a negative correlation of these parameters with the mean axial depth was found.
While a change
of mantle temperature explains the obserïed changes in subsidence rate and asial depth. it also calls for an increase in crustal thickness for ridges with a tiigh mantle temperature (see Figure l.5), due to an increwe in melt fraction in the upivelling region that eventually forms the oceanic crust.
Seismic experiments. on the other
tiand. show tliat the thickness of the crust remains essentially constant. at around
6 - 7 km. for
al1 ridges regardless of spreading rate. This indicates that the simple c~ualitative mode1 sketched above is not sufficient. In the following section quantitative mantle opwelling models are presented and investigated in more detail.
1.3 Mant le Upwelling
In the past decade or so, attempts have been made to model the mantle upwelling process. A goal of these studies was to compare model predictions with observations and to provide information about the most important processes occurring during the upwelling process.
To a first approximation, and at
Ieast
at fast spreading ridges. the floiv of the mantle beneath the MOR can be modelled as a 2D process. However. the assumption of two dimensionality is harder to justify for the migration of melt beneath the axis which may be distinct lrom the flow of the residual mantle. Although the topography of the ridge is two dimensional on a scale of a few 100 km. small scale topographic variations seg- ment the ridge, hinting that three dimensional mantle upwelling processes may occur. -4s ment ioned above, small scale topographic variations are associated ivit h variations in the melt supply. Although the general mantle upwelling flow might be predominantly two dimensional, the flow of the melt that is produced during the upivelling prscess is almost certainly more cornplex. However the study of mantle upwelling in two di- mensions gives a useful insight into the processes involved and into the production and emplacement of melt into the crust.
1.3.1 Passive Mant le Upwelling
The most basic mantle flow model considers passive upwelling. shown in Figure 1.6. It assumes diverging plates viscously coupled to the as t henosp here represent ed as a Iialf- space with a constant viscosity. The divergence of the plate will clrive a so-called corner flow, i.e. the asthenosphere rises to fil1 the void while, al the same time. it is advectecl sideways. The stream lines of the corner flow are depicted as solid
Iines
in Figure 1.6. The mantle upwelling region outlined by the stream lines has a broad prismatic shape. where the bottom of the prism has a typical width of 100 km and a top-width of several tens of kilometers. The exact dimension of the upwelling region depends on the value of viscosity chosen.
Mantle
rising underneath the ridge will follow an adiabatic gradient that is smaller than the adiabatic gradient of the solidus, thus partial melting occurs. Figure 1 .G: Steady state flow and distribution of liquid and solid in a passive flow mode1 cross section of a spreading centre ivith a spreading velocity of 1.5 cm/yr. Daslietl line. porosity in
1% contours (function of melt present); solid line, stream line of mantle flow:
Iieavy solid line, degree of melting in 4% contours ( represent masimum degree of melt ing reached up to that point along the streamline. not the fraction of nielt present ): lieavy dashed line, solidus. The rate of eruption is depicted at the top. It represents the i-ertical integration of the melt production beneatli each point. Scott and Stevenson (1989). The degree of rnelting is depicted as heavy solid horizontal lines and the solidus as the heavy dashed line in Figure 1.6. The isolines of degree of melting are horizontal. since the degree of melting is a function of the lithostatic pressure that changes wit h height only.
The presence of melt will
increase the porosity of the made material. Wi th increasing melt fraction the pore space will become connected and such that the permeability of the material increases.
Since
the melt is more buoyant the melt can migrate riptvards. The melt migration flow is modelled using Darcy's law. where the permeability is a function of melting and porosity. The velocity of the migrating melt is distinct from t hat of the residual mantle.
Dashed
lines in Figure
1.6 show the melt fraction. e-g. the arnount
of melt present in the upwelling region, which. due to melt migration. is different irom the extent of melting. The rate of eruption depicted at the top of the mode1 sliows the amount of melt t hat is emplaced at the surface as a funct ion of distance from the ridge aris. It is calculated from the difference in the vertical Rorv velocity of the solid and molten materiai.
Obviously
model predictions will clepend on value and functional behavior of various model parameters. For example the shape of the solidus. which is not known elractlj-. will have a direct effect on the amount of melt produced in the upwelling region. tvhich lias an impact on the porosity. permeability and t herefore melt fraction distribution in the upwelling region.
However,
for a given melt estent or shape of solidus. the geo- metric distribution of the pore space. which is also not known. will strongly influence the permeability and the melt Aow. also altering the melt fraction distribution. Fiir- thermore a change in the initial mantle temperature. will. as discussecl already earlier. have strong impact on the extent of melting and therefore melt migration. Since most of the mode1 parameters have not bcen measured and are not known preciselj-. interrial model parameters have been varied. such that the predicted mode1 results agree with the few geophysical and petrochemical observations. Since the mode1 parameters are not independent and the observations against which the model predictions can be coniparetl are few, the mantle upwelling modeiing problem is severely underdetermined.
However.
we can check for a general consistency of a passive mantle upwelling mode1 predictions and observations.
1.3.2 Discussion of the Passive Mantle Upwelling Mode1
As discussed previously. the passive mantle upwelling model cont radicts the observation that the crustal thickness is approximately 6 km and is more or less independent of spreading rate. An increase in spreading velocity produces a higher mantle upwelling velocitv as the conservation of mass has to be obeyed. The temperature in tlie upwelling region is controlled by conductive cooling to the ocean. Assuming a constant initiai mantle tem- perature. higher upwelling velocities imply that the upwelling region is hotter due to a decrease of the conductive cooling effect to the ocean? such that the melt fraction and porosities throughout the partially molten region will increase. The associated increase in permeability allows the liquid to escape more easily and consequently a thicker crust is produced if the spreading velocity is higher. As explained previously. a change in the initial mantle temperature between ridges. implied by the correlation of axial depth and subsidence rate. also produces a change in the criistal thickness. .A possible remedy to the spreading rate and rnantle temperature dependence of criistal t liickness is the *Diopside-out" hypot hesis (Hess. 1991 ). If melting normally proceecls until al1 the Diopside. which is the mineral that melts most easily. is depleted and the melt is reniokred as fast as it is created. then the primary control of melt proc!uction would be compositional and independent of spreading rate. The upwelling rnantle niay alreatly have been depleted of al1 easily melted components before it reachrs a depth affectecl bv conductive cooling. The rapid removal of melt calls for larger. vein like conduits. evidences of which were observed recently on the Oman Ophiolite.
The most st ringent const
raint on the passive upwelling model. hoivever. is the narrowness of the neovolcanic zone. As depicted in Figure 1.6 the modelled melt emplacement takes place on a scale of tens of kilometers. In fact, as the spreading velocity increases. the region of melt emplacement broadens even further, as tlie melt is advected aw- from the ridge crest faster. In order to account for the narrow neovolcanic zone. two approachcs are considered. Dynamic models predict that the modelled upwelling region is too broad ancl that the mantle upwelling region is much narrower.
A narrower upwelling zone ivo~ild require
that the vertical upwelling of the solid as well as the melt velocity is larger t han in the passive upwelling model, as conservation of mass has to be obeyed. An increase of the upwelling velocity of the entire mantle in this region can be achieved by buoyancy forces and changes in viscosity. Melt migration models on the other hand. predict. that the mantle upwelling itself is broad. yet the melt migration is focused.
These
models predict a change in the flow pattern of the melt onl.
1.3 -3 Focused Melt Migration Model: Passive Upwelling
Sparks and Parmentier (1991) propose that meit freezing will occur along a P-T solidus that slopes, like near-ridge isotherms. away from the the ridge asis. They suggest t hat large melt fractions will build up within roughly one viscous compaction length of the freezing boundary and that melt will migrate along this "channel" to the ridge axis. /ln alternative melt migration scheme ivas suggested by Phipps Morgan ( 1987). Melt. once formed. builds networks dong rvhich melt migrates. As the networks deform diiring spreading by the near ridge turning from vertical to horizontal flow. an anisotropic fabric is formed. tliat will preferentially align towards the ridge axis and which cliannels melt tomards the neovolcanic zone on the ridge axis.
Since
melt needs to be able to move freeiy t hrough the mant le upwelling zone. a me! t migration mode1 recpires part iai melt geometries characterized by well connected pore space. The melt migration model esplains the narrow neovolcanic zone. It is additionally compatible irit h t lie existence of the observed off-axis volcanism.
1.3.4 Focused Mantle Flow Model: Dynamic Upwelling
The alternative scheme to focus the melt emplacement is to look for a niechanism whicli focuses the upivelling region altogether. In order to achieve a narrower upwelling zone. the upwelling velocity has to be increased. such that the mass conservation is obeyed.
The dominant force
whicli wvould increase the upwelling velocity is buoyancy.
The niost
plausible source of buoyancy is composi tional buoq*ancy due to melt retention. t hus re- quiring that partial melt, once formed. is unable to form networks and a partial irielt geometry that is characterized by a lack of connected pore space.
If large melt fractions
(in the order of
40%) are retained in the solid matrix to siipply added biioyancy to the
upweliing mantle and at the same time the viscosity of the partially molten region is Figure 1.7: Dynarnic mantle upwelling mode1 driven by buoyancy forces including I>iioy- ancy of solid liquid mixture and compositional differentiation. C~iri-e labels as in Fig- ure 1.6. Scott and Stevenson (1989). drastically reduced the modelled upwelling zone is narrowed to a size of a few tens of kilometers (Buck and Su 1989, Su and Buck. 1991) as shown in Figure 1.8. SIodeling results presented by Buck anil Su (1989) and Su and Buck ( 1991) showed also t.hat an increased upwelling velocity reduces the dependency of the produced crustal tliick- ness on spreading velocity and potential temperature due to a predicted downward flow component in the dynamic model.
A "petrochemical remedy' such as the --Diopside
Out" hypothesis by Hess (1991) is therefore not needed to bring modeling results and observations into agreement.
However
there are some arguments against dynamic models. .A ctiemical analysis of
4I0RBWs
suggests that melt does not seem to be in equilibrium with its paternal material and that melt fraction is on average on the order of
10 to I.576. which weiglis against a
retention of large melt fractions in the solid material. Furthermore. the narrorv ~ipwelling zone does not provide an explanation for off axis volcanism.
1.3.5 Summary and Conclusion
:\ simple quantitative analysis shows that two classes of model. the melt migration model and tlie dynamic model. yield predictions that are in general agreement wi th observa- tions.
Since
the mantle upwelling process is a comples non-linear process and nianj- of the important parameters are correlated and t heir values and funct ional behavior are not sufficiently well known. a distinction regarding which one of tliese moclels represerits true aspects of the upwelling rnechanism underneath the ridge cannot be tlecided by Iiirtlier flow modeling efforts alone. -At t his stage the modeling needs addit ional geophysical information and input. The important Feature that distinguishes the melt migration and dynamic model is the size of the upwelling zone and the partial melt geometry. As dis- cussed above. the melt migration mode1 preclicts a large prisrn - shaped upwellirig zone and well connected pore space containing melt to enhance the formation of networks. along which melt migration occurs. The dynamic rnodel predicts a narrow itpn*elling zone and a partial melt geometry in which melt filled pores are connected poorly to al- low retention of large melt fractions in the upwelling zone to supply the buoyancy neecletl to focus the upwelling flow. .A suitable parameter, which is sensitive to the amount of melt present and the partial melt geometry is the electrical conductivit.
Since
the size of tlie upwelling region and t lie partial melt geometry between these models are distinct. a rneasiirement of elect rical Figure 1.S: Dynamic mantle upwelling mode1 driven by buoyancy forces inclutling btioj-- ancy of solid liquid mixture and compositional differentiation with recluced viscosity in melt region. Curve labels as in Figure 1.6. Scott and Stevenson (1989). conductivity underneath the MOR will help to determine the upwelling niechanism and distinguishing whether upwelling underneath the ridge is of a dynamic nature or not. In the following chapter 1 dl derive an electrical test-models. rvhich represents the melt migration model and the dynarnic model. The electromagnetic response of t hese models will in a later section be compared with data rneasured on the .Juan de Fuca ridge. The electrical rnodels consist of a normal layered electrical model representing the ocean floor. to which a 2D region representing the upwelling region is added. While the appropriate size and shape of the 2D region representing the dynamic and migration model is relatively easy to determine. the effective conductivity of a two phase material of solid and melt with different partial melt geometries needs to be discussed in detail in order to derive appropriate conductivity values for the dynamic and the migration model.
Chapter 2
Electrical Conductivity Model of a
Mid Ocean Ridge
In this section the electrical conductivities of the mantle uprvelling region and its sur- rouoding are introduced. First the "normal- oceanic model. which is t he background layered structure to the -anomalousb' mantle upwelling region. is discussed ancl derived. For the electrical conductivity in the upwelling region. the eiectrical conductivity of a two phase mixture has to be considered and is discussed iising a range of mising laws.
2.1 Layered Normal Model
.\t relatively low temperatures and pressures. such as in the vicinity of t lie earth's surface. the electrical current flow in rocks is due to ionic transport in fliiids containecl in the rock. The effective conduct ivity is tlierefore a function of the mobility and conccnt ration of ions. For a porous rock containing water. the effective conductif-it5- has been found ernpirically by Archie (1942) and is given by a, = a@mSn~, mhere O is the fraction pore volume (porosity). S the fraction of pores containing water. ow the conductivity of water. rz = 9 and a and rn constants with 0.5 5 ci 5 2.5 ancl
1.:3 5 rn < 2.5.
The electrical conduct ivi ty structure beneat h the seafloor in general can be approxi- mated to the first order to change with depth only. I start the discussion of changes ocean layer 0.3 ohm m 1 cru* 5G00 ohm m sediments
10 ohm m
lifhosphere ! O km
1.5 km
1000 ohm m
asthenosphere 100 ohm m Figure 2.1: ID lqered model thought to represent the normal electrical layerecl mode underneat h the seafloor. in the electrical conductivity and thickness of the lavers by presenting the model t llat
1 think represents the Iayered normal seafloor and then validate and tlisciiss eacli Iayer
separately. The standard normal layer is depicted in Figure 2.1. It consists of four major units. the ocean. the crust. wit h a shallow high conduct ire region. t lie lit liosphere ancl the ast henosphere. The knowledge of elect rical conductivities in t tiese uni ts is derived either [rom marine ES1 experiments. borehole measurements or laboratory experiments.
Available
data from either or al1 measurement types are included in cliscussion of the electrical conductivity of each unit.
2.1.1 Ocean Layer
The conductivity of seawater ranges from as much as 5 S/m near the surface to about
9.2 S/m below the main thermocline. ivhich occurs at a depth of a feiv huntlred rneters.
It is mainly governed by temperature and salinity. A reasonable approximation of the
LhhOW A-t mfstivity (ohmm)
Figure 2.2: üpper crustal resistivities and porosities derived from ODP hole 504h.
Becker ( l9d5).
conductivity of seawater in S/m as a function of temperature T in Celsius is given by where K is 10 S/(m OC,') (e-g. C'have et al.. 1988). The ocean laver in the normal mode1 used throughout this thesis is assumed to be 2.5 km deep with a conductivity of 3.3 S/m. The occan depth of 2.5 km has been chosen as it represents the average ocean deptli of the Juan de Fuca ridge. and moreover is a typical depth of MORS worldwide.
2.1.2 Crust
Borehole measurements of crustal conductivity were obtained during the continuing
Ocean
Drilling
Project
(ODP). Figure 2.2 shows the borehole drilling results of ODP hole 504b. drilled at the Costa Rica rift (Becker. 1985). The crustal region can he fur- ther subdivided into four units for the purpose of determination of electrical conductivity. The upper unit consists of seawater-saturated sediments. which are usually less than a few hundred metres deep close to t,he ridge. Due to high porosities and poor consoli- dation. the conductivity in this region is only slightly Less than that of seawater, and is in the range of 1 S/m. The underlying unit. with a thickness of approximately 700 rn consists of pillow lava and minor lava flows which have a slightly reduced conductivity. in the order of -1 S/m. In the subsequent transition from pillow lavas to sheeted dikes and massive units the conductivity reduces. The overall decrease in conduct ivity car1 be explained by a decrease in porosity due to an increase in Lithostatic pressure and compaction. The conductivity in the transition zone to the sheeted dykes and massive units ha a value in the order of .O1 S/m to -001 S/m and shows a further decreasing trend.
While
borehole measurements give an important and detailed est imate of the conduct iv- ity at a point value and on a small scale. it is important to investigate the conductivit~ measured over longer scales. as only the regional conductivity will have an impact on the response of deep penetrating methods such as NSEM. The regional condiictivity might differ on average from the local conductivity due to the presence of fluid fillecl cracks. Cont rolled source Eh1 measurements obtained by receiver-t ransmit ter separa- tion of up to .? km (Evans? 1991) show. that the upper Iayer of approxirnately 1 km depth is indeed very conductive, in the range of 10 S/m to .O1 S/m. Thus controllecl source EM results confirm the point value obtained by the ODP5O-! borehole on a larger scale. Evans ( 199 1 ) also reports t hat the electrical conductivity profiles do not Vary with distance in the vicinit- of the ridge axis.
Conductivit-
estimates obtainecl on the ridge axis as well as on 100 000 year old crust (approximatel?; -5 km off the ridge asis) are similar. For the normal model. I separatecl the crust into two electrical regions. The upper most sedinient and high porosity pillow lava region is summarized into a laycr of 1 lin1 thickness with a conductivity of -1 S/m. The thickness of the second layer is set to 6 km. as recluired by the seismic measurements. Since the conductivity at a depth of 1'100 ni is in the range of .O02 S/m and shows a decreasing trend. the bulk conductivity of the entire second layer will be loaer. I chose a value of .O002 S/m for the remaining crost. wliich is also in agreement with the results of other deeper penetrating controlled sotirce EM experirnent such as by Cos et al. ( 1986) discussed in the nest section.
2.1.3 Lithosphere
Due to increasing lithostatic pressure the porosity and water content of the lithosphere is low. such tliat the effective conductivity ceases to be governed hy electrolytic pro- cesses. Instead, the conductivity is mainlg influenced by the mobili tu and abundance of charge carriers in the crystal structure, which is typically an exponential function of temperature, Le. a = 00 exp(-Ei/bT), (2.3) where Ei is an excitation energy. The conductivity of the lithosphere has ben investigated by a controlled source Eh1 survey (Cox et al.. 1986) and by laboratory experiments (Constable and Duba. 1990). Based on an analysis of the data. Cox et al. ( 1986) report that the lithosphere underneatti the crust is extremely resistive, and assign an average minimal conductivity of 10-' S/m to this region. The high lithospheric resistivity is a direct result of the fact that a transrnitted electrornagnetic signal bas been observed sorne :30 km or more away from the transmitter.
Laboratory experiments
aim to measure the dependence of the electrical conductivity in the mant le region as a function of temperature. The identitp of the charge carriers is poorly known even for the niost abundant crystal in the rnantle. olivine. Plausible laboratory conductivi ty values are derived from measurernents usually carried out on single olivine crystals, which can be shown to be compatible with polÿcrystalline mea- surements (Constable and Duba, 1990). Figure 2.3 shows. that the conductivities of a single olivine crystal taken frorn the Red
Sea and San Carlos Peridot and a polycrys-
talline rock cornposed prirnarily of olivine (a dunite collected at Jackson.
C'oiinty,
S.C..
USA) are in good agreement. The conductivity of the mantle is sensitive to temperature and increases roughly three orders of magnitude. from
10-' S/m to 10-~ S/m as the
temperature is doubled from from TOO OC to 1400 'C. Tliese loiv conductivity values are in agreement with the bulk value of LO-~ S/m measured by Cox et al. (1986). Olivine conductivity is the sum of several thermally activated processes. /\t least three processes have been identified by changes in slope of the logarithm of measured conduc- tivity versus temperature, marked by arrows in Figure 2.3. The mechanism dominant in the middle temperature range is thougiit to be moving holes in the valence band. cliarig- ing Fe+* to FeC3. The other two processes dominating the lorver and higher temperattire range are not known. I summarized the lithosphere in rny normal electrical model into one layer of 53 kni thickness with a resistivity of 1000 ohm m. hlthough the resistivity of the lithospliere varies by three orders of magnitude from
106 ohm m to 103 ohm m (see Figure Z3),
Olivine Conductivity us Temperature
10-2 \
Duba rr d. ( 1974)
JieLsoa Co. Dunitr
7 8 9 10
Reciprocal Temperature. 104/K
Figure X3: Electrical conduct ivity of duni te versus temperat ure. Constable and
Duba (1990).
the very resistive layer only constitutes less than half of the lithosphere. Furthermore. since the tempeiature in the vicinity of the ridge is most likely increased. I regard a less resistive value of the lit hosphere as the most appropriate.
2.1.4 Asthenosphere
The astlienospheric conductivity is the biggest unknown. Marine NSESI experiments conducted by Filloux ( 1981), Law and Greenhouse (1981). Ferguson and Edwards ( 1996) and EklSLAB initiative (Wannamaker. 1989). require that a relatively resistive litho- sphere is underlain by a layer with an abrupt increase in conductivity to 0.002 S/ni (Oldenburg et al., 1984) or 0.2 S/m (Wannamaker et al.. 1989). No laboratory esper- irnents have been able to mimic asthenospheric conditions and give an estimate of the conduct ivi ty. A change of conduction mechanism in the laboratory experiments ment ioned above lia
Oceans Documents PDF, PPT , Doc