JOURNAL
OF GEOPHYSICAL
RESEARCH,
VOL. 94, NO. B10, PAGES 14,111-14,125, OCTOBER 10, 1989
MagnetotelluricObservationsAcrossthe Juande Fuca SubductionSystem
in the EMSLAB Project
PHILIPE. WANNAMAKER,
1 JOHNR. BOOKER,
JEANH. FILLOUX,ALAN(3. JONES,
GEORGE
R. JIRACEK,
ALAN D. CtIAVE, PASCALTARITS,HARVES. WAFF, GARYD. EGBERT,CHARLES
t. YOUNG,
JoHNA. STODT,MARIOMARTINEZG.,LAWRIEK. LAW, TAKESI YUKUTAKE,
JIRO S. SEGAWA,ANTHONY WI-R'IE, AND A. W. GREEN,JR.
A magnetotelluric(MT)transect has been obtainednear latitude 45øN from the active Juan de Fuca
spreadingcenter,acrossthe subduction•x•neand Cascadesvolcanic arc, and into the back arc Deschutes
Basin region. This paperpresentsthe MT data set and describesits major characteristics
as they pertain to
the resistivity of the subduction system. In addition, we discussthe measurementand processing
proceduresemployedas well as importantconcernsin datainterpretation. Broadbandaudiomagnetotelluric
(AMT)/MT soundings(approx. 0.01-500 s period) were collectedon land with considerableredundancyin
site location, and from which 39 siteswere selectedwhich conattainupper crustalheterogeneitybut sense
also into the uppermantle. Fifteen long-periodMT recordings(about 50-10,000 s) on land confirm the
broadbandresponsesin their commonperiod range and extendthe depthsof explorationto hundredsof
kilometers. On the Juan de Fuca plate offshore,33 out of 39 sea floor instrumentsat 19 locationsgave
good results. Of these locations, five magnetotelluric soundingsplus two additional geomagnetic
variation
sites,covering
theperiodrange200-105s approximately,
constitute
theoceanbottomsegment
of our profile. The featare of the land observationswhich probablyrelatesmost closelyto the subduction
processis a peak in the impedancephaseof the transversemagneticmode around30-50 s period. This
phase anomaly, with a correspondinginflection in the apparentresistivity, is continuouseastwardfrom
the seacoastand ends abruptly at the High Cascades. It signifies an electrically conductive layer in
otherwiseresistivelower crust or upper mantle, with the layer conductancedecreasingeastwardfrom the
coastto a minimum under the Coast Range but increasingsuddenlyto the east of the central Willamette
Basin. The higher conductanceto the east is corroboratedby the vertical magneticfield transferfunction
whosereal componentshowsnegativevaluesin the period range 100-1000 s over the samedistance. The
transverseelectric mode apparentresistivity and phaseon the land display a variety of three-dimensional
effects which make their interpretationdifficult. Conversely,both modes of the ocean floor soundings
exhibit a smoothprogressionlaterally from the coastal area to the spreadingridge, indicating that the
measurementshere are reflecting primarily the large-scaletectonic structuresof interest and are little
disturbed by small near-surface inhomogeneities. The impedance data near the ridge are strongly
suggestive of a low-resistivity asthenospherebeneath resistive Juan de Fuca plate lithosphere.
Approaching the coastline to the east, both impedance and vertical magnetic field responsesappear
increasingly affected by a thick wedge of depositedand accretedsedimentsand by the thinning of the
seawater.
INTRODU•ON
Juande Fuca spreadingridge throughthe subductioncomplex,
the Willamette Basin, the Cascades volcanic arc, and into the
Our goal in acquiringmagnetotelluricmeasurements
is to
increase understandingof the physicochemicalstate of the
Earth's interior through its influence on the electrical
resistivity. Generationand subductionof the lithosphereare
amongthe most fundamentalprocessesshapingthe earth and,
for investigationby inductionresearchersin North America,
the Juan de Fuca system is fortuitously close at hand.
Consequently, we have collected a profile of tensor
magnetotelluric(MT) soundingstrendingapproximatelyeast-
back arc Deschutes Basin region. Consideration of the
electric field with the magnetic in the MT method (i.e.,
measuringthe plave wave impedanceof the Earth) increases
our resolution of buried structure, especially layerlike
geometries,over using the magnetic field alone. This also
makes interpretation more difficult, however, because the
electric field can be strongly influenced by shallow crustal
structureswhich are not our primary targetsof investigation.
The landwardsegmentof the profile (Figure 2) consistsof
west
39
and
located
near
the
center
of
the
EMSLAB
magnetovariationarray (Figure 1) [EMSLAB Group, 1988;
Gough et al., this issue]. Because the profile crossesthe
northernOregon coast near the town of Lincoln City, it is
referred
to as the Lincoln
Line.
The
measurements
are
intendedto yield a detailedresistivity crosssectionfrom the
1Theaffffiations
of theauthors
maybefoundonthelastpageof this
paper.
Copyright 1989 by the AmericanGeophysicalUnion
selected
broadband
AMT/MT
observations
(approximately 0.01-500 s period) augmentedby 15 longperiod MT recordings(about50-10,000 s). A purposeof the
broadbandsoundingsis to accountfor the effects of upper
crustal lateral heterogeneity,but such soundingsalso sense
throughthe crust and well into the upper mantle. The longperiod MT data confirm, but are generally of higher quality
than, the broadbandresponsesin their commonperiod range
and extend the depths of exploration to hundreds of
kilometers. On the sea floor portion of the profile (Figure 1)
[EMSLAB Group, 1988], we have five MT soundingsplus
two additional magnetovariationsites which collected data
Paper number 89JB00680.
over the periodrange200 s to approximately
105s. The
0148-0227/89/89JB-00680505.00
coarsersite spacingrelative to the land in part is dictatedby
14,111
14,112
WANNAMAKER
ET AL.:MAGNETOT•LURIC
OBSERVATIONS
ACROSS
JUANDE FUCASYSTEM
50o 132
ø
124
ø
120
ø
1160
50 ø
PACIFIC
48o
-
46 ø
128
ø
//
JUAN DE
PLATE
MT.A
aT,;,p
Z
•o•// .M,P
.MT,
P
• -- .
ß
.... • 446
ø
FUCA
'M .
•//
P eM,P M,VMT
// eMT,
44 ø
%oPLATE
M•
•&
O
M
Magnetometer
V Ve.icaITeIIuric
.•//
U O
( • 44ø
MTMagnetotelluric
42 ø
PVolcano
Pressure
Sensor
A
GODA
40 ø
BASI N
I _._•42
ø
Mendocino
Fracture
Zone
;32ø
128
ø
124
ø
0
I
120
ø
4oø
1160
Fig. 1. Instrument
locations
of entireEMSLABarray.Onland,theunlabeled
dotsindicate
locations
of three-component
magnetovariometers.
The E-W trendingconcentration
of dotsin northwestern
Oregondenotesthe long-period
MT
recorders
alongthelandportionof theLincoln
Line. Seafloor
arraysitesarelabeled
according
toinstrument
typeandsites
0 through8 aregivenfor theseafloorextension
of theLincolnLine.No datawasrecovered
fromseaf•oorsites0 and6.
the muchgreaterexpenseof instrument
deployment
by ship
but moreimportantly
reflectsthe muchgreateruniformityof
SourceFieldsandTensorRelationships
the near-surfacegeologicstructureon the seafloor. Resultsat
The MT methodmakes use of naturally-occurring
electromagnetic (EM) fields as sourcesfor exploring
periodsshorterthan about200 s inevitablyare lackingin
resistivitystructure[Vozoff,1986]. Usually,the
oceanbottom MT becausethe magneticsignalsespecially subsurface
remoteoriginof thenaturalfieldstogetherwith theveryhigh
are attenuatedby the overlying conductiveseawater. The
index of refraction of the Earth relative to the air allows us to
primary purposes of this paper are to present the
measurementstaken by the numerousresearchersinvolved as
a unified, coherent data set and to describe major
characteristics
of thedataas theypertainto theresistivityof
treat
theincident
vector
electric
(E)andmagnetic
(/•) fields
as planar and propagatingvertically downward into the
ground[butseeEgbertandBooker,thisissue].Owingto the
the Juande Fuca subductionsystem. In addition,we discuss dominanceof conduction
over dielectricdisplacement
for the
themeasurement
andprocessing
procedures
employedaswell
rangeof resistivitiesandperiodsof interest,the propagation
as importantconcerns
in datainterpretation.Severalpapers of EM fieldsin the earthis diffusive. The amplitudes
of the
on interpretation follow this one which make use of all or
electricand corresponding
orthogonalmagneticvectorsof a
part of the Lincoln Line data set.
planewaveenteringa uniformconducting
half-space
decrease
A brief reviewof MT theoryis givenhereprimarilyto
over
adistance
called
theskin
depth,
6'"503
•] pT
define terms and conventions. The conceptsintroduced by1/e
in
meters,
where
p
is
the
resistivity
(inverse
of
conductivity
provide a foundation for the detailed discussion of our
observations which follows.
We then describe the
cr)in ohmmetersand T is theperiodin seconds.We seehere
measurementand processingtechniquesemployedby the
that the downgoingincidentfields can reach the deeper
researchers
who gathereddataalongour profile. This should structures
andbe scattered
backwith significantamplitudefor
indicatethe difficulty of the measurements
we make and the
measurement,
only at the longerperiods. This ability of MT
rigorof themethods
requiredto achievegoodresults.
to discriminatein depth holds fundamentallyeven for
WANNAMAKER
ET AL.: MAGNETOT•L•C OBSERVATIONS
ACROSS
JUASDE FUCASYSTEM
123 ø
124 ø
14,113
122 ø
121 ø
/
\
!
I
,•
/
D
•
AH
/
/
/
/
•
/•
s
•
/
A
'%..
.......
/
]I Deschutes•
I
m
•
• / Um
..................
atillaI
Plateau
i
•
] '•+
i
Basin
•,• ++• ',,• 1{
,
I
I
Lava
)1
I
• Plains
o
+
Broadband
MT
ß Long Period MT
Fig. 2. Broadbandand long-periodmagnetotelluricsoundingsalongthe landwardportionof the Lincoln Line and also
along the mini-EMSLAB test traversefarther south[Younget al., 1988]. Urban areasincludePortland(P), Salem (S),
Eugene(E), Newport (N), Lincoln City (L), Bend (B), Madras (M), and The Dalles (D). Cascadesvolcanoesare labeledas
triangles:Mount Hood (H), Mount Jefferson(Y),andSouthSister(S).
complex natural settings, although often the effects of
shallowstructures
may persistin the data from shortthrough
long periods(seediscussion
of observedMT responses).
The secondaryor anomalous EM fields induced around a
three-dimensionalstructure in general may be polarized
arbitrarily with respectto the incidentfields. Therefore, the
mathematicallyto extremizethe tensorelementsand thereby
obtain estimatesof geoelectricstrike [Word et al., 1971]. In
natural three-dimensionalsettings,however, the principal
direction or strike resulting from the impedancemay vary
independentlyfrom that due to H z, with both period and
relation between the total (E)and (H) at the surface for a
plane wave sourcemust be expressedby a complex-valued
tensor impedance,i.e.,
Ratherthan work directly with the impedanceelementsas
they appearin (1), we insteadconstructsimplefunctionsfrom
each. One useful functionis apparentresistivity,generically
denotedby Pa, and definedfrom the elementmagnitudesin
location.
(1). Specifically,
from Zxy, forinstance,
it is
"-
Zyy..
a
(1)
i
which can vary with period, from soundingto sounding,and
with orientation of the x-y coordinates. Similarly, the
relationship
(2)
can be definedfor the verticalmagneticfield. A right-handed
coordinate system with z positive down is assumed. With
field data, one often rotates the original x-y coordinates
tO/to
2
I z,,.,,I
(3)
in unitsof ohmmeters,whereto= 2•r/T is radianfrequency
and /t o is the magnetic permeability of free space. This
operationwhen applied to the impedanceon a uniform halfspace returns the intrinsic resistivity p. For more
complicatedgeometries, Pa versusT usually resemblesa
smoothed version of the true resistivity over increasing
distancesbelow the measurement
site. It is commonpractice
to substitutemeasuredapparentresistivity values into the
14,114
W•••
ETAt..:1VI•a•,•E'r••c
O•SEaV^T•ONS
ACROSS
JUaNDEFuc• SYSTEM
previous formula for skin depth to obtain crude estimatesof
depth to structure. Complementaryto the apparentresistivity
electronicsensorand amplifier noise, chemical instability in
is thephaseof theimpedance,
e.g., ½xy.Overa uniformhalfspaceof anyresistivity,½xyis 45ø. Parenthetically,
½yxin
al., 1983]. Cultural noise and thermaldrift presumablyare not
seriouson the oceanbottom but attenuationespeciallyof the
magnetic field signals by the overlying seawater prevents
collectionof MT data at periodsshorterthan a few minutes
[Filloux, 1987]. Sources of noise notwithstanding, the MT
fields we measurealso departsometimesfrom the assumption
of a plane wave geometry [Egbert and Booker, this issue].
Such departuresoccur toward long periods due to solarmagnetic storms in the magnetosphereand ionospherebut
also at periods even less than 1 s during times of nearby
thunderstorm activity.
These characteristics of the
measurementproblemnecessitateultra low-noisesensorsand
signal conditioning plus sophisticated time series
processing. Relevant features of the individual recording
this case is -135ø due to the right-handed coordinatesystem,
but we generally rotate this phase to the first quadrantfor
display and interpretation.Over any one-dimensionalearth, a
Hilbert transform relationship between the apparent
resistivityand the impedancephasecan be proven [Boehl et
al., 1977]. This proportionality between the impedance
phaseand the slope of the apparentresistivityversus T is
observed almost invariably over heterogeneousearths as
well.
A similar
Hilbert
transform
relation
between
the real
and imaginarycomponents
of M•x and M•y is usually
observedin keeping with the approximatelycausalnature of
the linear system(2).
The structureof the tensorrelations (1) and (2) simplifies
for special cases of resistivity structure. Over a two-
dimensional
earth,elementsZxx,ZyyandMzx vanish
if thex
andy axes are alignedwith and acrossstrike respectively(the
principal directions). This constitutesa decouplingof the
total fields into two independent modes: the transverse
electric(TE, or E-polarization
with Ex, Hy, Hz) andthe
transverse
magnetic
(TM, or B-polarization
withHx, Ey,Ez).
In other words, the TE mode comprises electric fields
(currents)parallel to strike while the TM mode comprises
electric fields (currents) perpendicularto strike. Since we
could carry out only a single well-sampledprofile of MT
soundingsin EMSLAB, we mustderivemeaningfulresistivity
models using primarily a two-dimensional analysis [see
the E field electrodes, vibration, and thermal drift [Clarke et
instruments used in EMSLAB
are summarized
in Table
1 and
explainedin more detail below.
The electric fields invariably are obtained as voltage
differences over orthogonal grounded bipoles, which
terminate in quiet and stable chemical electrodesconsisting
of a base metal and its salt [Petiau and Dupis, 1980]. The
combinationusedmost often for the land data was Pb-PbC12
but goodresultswere obtainedalsowith Cu-CuSO4. To boost
the E field signalof the land measurements,
bipole lengthson
the order of 100 m were laid out. A variety of types of sensor
are availablefor measuringthe magneticfield. The broadband
AMT/MT
instrumentsin our study mainly utilized large,
ferrite-cored induction coils for the three componentsof (B)
although one systemused a cryogenic SQUID. The longWannamaker et al., this issue; Jiracek et al., this issue]. To
periodrecordersemployedflux gate devices. Two additional
facilitate two-dimensional modeling, the Lincoln Line was
channelsof horizontalmagneticfield were recordedor taken
designedto crosstectonicunits, includingboth upper crustal from a neighboringsite for use as independentreferencesfor
noise cancellation in the broadband systems (see below).
structure and the subduction system, whose preferred
Output from the electric and magneticfield sensorswas fed
orientationappearedpredominantlyN-S.
Equations(1) and (2) reducemuch furtherwhen considering immediately through high-gain analog amplification and
band limiting, usually with custom-built electronics. The
just one-dimensionalgeometries. Now, the polarization of
theincidentEM fieldsis immaterial
sothat Zxy=-Zyx = Z1D, broadbandsystemsdivided their signalspectruminto at least
three overlapping period bands. This was necessary to
which is the scalar impedanceof the one-dimensionalearth,
while Zxx,Zyy and H z all equal zero. One-dimensionaleconomize
conductivity models of individual MT soundingsare very
popular because the algorithms fit easily on desktop
computers,but these models frequently are suspectdue to
commondeparturesfrom one-dimensionalityin nature. They
can be valuable in constructinginitial guessesfor subsequent
two-dimensional modeling [Waft et al., 1988, Livelybrooks
et al., this issue;Youngand Kitchen, this issue].
Measurementand ProcessingTechniques
A minimum of five channelsof electromagnetictime series
(Ex, Ey,Hx,Hy,Hz) needto be collected
at a surveysitein
order to estimate the MT transfer functions (1) and (2). Of
foremost importance,the natural fields we measure are very
low in strength;full-scale temporalvariationsof only about a
nanoteslafor B =ktoll and a microvolt per meter for E are
typical. The signalspectrummoreoveris far from white with
strengths in the so-called mid or dead band (0.2-20 s
approximately) often 10-100 times weaker than at periods
above or below. Sources of electromagnetic noise on land
which can rival or overwhelm the signal and so degradeour
estimates
include
man-made
or
cultural
interference,
on the volume
of time series to be collected
and
also allowed extra amplificationof the weak midbandsignal.
After antialiasfiltering, the signalswere digitized and results
stored, sometimespartially processed,on floppy diskettes
for the broadbandsystemsor on cassettetape for the longperiod recorders. In addition to high sensitivity and
complicatedin-field processingcapability, it was essential
that all MT systemswere sturdily constructedto withstand
adverseweather, dust, and mechanicalshock during transport
and deployment.Also, the long-periodMT systemshad to be
securelyconcealedto preventtheft or vandalismduringtheir
approximately2-month durationof recording.
A very thoroughdiscussionof the difficult measurement
of
sea floor electromagneticfields is given by Filloux [1987].
Electric fields again are obtainedas voltage differencesover
finite bipoles, but for deployment from ships, it is
compelling logistically to use rigid bipoles of short length
(6.4 m typically and in EMSLAB). The length must be
sufficienthoweverto escapeperturbationsof the electricfield
in the vicinity of the instrument case. These arise from
possible electronic leakage currents and electrochemical
reaction between the case and the seawater, and because the
case itself constitutes a resistive inhomogeneity in the
WA•AMAI•ERET At..:MAOt•'rOTELLURIC
OI•SERVATIOSS
ACROSS
JUA•DE FUCASYSTEM
14,115
TABLE1. Summary
of Attributes
of Magnetotelluric
Recording
Systems
Usedin theEMSLABProject
Broadband
Land
Michigan
Techical University
University
Manufacture
Periodrange
(nominal),s
Magnetometer
Electrodearray
PhoenixLtd.
0.003-1800
coils
(Phoenix
Ltd.)
cross
(30mbipoles)
LongPeriodLand
SDSU/CICESE
University
of Oregon
GSC/Ottawa
of Utah
In-house
0.025-600
In-house
In-house
0.025-1000
0.008-4000
squid
coils
coils
(SHECorp.) (Geotronics
Ltd.) (Geotronics
Ltd.)
Seafloor
University
of
In-house
In-house
40-30,000
32-175,000
128-DC
fluxgates
(EDALtd.)
fluxgates
(EDALtd.)
suspended
magnets
(in-house)
L
cross
cross
cross
cross
(100m)
(loom)
(loom)
(3OO
m)
cross
(6.4 m)
Electrodetype
Pb-PbC12
Pb-PbC12 Cu-CuSO4
Pb-PbC12
Pb-PbC12
Pb-PbCl
2
Reference
local
(wirelink)
local
(wire link)
local
(wirelink)
singlesite
singlesite
Processing
cascade
FFT
FFT
cascade
decimation
decimation,
Institution
In-house
(150m)
remote
(radiolink)
Scripps
Washington
Ag-AgCI
neighboring
site
cascade
decimation,
cascade
FFT,
decimation,
robust
coherence
sortingrobust
processing
robust
processing processing
In-fieldcomputer
HP 9845B
DEC PDP 11/23 DEC PDP 11/23
DEC PDP 11/23
Datel Ltd.
cassette
logger
SeadataLtd.
cassette
logger
PhillipsLtd.
cassettelogger
Reference
to a company
orproductnamedoesnotimplyapproval
or recommendation
of theproductby members
of EMSLABor
by theagencies
whofundedtheprojectto theexclusion
of otherproducts
thatmaybe suitable.
electrically conductive ocean host. An additional design
criterion for the ocean floor electric field recording was the
ability to measure the long-term field for oceanographic
purposes [Chave et al., this issue]. To extract the minute
signal from potentiallylarge electrodeoffset voltages,which
themselvesmay drift in time, the electrochemicaloffsetswere
estimatedindependentlyby an alternatingshort-circuitingof
the voltage electrodesthrougha salt-bridgeswitch or chopper
[Filloux, 1987]. The H field sensorsof four of the sea floor
magnetometerson the Lincoln Line were suspended-magnet
devices. These have a sensitivity of about 0.2 nT (for B),
adequatefor long-periodfield variations,plus the advantage
of extremely low power consumption. The other ocean
bottom magnetovariation recorders on the Lincoln Line
utilized flux gate sensors. Finally, one cannotoverstatethe
extreme care required in fabricating sea floor
MT
instrumentation to withstand the high-pressure, corrosive,
and conductive environment of the ocean bottom [Filloux,
1987]. That only one instrumentpackagefailed to return at
all and that 33 out of 39 instrumentsin EMSLAB gave good
data is a tribute to the growingreliability of long-termEM
experimentson the oceanbottom.
Spectral analysis methods used in EMSLAB have been
reviewedby Joneset al. [this issue]and comparedusinglongperiod MT time seriesgatheredfor the project.One approach
of long standingis simply to apply a fast Fourier transform
(FFT) to each channelof EM time series,form cross-power
and autopowerspectralaveragesover a finite bandwidth, and
computeimpedanceand vertical field transfer elementsfrom
the appropriate spectral combinations [Sims et al., 1971].
More recently, analysisusing cascadedecimation [Wight and
Bostick, 1980] has become popular because it allows
virtually real-time spectral estimation. Bias errors in the
transfer
function
estimates
of the broadband
MT
data were
reduced or eliminated using the aforementioned reference
fields, with error boundscalculatedaccordingly[Gamble et
al., 1979a,b]. Most of the broadbandsystemsutilized local
reference fields, measuredonly about 200 m from the base
sensors, which worked well in most instances but which
sometimeswere inadequatein the WillametteBasin areawhere
cultural
noise
could
be correlated
over
several
kilometers
distance. Both FFT and cascadedecimationapproacheshave
been modified to sort spectrafrom subsetsof the EM time
series according to quality and to reject outliers, thereby
treating the nonstationary and sometimes nonuniform
behavior of the signal and noise [Stodt, 1983; Egbert and
Booker, 1986; Chave et al., 1987; Chave and Thomson, this
issue]. To ensure consistency in sensor reliability,
calibration,and data processingfor both broadbandand longperiod MT systems,a preliminaryfield test at six sitessouth
of the Lincoln Line was cardedout in August1984 [Younget
al., 1988] (Figure 2). Close agreementbetween the different
broadbandinstrumentsand with the long-period results in
their common period range confirmed that the various land
MT systemscould yield equivalentdata at a particularsite on
different days.
The sea floor data in particular were processedwith a new
robust, remote reference algorithm with jackknife error
estimates that is described by Chave and Thomson [this
issue]. The periods of interest for sea floor studiesextend to
105 s, a decademorethanon land. Sincethereare serious
questionsaboutthe homogeneityof sourcefields at very long
periods and because it is possible that contaminating
electromagneticfields of variable wavelength induced by
oceanmotionsmay be present,a carefulstudyof the behavior
of remote reference
estimates
as a function
of the reference
type was conducted. Using sea floor site SF3 as the local
station, robust remote reference estimateswere computed
using electric and magnetic referencesat nearby (SF4) and
distant(SF7 and SE3) sea floor sites and magneticreferences
14,116
from
WANNAMAKER
ET AL.:MAat•'rOTELLURIC
OI•SERVATtONS
ACROSS
JUANDE FUCASYSTEM
land stations at Victoria,
British
Columbia,
and
broadband values over their common period range (see
discussion of static effects below). Overall, the MT
response along our profile on land appears adequately
sampledbecausethe soundingsare isotropicat short periods
periods(beyond104s) for distantlocations.
Thisis probably and tend to group distinctly accordingto the major geologic
features crossed, in particular the upper crustal ones. To
due to motional electric fields from gravity waves, and
electric references were rejected. Distant land magnetic define the electrical structure of the Juan de Fuca plate
referencesproducedbiasedresponsefunctions at long periods offshore, there are five MT soundingsplus two additional
(greaterthan 104 s), suggesting
sourcefield heterogeneity geomagneticvariation sites. Sampling demandsfortunately
over the 200-1000 km distance separatingthe measurements. appear much lower on the sea floor due to the greater
Identical bias effects were seen with distant sea floor sites.
uniformity of geologicconditionsat shallow depth. Values
By contrast,remote referenceestimatesusing a nearby sea from smoothedindividual responsecurveshave been used to
define pseudosections
for the sea floor data also.
floor magnetometer were essentially identical to
conventionalsingle-siteresults at periodsover about 1000 s
yet dramaticallychangedthe estimatesat shortperiodswhere TransverseMagneticImpedanceFunctions
American Bottom, Oregon. The results show that sea floor
electricreferencesproduceerraticresponsefunctionsat short
periods(below 1000 s) for nearbyreferencesitesand at long
magnetometernoise is expected to increase as the external
magneticfield becomesmarkedly attenuatedby the overlying
seawater. In most instances, changes of over an order of
magnitude in apparentresistivity were observed,and usable
responsesto periodsof about200 s couldbe found. Basedon
these tests, all of the sea floor data were processedusing a
nearby sea floor magnetic site located at the same
geomagneticlatitude for a reference.
OBSERVED MAGNETOTELLURIC RESPONSES
Theobserved
transverse
magnetic
quantities
Pyxandcpyx
will be presentedat the outsetfor two reasons. First, as we
will describe, the assumption of a two-dimensional
resistivity variation beneath the profile is more immune to
common
three-dimensional
effects
for the
TM
mode data
than for the TE. Second,the TM mode appearsto exhibit
more clearly the effects of resistivity structurerelated to the
Juande Fucasubductionsystem.
The TM mode can be complicatedto understanddue to
potentially
widefluctuations
of theelectricfield Ey over
We now presentthe magnetotelluricmeasurements
gathered
along the Lincoln Line. The majority of the data is
summarized in the form of pseudosections,where east-west
location is the abscissaand log period is the ordinate for
contour plots of apparent resistivity, impedance phase and
sometimes
short
distances.
These
fluctuations
arise
from
profile of concentrated MT observations,and the geologic
structures beneath trend approximately N-S, a twodimensionalframe-work will be emphasizedhere over a full
three-dimensionalanalysis. The measurementsthereforewill
be described in terms of their analogous two-dimensional
modes of polarization, with the applicability of a twodimensionalassumptiondiscussedfor each mode. Consistent
with this, we define the x and y axes to coincide with
geographic north and east for all stations and all periods.
Following each set of pseudosections,examples of the raw
data at several individual MT sites are given to convey the
statisticalquality of our observationsand to substantiatekey
characteristics of the response as they pertain to earth
electric charge distributionsoriginating wherever the field
crosses resistivity boundaries [Price, 1973; Jones, 1983;
Wannamaker et al., 1984]. The charge in turn reflects the
requirement that electric current across boundaries be
continuous.Toward longerperiods,the influenceof boundary
charges around two-dimensional or three-dimensional
structures can remain strong but becomes independent of
period as the governingHelrnholtzequationreduceslocally to
Poisson'sequation.The result is a near-fieldor static,roughly
multiplicative distortionof the impedancewhich would exist
if the structure were not present. Therefore, while deep
resistivity variationsaffect MT observationsonly at longer
periodsthroughthe skin-depthconceptexplainedpreviously,
near-surfacechangesboth may figure in the observationsat
short periods and superimpose their effects upon the
impedanceto arbitrarily long periods. This situationhas lead
to repeatedcalls for the ability to samplethe MT response,
especially the electric field component,more finely when
structure.
needed
therealandimaginary
partsof Mzy.Because
wehavebutone
The descriptionof the measurementson land is separated
somewhat
from
that
of
the
sea floor
observations
as the
period range and stationdensityof the two data setsare quite
different. Values defining the pseudosections
of the land data
were derived from hand-smoothing(by P. E. W.) of the field
soundingcurvesat the 39 broadbandand 15 long-periodsites.
The broadbandsoundingsusedwere selectedfrom a total of 73
site occupationsfrom three separatefield excursionsby the
four institutionsaccordingto data quality and freedom from
local three-dimensionaleffects. For pseudosectiondisplay,
we wish to preserve, through the long-period range, the
lateral variation in the apparentresistivity as sampledby the
more numerous broadband soundings.The smoothed longperiod data hence were interpolatedlinearly to give values at
the 39 broadband locations and the resulting long-period
apparent resistivities were static-shifted to merge with the
or even to make
continuous
E field
measurements
along one's profile [e.g., Bostick, 1986].
Fortunately,it is only the impedancemagnitude,and hence
the apparent resistivity, which is distorted in this manner.
As the distortion becomesperiod-independent,impedance
phase anomalies diminish to zero. Small-scale, shallow
inhomogeneityconsequentlyaffects the phase data only at
relatively shortperiodswhile larger, deeperbodiesaffect the
phaseat longerperiods.This very much easesthe recognition
of deep target structuresbeneath extraneoussurficial clutter.
Nevertheless,the apparentresistivity must be incorporated
into the modeling processin order to resolve absolutevalues
of crucial parameterslike depth and resistivity.To our great
advantage,
Pyxand ½yxarenotverysensitive
to variations
of the resistivity structure off-line from a profile (threedimensional effects); it is the boundary charge near to the
measurementsite which most strongly affects the electric
WANNAMAKER
ET AL.: MAGNETOTELLURIC
OBSERVATIONS
ACROSS
JUANDE FUC^SYSTEM
OBSERVED
field and hence the impedance[Wannamaker et al., 1984].
Given the prevailing N-S structural trends beneath the
Lincoln Line, we believe that an accurate cross section can
14,117
w
CR
WV
WC
HC
DB
-2
result from two-dimensional modeling of the TM mode
response [see Jiracek et al., this issue; Wannamaker et al.,
this issue].
Figure3 andPlate1 displaythepseudosections
of Pyx and
•Pyxobserved
alongthe landportionof the LincolnLine.
(Plate 1 can be found in the separate color section in this
issue.)
Characteristic of the apparent resistivity
pseudosection
is a strongvertical orientationof the contours,
indicating that several major structural areas have been
crossedover the line's 200 km length. The persistenceof
this orientation to the longest periods is a clear
manifestationof boundarychargeeffects as describedabove.
øl//II I
-•+2 •
+•
•
i1[ i i ] i i i i i i i
L [ i i i
Ifil
Ill
I I I I I I I I
•
•
i i i I I i [ i 11i
I I
-Pyx
I I I I [ I I I II I [ ii
-1
_
i j ]
I
I
I
[
0
bo
Porous, water-saturated sediments of the Willamette Basin
(•yx
constitute the most important upper crustal inhomogeneity
affectingthe dataon land. They resultin valuesof Pyx
typically below 10 ohm m throughoutnearly the entire period
range extending over about the central third of the profile.
Moderately higher values ~30 ohm m at short periods)near
the middle of the basin correspondto overlying Columbia
River basalt flows of Late Tertiary age. Flanking the
expression of the Willamette Basin are high apparent
resistivities, of the order of 100-1000 ohm m, over the Siletz
River marine basaltsof the Coast Range to the west and the
volcanic
and intrusive
rocks of the Western
Cascades
to the
east.
More near-surface material of low resistivity is evident
between the Westem and High Cascadesin Figure 3 and may
represent sediments and volcanics accumulated in the
Cascadesgraben. Beneath the easternmostfour stationsof
our profile in the Deschutes Basin, outcropping basaltic
+3 •
+4
30
; ;77 ;,,•77; 7;; ;7 ;', 7 7 ; ;7;; ; ; ;;7•,,,,
i
0
Fig.3.
i
50
• , , ,
i
100 km
Pseudosections
of transverse
magnetic
apparent
resistivity
Pyxand
impedance
phase
•pyx
observed
along
thelandward
portion
of
the LincolnLine. Importantphysiographic
and structuralregions
crossedinclude the Coast Range (CR), Willamette Basin (WB),
Westem Cascades(WC), High Cascade(HC), and DeschutesBasin
(DB). Zero tick on scalebar corresponds
to coastline. Numerical
valuesfor contouring
arefromfield dataof 39 broadband
and 15 longperiodMT sitessmoothedand mergedby P. Wannamaker. Color
versionsof these pseudosections
appearin Plate 1 in the separate
color section of this issue.
rocksgiveriseto valuesof Pyx around100ohmm at the
shortestperiods. They fall rapidly to about 20 ohm m as
period increasesto 1 s, however, with correspondinghigh
aboveandbelow.The subtleinflectionin Pyx belowthe
values
of rpy
x. Based
justonskindepthcomputed
usingPyx, Coast Range which correspondsto the small phase peak
these data indicate a low-resistivity layer only several
hundred meters down. Examining more closely the Coast
Rangedata,oneseesPyx fallingfromtheshortest
periods
to
a subtle minimum around 1 s, along with phases usually
exceeding 45ø over the sameperiod range. Based again on
describedis not discernedeasily in the apparentresistivity
pseudosection.
When Pyx is plottedin soundingform,
however, the inflection is clearly above the noise (Figure 4).
East of the central Willamette Valley, an actual minimum in
Pyx at around100 s can be seenboth in the apparent
skin deptharguments
using Pyx, thesedataindicatethe
resistivity pseudosection and the sounding curves (e.g.,
existenceof an electrically conductiveunit or layer over a
depth range roughly of 1-2 km beneath the Siletz River
volcanics. Structure off-line from the profile cannot be
responsiblesince a minimum at around 1 s is seen in both
Figure5). Theanomaly
in Pyx and Oyx strongly
suggests
1-10
eastward increase in conductance, is visible in the mini-
the presence of a conductive layer whose conductivitythicknessproduct(conductance)increasesmarkedlyfrom west
to east. Given the range of periods and apparentresistivities
Pyx andPxy..Beneath
all theconductive
structures
described characterizingthe anomaly,skin depth calculationssuggesta
depth of a few tens of kilometersfor the layer. Note also that
aboveliesa resistive
basement
andmiddlecrustcausingPyx the layer appearsto end or change abruptly near the High
toriseandepyxtofallto30øorlessovertheperiod
rangeof Cascades. Similar evidence for the layer, including its
s.
The feature of the observationswhich has intrigued us the
EMSLAB results to the south [Young et al., 1988]. The
mostis theridgein cpyxjustbelow100s periodextending amplitudeof our anomalyis considerablyweaker, especially
across most of the Lincoln Line profile (Figure 3).
The
in the Coast Range, and occursat somewhatlonger periods,
amplitudeof theridgeis nearly15ø at thecoastbut decreases than that observed by Kurtz et al. [1986] on Vancouver
to only several degrees below the Coast Range and
westernmostWillamette Valley (see Figure 4). However, the
phaseanomalyincreasesabruptlyto the east startingbeneath
the central Willamette Valley and reaches values under the
WesternCascadesthat exceedby 25ø thoseat periodsa decade
Island. This probably resultsfrom much higher resistivities
above and below the conductive layer beneath Vancouver
Island, relative to thosebeneaththe Coast Range, as implied
by the generallyhigh values of apparentresistivity of Kurtz
et al. At presentwe considerthe phaseanomalyof Figure 3 to
14,118
WANNAMAKER
ETAL.:1VL•GNE'FOTELL•C
OBSERVA•ONS
ACROSS
JUAN
DEFUCA
SYSTEM
layer, the deephigh-resistivitymaterialextendstowardand
may terminateunderthe High Cascades.
The steepupturnin
+3
rpy
x at theverylongest
periods,
5000s andbeyond,
plusthe
incipientrolloverin Pyx here,showthat the resistor
E +2
possesses
a lower limit beyondwhich resistivitydecreases
again.This final mediumto whichour datais sensitive,
if the
data can be shown not to be respondingto some off-line
(three-dimensional)
structure,shouldlie at depthsof hundreds
o
'"" + 1
+13
of kilometers.
The resultsat longerperiodseastof the High Cascades,in
the DeschutesBasin, are very different from thosedetailed
above. We have no long-period MT impedancerecordings
east of Mount Jefferson but broadband data to 800 s have been
obtained(Figure6). At periodsbeyondabout30 s, a decrease
in Pyxisquiteapparent
whilerPyx
riseswellabove
50ø. Our
1 og T
measurementshere in the east, however, show no indication
(s)
Fig. 4.
Broadband(circles) and long-period(squares)MT data
collectedby the Universityof Utah ResearchInstitute(UURI) andthe
Geological Survey of Canada(GSC) some48 km inland on Lincoln
Line. Displacement of about 1 km between broadbandand long-
periodsitesis responsible
for smallstaticoffsetin p•. Thelongperiod
anomaly
in •yx presumed
tobeassociated
withsubduction
(A)
is at its weakest in this area.
Error bars are one standard deviation
wherethey exceedthe size of the plottingsymbols.
+3
•-
xu
'
i
•,-
i
ux
EMSL
, 123
i
mmmm
mmm
• .... •..:..."
o
'-'+1
i
i
i
i
i
i
I
I
I
I
I
I
'90
•
60
'-'
30
q3-
model in this region should be treated with caution. The
geomagneticarray coveragemay be especiallyuseful here in
constraining our model structures. Nevertheless, it is
obvious that a fundamentalelectrical boundary exists almost
directly underthe High Cascades.
In Figure 7 and Plate 2 are the pseudosections
of apparent
resistivity and impedance phase of the transversemagnetic
mode of the sea floor data. (Plate 2 can be found in the
separatecolor section of this issue.) In contrastto the land
E +2
+0
of the deep,high-resistivity
materialso obviouswestof the
High Cascades. Interpretation of these data may be
complicatedby their proximity to a major east trending
boundaryin resistivitybetweenthe Blue Mountainsblock to
the north andthe Basinand Rangeto the south[Goughet al.,
this issue], so that the assumptionof a two-dimensional
apparent resistivity, we see only a smooth variation from
west to east in the offshore results, suggesting minimal
effects by shallow lateral inhomogeneity local to the
individual measurementsites. The westerntwo soundingsat
the shortestperiods exhibit high values of impedancephase
and apparentresistivitieswhich fall rapidly with period from
values around 100 ohm m. Based on skin depth using
apparent resistivity, these results indicate a pronounced
•"%...ß..."ß.•
0
-'1
•
1 og T
i
i
i
+1 +'2 +3 +4
'
- xu
I
•,-
I
i
ux
i
i
EMSL 200
'
(s)
ee
Fig. 5.
Broadbandand long-period MT data collected by the
University of Oregon (circles)and GSC (squares)about 123 km
inland on the Lincoln
Line.
I
i
i
i
be the most obvious manifestation of resistivity structure
which may be associatedwith the subductingJuande Fuca
platebelowOregon. Especiallycloseattentionwill be given
to this anomalyin the interpretationpapersin this issue,to
define the permissiblerange of causativestructureand its
significancefor physicalconditionsin the subduction
zone.
A very substantial
volumeof resistiveearthmateriallies
below the conductivelayer just described,presumablyin the
ee me em el •le
I
+2
log
T
i
+3
[s)
Fig. 6. Broadband
MT sounding
collected
by UURI about200 km
uppermantle.Thatis evidentfromthesteepupturnin Pyx, inland
on the Lincoln Line (its easternmost site of Figure 2).
plusthedropin Oyxtolessthan30ø,atperiods
of about100
Observethe lack of evidencefor the deepconductivelayer which was
s andlonger(Figures3-5). As wasthecasefor theconductive apparent
in Pyx andcpyx
westoftheHighCascades.
WANNAMAK]•
ETAL.:MAONEFOTFiLURIC
OBSERVATIONS
ACROSS
JUA•DEFUCA
SYSTEM
OBSERVED
measured. Contributing to this effect also may be the
thinning and termination of the seawater toward the coast
[Ranganayakiand Madden, 1980]. A graphicconsequence
of
the wedge and seawater edge structuresis the more vertical
orientation of the apparent resistivity contours as we
approach the eastern end of the profile. Finally, all five
W
+2
SF8
SF7
I
I
SF5 SF4 SF3
I
I
14,119
i
Pyx
soundings
at the very longestperiods,greaterthan 104s,
show ½yx increasing
againto valuesover 60ø with Pyx
decreasing. This behavior occursmore obviously in the TE
mode (e.g., Figure 8). On the face of it, the deepeststructure
we have sensed on the sea floor, with a scale of hundreds of
+5
I
I
!
I
kilometers, has resistivity which is very low.
The
relationshipbetweenthis conductorand the deepestunit, also
conductive,inferredfrom the land data is not clear at present.
I
+2
Transverse
ElectricImpedanceFunctions
•+3
Over purely two-dimensional structures, the
response
.functionsPxyand cpxybehave
muchmoresimply
o +4
thandothe•I'M functions.
Thisisbecause
theelectric
field
Ex is parallel to resistivity contacts everywhere so that
boundary charge effects do not arise. Consequently,the
+5
Fig. 7.
TE mode
0
100
I
I
influence
uponboth Pxy and cpxyof relativelylarge,deep
200 km
I
Pseudosections
of transverse
magneticapparentresistivity
py•,andimpedance
phase•y•,observed
at thefiveMT soundings
on
the ocean bottom segmentof the Lincoln Line. All electric field
recording instrumentsand four of the magnetic recorderswere
designedand built by J. Filloux. The magneticvariometerat SF8
was constructed
by T. Yukutake,J. S. Segawaand colleagues,that at
SF1 wasproducedby L. Law, andthatat SF2 by A. White. Color
structuresmay emerge toward longer periods without static
distortion by small-scale resistivity variations near the
surface. The two-dimensional TE mode responsein that
senseresemblesthe responseof one-dimensionalor layered
earths. Basedon this characteristic,
the TE mode impedance
data, identified from field data by its correlation with H z,
often is used to constructone-dimensionalresistivitymodels
versionsof thesepseudosections
appearin Plate 2 in the separate
color section of this issue.
i
+3
,= +2
plate. In the TE dataat SF7, a minimum
in cpxy
and an
inflection
in Pxy around
2000s periodat firstglancesuggest
o
that somewhat more resistive mantle material
aforesaid conductive zone.
i
i
SF7
• -
transition from low resistivity in the upper few tens of
kilometers to high resistivity below. The high resistivity
region may representthe asthenosphere
of the Juan de Fuca
i
'
•x
oo
oo
ooooOøøøøøoO
+!
lies beneath the
This subtle characteristic however
is not sharedby the site closestto the ridge, SF8, nor by the
westernmost
sites SG4 and SE3 off the Lincoln
Line and thus
cannotbe consideredtypicalof the sea floor soundings.
Toward the east end of the ocean bottom profile, the
soundings appear increasingly affected by conductive
material at relatively shallow depths. The apparent
resistivitiesat the shortestperiods,say 200 s, progressfrom
+13
c90
I
I
I
I
I
I
m
1
:..
.
about10ohmm atSF5to3 ohmm atSF3.Thevalues
of ½yx
lessthan45øandtheuptumin Pyxatperiods
below1000s
3O
at these three sites is indicative of underlying resistive
material. The high-level but large-scaleconductormanifest
here probably is the thick wedge of depositedand accreted
•
i
i
+2
sedimentsassociated
with the subducted
slab. The response
of this wedgeconductorappearsto be screeningevidencefor
any possiblelow-resistivity asthenosphere
as suggestedat
the western sites. In the TM
•
I
i
i
+3
log
+4
i
+5
T
mode, the E-W limits of the
sedimentary
conductor
acttokeepPyxlowat theeastern
sea
floor sites from the shortestthrough the longest periods
Fig. 8.
Apparent resistivity and impedancephase responsesat
ocean bottom site SF7. Error bars are one standard deviation where
they exceedthe size of the plotting symbols.
14,120
WANNAMAKER
ETAL.'MAG•••C
at individual
MT
sites when two-dimensional
effects
OBSERVATIONS
ACROSS
JUAN
DEFUCASYSTEM
seem
apparent in the measurements [Word et al., 1971].
Unfortunately, the TE results are very susceptible to
deviations of the structure along strike (three-dimensional
effects), which are apparentat least in the surficial geology
beneath our transect and are likely to exist in the deep
structureswe are studying.Boundary chargesdo arise at such
deviations
andleadto staticeffectsin Pxy thatcannotbe
accommodated in purely two-dimensional TE modeling
approaches [Wannamaker et al., 1984]. Once again, the
impedancephase is worth checking for evidenceof any deep
target, even though its contributionto the responsemay not
be modeledcorrectlyin terms of a two-dimensionalstructure.
Pseudosections
of Pxy and cpxyobserved
alongthe land
segmentof the Lincoln Line appear in Figure 9 and Plate 3.
(Plate 3 can be found in the separatecolor section of this
issue.) The pronouncedvertical orientation of the apparent
resistivity contoursis as obvious here as it was in the TM
modedata. Whereasthe lateralvariationin Pxy at short
periods reflects differencesin near-surfaceresistivity along
the profile, its preservationto the longestperiodsmeans that
these near-surfacestructuresprobably also vary substantially
perpendicular to the profile as discussedjust above. The
signatures of the Willamette Basin and Cascades graben
conductorsperhapsprovide the clearestexamplesof suchend
Noteworthy is the lack of an obvious expressionof the
subductionsystemin the TE mode resultsof Figure 9. In
contrast
to the TM mode,nosharppeakin cpxy
or inflection
in Pxy is visiblein theperiodrange30-100s. Thereis
instead an overall decreasein the TE
apparentresistivity
plusgenerallyhigh valuesof phase(-50 ø) at periodsof 100 s
and longer. One possiblefactor contributingto the TE
behavior may be variations in the inferred deep conductive
layeringalongstrike. In fact, thereis a major westwardshift
in the Cenozoic volcanic front [EMSLAB Group, 1988] only
50 km north.
The mini-EMSLAB
results 50 km to the south
describedby Younget al. [1988] supportthis conjecture. At
site5 of thatstudy,oneobserves
anactualpeakin cpxy
and
aninflection
in Pxy around100s periodsuggesting
thatthe
deep conductiveunit is respondingin a somewhatmore onedimensional manner farther away from the possible
discontinuity. The TE mode data of Kurtz et al. [ 19 8 6]
appearto show the conductivelayer below VancouverIsland
more clearly, perhapsagain due to higher resistivitiesabove
and below that layer.
Thereareotherpuzzling
complexities
observed
in Pxyand
Oxy at longperiodson land.Thehighvaluesof phaseand
steeply declining apparentresistivitiesaround 100 s beneath
the Western Cascadesmay primarily representthe decaying
influenceof high resistivitiesin the upper and middle crustof
effects.Valuesof ½xyof 30øor lessaround
10s periodplusa
this area.However,the narrowanomalyof high ½xyand
corresponding
upturnin Pxy signifytheresistive
basement
rapidly
droppingPxy at theseperiodsabout40 km inland
and middle crust. The rather shallow conductive layers
fromthecoast,plusthoseof increasing
Pxy andlow cpxy
at
beneath the Coast Range and the High Cascades and
Deschutes Basin volcanics, inferred from the TM
mode data
at shortperiods,
areevidentin thepseudosections
of Pxyand
rpxyalso.
w
OBSERVED
WC
HC
CR
WB
DB
E
very long periods (2000-5000 s) around 25 and 135 km
inland are especiallyproblematicbecausetheir lateral extent
seems so limited relative to the penetration depth of the
incident fields. One may only speculateon the existenceof
some three-dimensional electrical connection or coupling
between narrow, shallow structures and much larger ones
below or off-line. We do not observesuch problems in the
TM mode data. Wannamakeret al. [this issue]presentmore
quantitative demonstrations of the ability to fit the
observations
of Pxyandrpxywitha purelytwo-dimensional
model
-2
-1
•_ +1
•
+2
+3
+4
I
0
I
50
I
100 km
structure.
The TE moderesponseon the seafloor appearsto be much
simplerthan that on land (Figure 10 and Plate 4). (Plate 4 can
be found in the separatecolor section of this issue.) Offprofile problems are much less likely becauseof the fairly
homogeneous
near-surfacesedimentresistivity. Like the TM
mode, the TE results show high resistivities in the
lithospherenear the ridge with a substantialconductorbelow,
as well as very low resistivitiestoward the east end of the line
as it passes onto increasing sedimentary thicknesses.
Perhapsthe greatestdifferencein the two modesis apparentat
moderate to short periods (less than about 3000 s) at the
eastern three sites. Rather than remaining low from short
through long periods, the TE apparentresistivity climbs
very steeplyas T increasesfrom the shortestperiodsand is
accompaniedby very low values of impedancephase around
1000 s (Figure 11). Over moderateperiods,3000-10,000 s,
Pxy reaches
valuesof 3040 ohmmetersacross
theentiresea
Fig. 9.
Pseudosections
of transverseelectric apparentresistivity
pm andimpedance
phase•m observed
alongthelandward
portionof
the Lincoln Line. Constructionand labeling otherwiseare similar to
TM data of Figure 3. Color versionsof thesepseudosections
appear
in Plate 3 in the separatecolor sectionof this issue.
floor profile. We may be observingnearly two-dimensional
behavior
by Pxyand ½xyat leastin theuppertwo-thirds
of
the periodrange. The absenceof problemsthat besetthe land
data, namely severe static distortions and other complex
WAN•AMAKEa
ETAL.' MAOSET•tmlC OaSERVA•ONS
ACROSS
JUANDEFUCASYSTEM
OBSERVED
ultra-deep conductiveregion appearssomewhatstrongerhere
in the TE mode than in the TM,
w
SF7
SF8
•
• •
Vertical Magnetic Field Functions
10•
A
-
ß +3
•30
o +4
•2•
Pxy
3O
2O
+5
I
I
I
I
I
As implied by equation (2), a vertical magnetic field exists
only when there are lateral variations of resistivity in the
Earth. Electric current flow concentratesin good conductors
relative to poor ones, resulting in a circulation of the
anomalousmagnetic field about the structuresin accordance
with Amp•re'slaw. For example,element Mzy acrossan
isolated
+2
m
xy
o +4
+5
I
o
lOO
I
I
I
200 km
I
Fig. 10. Pseudosections
of transverseelectric apparentresistivity
of the Lincoln Line. Color versionsof these pseudosectionsappear
in Plate 4 in the separatecolor sectionof this issue.
I
i
+• ' (•- xU
two-dimensional
conductive
axis exhibits
a reversal
of sign. If resistivitymerely increasesin one direction (-y or
+y) across strike, a simple peak (positive or negative)
results. Individual resistivity structures, including threedimensionalones,tend to respondover a limited period range
in a fashion similar to the impedancephase [Wannamaker et
al., 1984]. Thus the vertical magnetic field also favors
recognitionof deep structuresbeneath complex near-surface
heterogeneity.
Furthermore,Mzy is affectedlessadversely
I
Pxyandimpedance
phase•xy observed
alongtheseafloorsegment
I
as mentioned earlier with
Figures7 and 8.
SF5 SF4 SF3
• ,•00•_00_•
+2
14,121
I
$F 4
-
than the TE mode impedanceby interruptionsof a structural
trend along strike. Specifically, the basic geometry of the
anomalies (e.g., location and shape of peaks or crossovers)
tends to remain although their amplitudes and breadths of
period range can differ.
The vertical magnetic field response along the landward
portion of the Lincoln Line (Figure 12 and Plate 5) appearsto
support and augment inferencesdrawn previously from the
transversemagnetic impedance functions. (Plate 5 can be
found in the separate color section of this issue). In the
vicinity
ofthecoastline,
astrong
positive
anomaly
inthe
T:1=+1 ' •- Ux
eeeeeeeeee•.
real
component
ofMzy,
denoted
Re(Mzy
), shows
an
eee
r
e• --'"'""'
•
w
+0
60
WB
I
.._
ß
'moomemo
i
HC
DB
I
-'
•-
1 og T
(s)
FiB.
11.
Apparent
resistivi
W
and
impedance
phase
responses
•or
both• andT• modes
atoce•bottom
site
SF4.
+4
0
three-dimensional effects, gives us hope that a model
resistivity section is achievable for the ocean results that is
largely in agreementwith both modesof the MT impedance.
Forperiods
beyond
104s, ,oxyfallsmarkedly
and c)xy
exceeds
WC
....'...................................
I
90 .
OBSERVED
CR
70 ø over the entire ocean line.
The inference
of an
,
50
Fig. 12. Pseudosections
of vertical magneticfield transferfunction
M,y, real andimaginary
components,
observed
alongthe land
segment of the Lincoln Line. Contour values are dimensionless.
Color versionsof these pseudosections
appear in Plate 5 of the
separatecolor sectionof this issue.
14,122
W•••
aT AL.:IV[AGariC
OBSERVATIONS
AcrossJU•d•D• Fuc• SYSTEM
amplitude maximum around 300 s period, below which it
weakensbut spreadslaterally. By 10,000 s, valuesexceeding
0.1 underlie the entire line. The associated imaginary
component
lm(Mzy
) is moresubdued
in itsrangethanthe
real, showingnegative amplitudesaround 10-100 s near the
coast but becoming positive at longer periods. The most
obvious candidate to explain the aforesaid vertical field
responseis the conductive seawater of the Pacific Ocean.
That hypothesismay be verified using a two-dimensional
computer algorithm, however, since the bathymetry is
orientedpredominantlyN-S and the resistivityof the seawater
is well constrained[Filloux, 1987]. Whatever portion of the
response cannot be explained by the known seawater
providesinformationon the volume of conductivesediments
offshore,in the trenchmelangefor instance,and alsoperhaps
aboutthe deepasthenosphere
below the Juande Fucaplate.
Sedimentsof the Willamette Basin affect the responseof
I
I
I
I
I
I
I'
0.75
0.00
....................
ß'.,..,.l•O
OOO•%i•_
-
.t •
-0.75
I
-2
I
-1
I
0
I
+1
I
+2
I
+3
I
+4
1og T
Fig. 14. Real andimaginaryresponsecurvesfor elementsM•x and
Mzy in Figure12 overaboutthecentralhalfof theLincoln M•ymeasured
byUURIandGSCabout
114km inland.
Line and the period range 1-100 s. The negative peak in
Re(Mzy)associated
withthisconfined
conductor
liesabout50
krn inland with the valuesbecomingpositiveeastof about y
= 75 krn. While this sign reversal probably signifies the
point of greatest sediment thickness, the presenceof two
positivepeaksat y = 85 krn and from 120-135km indicates
either another depth maximum around 100 km inland or
imaginary components,greatly exceeds that of Mzx over
almost the entire period range, thereby substantiatingour
assumptionof N-S trending geoelectricstructuresat least in
this area. Obvious in the raw data here is the negative in
perhapsan increasein basementresistivityfarthereast. The
Re(Mzy
) around10-20s signifying
theWillametteBasin.
valuesnearzeroof bothreal andimaginaryMzy overthe
central Willamette Basin at periods shorter than about 1 s
reflect less the absenceof a structural responseat shallow
levels than they do the erratic and unreliablecharacterof the
datahere due to strongculturalinterference.The lossof these
results prevents confirmation of the thickness of the
Columbia
River basalt flows in this area as inferred from the
impedancefunctions.
Examplesof the observations
taken at individual MT sites
near the westernand easternmarginsof the Willamette Basin
are plotted in Figures 13 and 14. We observe first the
excellent agreementbetween the broadbandand long-period
data in their common period range of 60-300 s at both
stations.
In Figure13, theamplitude
of Mzy, bothrealand
0.75
0.00
Also, the positive in this quantity developingbeyond 100 s
is evidence for the Pacific Ocean and other large-scale
heterogeneity. On its easternside, the effect of the basin on
Mzx is nearlyas greatas on Mzy (Figure14) andthe
coordinate rotation which minimizes
M zx yields a
geoelectric
strike
of aboutN35øE.
Thepredominance
of Mzy
returnsbelow 100 s, however, showingthat deeperstructures
are still alignednearly N-S. Following a subtle excursionto
belowzerofrom100-1000s, Re(Mzy)at thelongest
periods
displaysthe positive valuesubiquitousto the Lincoln Line.
Thereis a strong
negative
response
in Re(Mzy
) associated
with the western boundary of the Cascadesgraben located
about 145 km inland (Figure 12). The lack of a positive
counterpartto the east indicatesthat the conductivefill of the
grabendoesnot terminateas abruptlyon this side but instead
merges, or nearly does, with the shallow conductivelayer of
the DeschutesBasin region as implied from the impedance
functionsof Figure 3, 6, and 7. This responseof the graben
structure attenuates rapidly as periods reach 100 s. At
somewhatlonger periodshowever, 100-1000 s specifically,
an additional
negativeanomalyin Re(Mzy
) is resolved
in
-0.75
I
I
I
I
I
I
]'--""="
0.75
.
0.00
-0.75
I
-1
I
0
I
+1
I
+2
I
+3
+4
1 og T
Fig. 13. Real andimaginaryresponse
curvesfor elementsMzx and
Mzymeasured
by UURIandGSCabout48 km inland.Errorbarsare
one standarddeviation where they exceed the size of the plotting
symbols.
this area (Figure 15) [see EMSLAB Group, 1988]. The
anomaly is strongestbetweenthe Western and High Cascades
but negative values extend west to within 100 km of the
coast (as was seen in Figure 14). Given the persistence
westward, the responsecannot be due solely to the abrupt
transition in deep conductivityfrom west to east below the
High Cascades, most obvious in Figure 3, although that
transitionprobably contributes.We believe this negative in
Re(Mzy
) thusis providing
independent
corroboration
of the
deep layer of high conductanceextending from the central
Willamette Basin to the High Cascadesbrought to light by
the TM mode observations. The peak negative response
near y = 145 km may in part alsobe due to a conductiveaxis
associatedwith the presentvolcanic arc, a hypothesisto be
testedin subsequentpaperson interpretation.
WANdAMAKER
ETAL.:MAO•-'rOTELLURIC
OBSERVATIONS
ACROSS
Jun• DEFUCASYSXEM
0.75
i i i i
0.00
-0.75
i
i
O.75
OBSERVED
.
ß1 : '1
i
I
w
.1_
.............
I
E
SF8
SF7
+2 _1
EMSL
147
I
14,123
o
I
SF5 SF4 SF3 SF2 SF1
I
I
I
I
I
I_
I
I
I
I
I
+3
-
• +4
0.00
-0.75
I, I
-2
-1
+5
I
0
log
I
+1
T
I
+2
I
+3
+
0
I
I
'4
(s)
Fig. 15. Real andimaginaryresponse
curvesfor elementsMzx and
Mzymeasured
by University
of Oregon
andGSCabout147km inland
on the Lincoln
Line.
On the ocean bottom, we have seven geomagnetic
variationsoundingsto characterizethe vertical field response
over the Juande Fucaplate alongthe Lincoln Line. By far the
most striking feature of the observations(Figure 16 and Plate
6) is the enormous
positiveanomalyin Re(Mzy) at the
easternfour sites some 50-100 km offshore,reachingunity
at SF2, with complementarybehavior by the imaginary
component. (Plate 6 can be found in the separate color
Fig. 16.
0
100
I
I
200 km
I
Pseudosectionsof vertical magneticfield transfer function
Mzy, realandimaginary
components,
observed
alongthe seafloor
portion of the Lincoln Line. Contour values are dimensionless.
Color versions of these pseudosectionsappear in Plate 6 of the
separatecolor sectionof this issue.
section
of thisissue.)Anomalies
in Mzy ontheseafloor,we
shouldpoint out, may be amplifiedby the overlyingseawater
at peak periods because current flow induced in the water
substantially reduces the horizontal H field below it. The
aforesaid anomaly occurs near the ocean floor-continental
shelfbreakwherewater depthstartsdiminishingmostrapidly
to the east. The vertical magnetic field neverthelessshould
also carry information about structure beneath the ocean
floor, for instance,the conductivesedimentmelange. By
constrainingthe effect of the seawaterin the same manner as
for the land data, the existenceof such structuremay be
ascertainedthrough two-dimensional simulations. However,
,.., 0.75
N
q•"=
0
GGGG
O.OO
I
Themajoranomaly
in Mzy justdescribed
decreases
rapidly
0.00
around
105s (Figure
16)TherateatwhichMzyresponses
for the material
I
I
I
I
I
I
i
.
-
GGGGG
-
-0.75
I
+2
include the conductive
asthenosphere
inferredpreviouslyfrom both the TM and TE
impedancefunctionsand perhapsalsothe very deepconductor
I
I
+3
log
evidentat periodsbeyond104s. Finally,a verylargescale
three-dimensional inhomogeneity is indicated at periods
-
zu
z•
0.75
as period falls below 10,000 s and appearsto reach zero
Candidates
••I
-0.75
associated
with a conductiveaxisbeneaththe spreadingridge.
This is supportedby other EMSLAB geomagneticstations
near the ridge crestnot on the Lincoln Line and alsoby earlier
observationsand modeling of Law and Greenhouse [1981]
acrossthe ridge northof the EMSLAB array.
the mantle.
fI I I I
•
thereis no discernible
response
in M zy that may be
decayat long periodsis a functionof the resistivitybelow the
causativestructure[Wannamakeret al., 1984]. The relatively
steep decrease visible in the pseudosectionsuggeststhat
conductivematerial of regional extent exists well down into
$F 5
t
+4
T
+5
(s]
Fig. 17. Realandimaginary
response
curves
for M:xandMzyat
greaterthan104s by elementMzx, whichwe includein the
ocean bottom site
plot of the responsefunctionsat site SF5 (Figure 17). For
they exceedthe size of the plotting symbols.
SF5.
Error bars are one standard deviation
where
14,124
WANNAMAKF_.R
ET AL.:]Vff_AGNET••C OBSERVATIONS
ACROSS
JUA•DE FUCASYSTEM
the x-y-z coordinatesystem adoptedin EMSLAB, which is
that specified with equations(1) and (2), the negative values
of Re(Mzx) and positive values of Im(Mzx) observedat
periodsbeyond104s in Figure17 andoverthewholesea
floor line imply that a lateral transition to higher deep
resistivity should exist south of the sea floor array. Such
behavior in Re(Mzx) appearsat the longestperiods of the
land data as well (Figures 13-15).
CONCLUSIONS
We have collecteda magnetotelluricdata set of high quality
that is richly informative about the resistivity expressionof
the Juande Fuca subductionsystem. Upper crustalconductive
units apparent upon inspection of the land observations
include massive sedimentaryfill of the Willamette Basin as
well as layers beneaththe Coast Range, from the Westernto
the High Cascades,and in the DeschutesBasin area. All are
underlainby resistivebasementand middle crust. Especially
relevant to the subductionprocessbeneath the continentwe
believe is a conductive,roughly horizontallayer at a depth of
a few tens of kilometers
in otherwise
resistive lower crust and
upper mantle which extends eastward from the coast and
changes abruptly at the High Cascades. A fundamental
transition in deep resistivity to that beneath the Deschutes
Basin occurs below the active High Cascadesvolcanic arc.
Seafloor MT soundingsover the Juande Fucaplate near the
spreadingridge imply a highly conductingregion at depthsof
some tens of kilometers that may represent the
asthenosphere. Toward the coast, a conductive wedge of
sedimentsincreasinglyinfluencesthe MT and geomagnetic
soundings. Probably our best prospect for determining an
accurate resistivity model from the observed data is to
interpret the data carefully assuming a two-dimensional
resistivity variation with a N-S strike. For such an analysis,
the TM mode impedancefunctions should be emphasized
since they are more robust to common deviations from the
two-dimensionalassumption,with the vertical magneticfield
and the TE impedanceexamined also for consistencyand
additional
information.
It
should
be
clear
from
our
discussions,however, that certain important featuresof the
data ultimately will be understoodonly in termsof fully threedimensional geometries.
Valuable
lessons have been learned
about the collection
of
magnetotelluricdata in complex natural environments. First,
observations of high statistical accuracy are required to
resolve certain key features of the resistivity structure.The
anomalyin Pyx around30-50s periodbeneaththe Coast
Range, which is the most likely manifestationof subduction
structurein this region, is only of several degreesamplitude.
Luckily, the responseoccurredat periods where the natural
fields are strong and cultural interference is relatively
subdued. We cannotcount on suchgood forme always, and
high quality data at shorter periods are important more
generallyfor understanding
upper crustalheterogeneity.It is
therefore highly desireablefor the future that the broadband
MT
instrumentation incorporate full remote reference
capability (sensor separations of 10 km or more) and
statisticallyrobustprocessingto ensuresoundingcurvesthat
are accurateover all periods. Second,the rate of collectionof
MT data by research institutions is painfully slow and
exhaustingdue again to the stateof the field systems.Highly
desireable improvements in broadband instrumentation
include acquisition of the mid- and low-band data
simultaneously,completein-field data processingto yield a
finished sounding on-site, and the ability to deploy
additional
electricfieldbipoles(especiallyEy) independent
of the magneticfield apparatus. Similarly, improvementsin
sea floor instrumentation to take advantage of modern
electronicdevelopmentswould increaseour ability to utilize
MT in the oceans. While better data analysishas increased
the usableperiod range on the sea floor, the shortestperiods
attainableat presentare limited by magnetometersensitivity.
Implementation of the foregoing recommendations is
essentialwe believe if the productivity of magnetotelluric
profiling is to approachthat of reflectionseismicprofiling.
Acknowledgments. We wish to acknowledge the following
sourcesof funding supportwhich made the project possible:U.S.
National Science Foundation, grants EAR-8410638 and EAR8512895; Natural Sciences and Engineering Research Council of
Canada, Collaborative Special Grant A6892; Geological Survey of
Canada;Ministry of Education,Scienceand Culture of Japan,Grantin-Aid for OverseasScientificSurvey61043014; AustralianResearch
Grants Scheme,Flinders University and Australian Departmentof
Science; Green's Foundation of the Scripps Institute of
Oceanography;
and AT & T Bell Laboratories.Second,we are very
grateful to the U.S. Forest Service, Oregon State Department of
Forestry, and the many private timber companiesand land owners
who permittedaccessto measurements
on their properties. Finally,
we are indebted to the numerous graduate studentsand technical
assistantswho worked on the field acquisitionand reductionof our
data. This is GeologicalSurveyof Canadacontribution23388.
REFERENCES
Boehl, J. E., F. X. Bostick, Jr., and H. W. Smith, An applicationof
the Hilbert transform to the magnetotelluricmethod, report, 98
pp., Electr. Geophys.Res. Lab., University of Tex. at Austin,
1977.
Bostick, F. X., Jr., Electromagnetic array profiling (EMAP),
extendedabstractof paper presentedat the 56th annualmeeting,
Soc. of Explor. Geophys.,Houston,Tex., 1986.
Chave,
A.
D.,
and
D.
J. Thomson,
Some
comments
on
magnetotelluricresponsefunction estimation, J. Geophys.Res.,
this issue.
Chave, A.D.,
D. J. Thomson, and M. E. Ander, On the robust
estimationof power spectra,coherences,and transferfunctions,J.
Geophys.Res., 92, 633-648, 1987.
Chave, A.D., J. H. Filloux, D. S. Luther, L. K. Law, and A. White,
Observationsof motional electromagneticfields during EMSLAB,
J. Geophys.Res., this issue.
Clarke, J., T. D. Gamble, W. M. Goubou, R. H. Koch, and R.
Miracky, Remote reference magnetotellurics: Equipment and
measurementprocedures,Geophys.Prospect,31, 149-170, 1983.
Egbert, G. D., and J. R. Booker, Robust estimation of geomagnetic
transfer functions, Geophys.J. Roy. Astron. Soc., 87, 173-194,
1986.
Egbert, G. D., and J. R. Booker, Multivariate analysis of
geomagneticarray data, 1, The responsespace,J. Geophys.Res.,
this issue.
EMSLAB Group, The EMSLAB electromagnetic sounding
experiment,Eos Trans. AGU, 69, 89, 98-99, 1988.
Filloux, J. H., Instrumentationand experimentalmethodsfor oceanic
studies,in Geomagnetism,1, edited by J. A. Jacobs,Academic,
San Diego, Calif., pp. 143-248, 1987.
Gamble,T. D., W. M. Goubau,and J. Clarke, Magnetotelluricswith a
remotereference,Geophysics,44, 53-68, 1979a.
Gamble, T. D., W. M. Goubau, and J. Clarke, Error analysisfor
remote reference magnetotellurics: Geophysics, 44, 959-968,
1979b.
Gough, D. I., D. McA. McKirdy, D. V. Woods, and H. Geiger,
Conductive
structures
and tectonics
array, J. Geophys.Res., this issue.
beneath
the EMSLAB
land
WANNAMAKER
ETAL.:MAG•••C
OBSERVATIONS
ACROSS
JUAN
DEFVCA
SYSTEM
Jiracek,G. R., J. H. Curtis, J. Ramirez, M. Martinez, and J. Romo,
Two-dimensional magnetotelluric inversion modeling of the
EMSLAB Lincoln Line, J. Geophys.Res., this issue.
Jones,A. G., The problemof currentchanneling:a critical review,
Geophys.Surv., 6, 79-122, 1983.
Jones,A. G., A.D. Chave,G. D. Egbert,D. R. Auld and K. Bahr, A
comparisonof techniquesfor magnetotelluricresponsefunction
estimation,J. Geophys.Res., this issue.
Kurtz, R. D., J. M. DeLaurier,and J. C. Gupta, A magnetotelluric
soundingacrossVancouverIsland detectsthe subductingJuande
Fuca plate, Nature, 321, 596-599, 1986.
Law, L. K., andJ.P. Greenhouse,
Geomagnetic
variationsoundingof
the asthenosphere
beneaththe Juande Fuca Ridge, J. Geophys.
Res., 86, 967-978,
1981.
Livelybrooks,D. W., W. W. Clingman,J. T. Rygh, S. A. Urquhart,
and H. S. Waft, A magnetotelluricstudy of the High Cascades
grabenin centralOregon,J. Geophys.Res., this issue.
Petiau, G., and A. Dupis, Noise, temperaturecoefficient,and long
time stability of electrodesfor telluric observations,Geophys.
Prospect, 28, 792-804, 1980.
Price, A. T., The theory of geomagneticinduction,Phys. Earth
Planet. Inter., 7, 227-233, 1973.
Ranganayaki, R. P., and T. R. Madden, Generalized thin sheet
analysisin magnetotellurics:An extensionof Price's analysis,
Geophys.J. Roy. Astron. Soc., 60, 445-457, 1980.
Sims, W. E., F. X. Bostick, Jr., and H. W. Smith, The estimation of
magnetotelluric impedance tensor elements from measured data,
Geophysics, 36, 938-942, 1971.
Stodt, J. A., Noise analysisfor conventionaland remote reference
magnetotellurics,Ph.D. thesis, University of Utah, Salt Lake
City, 220 pp., 1983.
Vozoff, K. (Ed.), MagnetotelluricMethods,Geophys.ReprintSeries,
vol. 5, 800 pp., Soc. of Explor Geophys.,Tulsa, Okla., 1986.
Waft, H. S., J. T. Rygh, D. W. Livelybrooks,
andW. W. Clingman,
Results of a magnetotellurictraverseacrosswestern Oregon:
Crustal resistivity structure and the subductionof the Juan de Fuca
plate, Earth Planet. Sci. Lett., 87, 313-324, 1988.
Wannamaker, P. E., G. W. Hohmann, and S. H. Ward,
Magnetotelluricresponses
of three-dimensional
bodiesin layered
earths, Geophysics,49, 1517-1533, 1984.
Wannamaker,P. E., J. R. Booker, A. G. Jones,J. H. Filloux, A.D.
Chave, H. S. Waft, and L. K. Law, Resistivity cross section
through the Juan de Fuca subductionsystem and its tectonic
implications,J. Geophys.Res.,this issue.
Wight, D. E., andF. X. Bostick,Jr., Cascadedecimation-a
technique
for real time estimationof powerspectra,in ProceedingsIEEE
International Conference on Acoustic Speech and Signal
Processing,pp. 626-629, IEEE, New York, 1980.
Word, D. R., H. W. Smith, and F. X. Bostick, Jr., Crustal
investigations
by the magnetotelluric
impedancemethod, in The
Structureand PhysicalPropertiesof the Earth's crust,Geophys.
14,125
Monogr.Ser., vol. 14, editedby J. G. Heacock, pp. 145-156,
AGU Washington,D.C., 1971.
Young, C. T., and M. R. Kitchen,A magnetotelluric
profile in the
CoastRange of Oregon,J. Geophys.Res., this issue.
Young, C. T., J. R. Booker, R. Fernandez, G. R. Jiracek, M.
Martinez, J. C. Rogers, J. A. Stodt, H. S. Waft, and P. E.
Wannamaker,Verificationof five magnetotelluric
systemsin the
mini-EMSLAB experiment,Geophysics,53, 553-557, 1988.
J. R. Booker,Geophysics
Program,AK-50, Universityof
Washington, Seattle, WA 98195.
A.D. Chave, AT&T Bell Laboratories,Rm 1E-444, 600 Mountain
Ave., Murray Hill, NJ, 07974.
G. D. Egbert,Collegeof Oceanography,
OregonStateUniversity,
Corvallis, OR, 97331.
J. H. Filloux, ScrippsInstitutionof Oceanography,
MS A-030,
Universityof Californiaat San Diego, La Jolla, CA 92093.
A. W. Green,Jr., U.S. GeologicalSurvey,DenverFederalCenter,
P.O. Box 25046, Denver, CO 80225.
G. R. Jiracek,Departmentof GeologicalSciences,
SanDiegoState
University,San Diego, CA 92182.
A. G. Jones,Lithospheric
Geophysics,
GeologicalSurveyof
Canada, 1 ObservatoryCrescent,Ottawa, Ontario, Canada,K1A 0Y3.
L. K. Law, PacificGeoscience
Center,Geological
Surveyof
Canada,P.O. Box 6000, Sidney,BritishColumbia,CanadaV8L 4BZ
M. Martinez G., Center for Scientific Research and Graduate
Educationof Ensenada(CICESE),AvenidaEspinoza843, 2732Ensenada,Baja Calffomia, Mexico.
J. S. Segawa,OceanResearch
Institute,Universityof Tokyo, 115-1 MinamidaiNakano,Tokyo 164, Japan.
J. A. Stodt,Universityof UtahResearch
Institute,391-C Chipeta
Way, Salt Lake City, UT 84108.
P. Tarits,Institutede Physiquedu Globe,Laboratoirede
Geomagnetisme,
4 PlaceJussieu,75252 Paris Cedex05, France.
l-I
•q_
Waft
---
,
I3o.rnartrnont
rff Gor•lr•ovoJ
ß
11nlve_r•itv
.....
J
rff/•roo•n
--o
Eugene, OR 97403.
P. E. Wannamaker,Universityof Utah ResearchInstitute,391-C
ChipetaWay, Salt Lake City, UT 84108.
A. White, Schoolof Earth Science,FlindersUniversityof South
Australia, Bedford Park, S. A. 5042, Australia
C. T. Young, Departmentof Geologyand GeologicalEngineering,
Michigan Technical University, Houghton,MI, 49931.
T. Yukutake,EarthquakeResearchInstitute,Universityof Tokyo,
Bunkyo-ku, Tokyo 113, Japan.
(ReceivedMay 13, 1988;
revisedApril 3, 1989;
acceptedApril 3, 1989.)
WATV•A•,A• ET A•.: lVlAO•'TOTm.CUlUC
O,•Sm•VATiOlVS
Across IUA•TDl• FUCAS¾STg•
.SER
ED
B
WC
CR
HC
14,271
DB
/%
,,,
-.2
-
,.
-1
1.o
. !
..
ß
ß
o
1o
+1
1oo
+2
'+3
+4
1o0o
_ I
I i
l ....l I!1 ....I i !
Ill,l
I,I
], !
I
!
..Ill
I {
I II.I
I Illl
I
.I.... i
i
(deg)
-
75
- 65
F-_-•.•
- 45
i[mTqz] _
25
15
+4
i
I
'
0
W
1.........
.50
I
I.
i
I
I00
km
E
Plate1. [Wannarnaker
et al. ]. Pseudosections
of transverse
magnetic
apparent
resistivity
P•x and
impedance
phase• observed
alongthelandward
portion
of theLincoln
Line.
Pyx
14,272
WA•A•vIA• E?At,.:MAo•'ronu.UnucOUSmVAVIO•/S
Across.IOA•I• FOCA
S¾s•
OBSERVED
SF7
!
•*2 •
SF5 4 3
I '
I
i
2 !
I
(œ.m)
I
- 1.0
10
Pyx
•...... - I00
,4
l
*5
!
_J__
I
i .... 1
1_
•
- 1000
(deg)
-55
-25
0
I00
260km
W
E
Plate2. [Wannamaker
et al.]. Pseudosections
of transverse
magneticapparent
resistivityP•xand
impedance
phase•b•xobserved
alongtheseafloorportionof theLincolnLine. Notedifferences
in
horizontalscaleand periodrangebetweenland and seafloor results.
WAn'IV•
ET AL.: MAO•TBTOnU.LLtr•c
O•SmVA•IO•tSA•OSS :3UA•t
DE FUCASYSTEM
14,273
OBSERVED
CR
WB
WC
HC
DB
(,-•' m)
-2
-
I.O
-1
1o
1oo
-
+4
Pxy
1ooo
(deg)
-2
-1
75
- 65
+I
xy
+2
-
+3
25
-15
+4
I
,,
!_
I
,
i
i
0
50
W
I
1
100 km
E
Plate3. [Wannamaker
et al.]. Pseudosections
of transverse
electricapparent
resistivityPxyand
impedance
phase•xyobserved
alongthelandward
portion
oftheLincoln
Line.
14,274
WA•U••
ETAL: MAO1,/L•T••½ OgSlmVATIO•S
ACROSS
JUAn/D•
FUCA
S¾•
OBSERVED
SF8
SF7
•
SF5 4 3 2 1
• ! •.........
•....
[•• • - i.0
p
xy
•===,-
lOO0
(deg)
-65
•3
•"- , -55
-*4
.......
:35
-'25
-15
+5
o
,6o
•6o•.,
W
E
Plate4. [Wannamaker
et al.]. Pseudosections
of transverse
electricapparent
resistivityPx• and
impedance
phase•Jx•observed
alongtheseafloorportionof theLincolnLine. Notedifferences
in
horizontalscaleand periodrangebetweenland and seafloor results.
WAlVNAlvIA• ET AL.: MAO••tmiC
O•S•VATIONSACROSS
JUANI• FUCASYsamuvi
14,275
BSERVED
CR
I
-2
I
I
I
!
ill
WB
II
I
III
I
1 I
I
WC
I
I
I
Ill
I I
I
II
HC DB
I
II!ll
1
!
I
(%)
I
-,•
- 60
-1
-
4O
+1
•
+2
--20
N
+3
+4
- '60
__. I
I I
.... I'1
!
I I
I 111 I I 1 ....II
I Iii
I!
I I. I i I
I I
I
I,,I..l i i
I !'"1"1"i"'1I I
I I
I
III
I.,.11...i I I11..i ...... I
I I I II
IIIIII
I
I
i
I
I
I
(%)
-2
-1
60
40
2o 3
O
+1
o
+2
- -20 N
+3
- -60
+4
I
I I
I
I III
I1
I
111
i
!
0
50
i i ! I
I,,I
i
,I
i I I l,, I
II
I..i Ill
I
...1.
l_
I
1
1oo km
W
Plate5.
I
E
[Wannamaker
et al.]. Pseudosections
of vertical
magnetic
fieldtransfer
elementM,y, realand
imaginaryparts.observedalongthe landwardportionof the LincolnLine.
14,276
WANdAMAKER
ET AL: 1VI_AGN•(7I'EI•URIC
OBSERVATIONS
AcrossJUA•D•- FucAS¾s•
OBSERVED
SF8
SF7
I
SF54
I
3
•
I
21
•
'i"•
90
.,
6O
3O
O
-30
i
N
-60
-90
(•)
+2 œ1
90
•-
-30
-
-6O
-90
O
w
100
200 k m
E
Plate6. [Wannarnaker
etal.].Pseudosections
ofvertical
magnetic
fieldtransfer
elementM,y,realand
imaginaryparts,observedalongthe seafloor portionof the LincolnLine. Note differencesin horizontal
scaleand periodrangebetweenland and seafloor results.