A deep tow magnetic survey of Middle Valley,
Juan de Fuca Ridge
Jeffrey S. GeeScripps Institution of Oceanography, La Jolla, California 92093, USA ( [email protected])
Spahr C. WebbLamont Doherty Earth Observatory, Palisades, New York 10964, USA ([email protected])
Jeffrey Ridgway, Hubert Staudigel, and Mark A. ZumbergeScripps Institution of Oceanography, La Jolla, California 92093, USA ( [email protected]; [email protected];
[1] Abstract: We report here results from a deep tow magnetic survey over Middle Valley, Juan de
Fuca Ridge. A series of track lines are combined to generate a high-resolution map of the magnetic
field anomaly within a 10 � 12 km region surrounding the Bent Hill massive sulfide (BHMS)
deposit. A uniformly magnetized body (5 A/m) with a cross section approximating the body inferred
from Ocean Drilling Program (ODP) drilling can account for the observed near-bottom magnetic
anomaly amplitude. Assuming this magnetization is entirely induced, the average susceptibility (0.11
SI) corresponds to �3.5% magnetite + pyrrhotite by volume, consistent with the abundance of thesephases observed in drill core samples. However, this uniform magnetization model significantly
underestimates the magnetic anomaly measured a few meters above the seafloor by submersible,
indicating that the upper portion of the sulfide mound must have a significantly higher magnetization
(�10% magnetite + pyrrhotite) than at deeper levels. On a larger scale, the near-bottom magneticanomaly data show that basement magnetizations are not uniformly near zero, as had been inferred
from analysis of the sea surface anomaly pattern. We interpret this heterogeneity as reflecting
primarily differences in the degree of hydrothermal alteration. Our results highlight the potential of
magnetic anomaly data for characterizing hydrothermal deposits where extensive drill core sampling
is not available.
Keywords: Magnetic anomaly; hydrothermal alteration; mid-ocean ridge.
Index terms: Magnetic anomaly modeling; spatial variations attributed to seafloor spreading; hydrothermal systems;
midocean ridge processes.
Received April 24, 2001; Revised July 20, 2001; Accepted July 25, 2001; Published November 15, 2001.
Gee, J., S. Webb, J. Ridgway, H. Staudigel, and M. Zumberge, A deep tow magnetic survey of Middle Valley,
Juan de Fuca Ridge, Geochem. Geophys. Geosyst., 2, 10.1029/2001GC000170, 2001.
1. Introduction
[2] At sedimented ridges, hydrothermal circu-
lation is confined beneath relatively imperme-
able sediments, which restrict heat and fluid
flux out of the system [Davis and Fisher,
1994], and consequently discharge is focused
G3G3GeochemistryGeophysics
Geosystems
Published by AGU and the Geochemical Society
AN ELECTRONIC JOURNAL OF THE EARTH SCIENCES
GeochemistryGeophysics
Geosystems
Article
Volume 2
November 15, 2001
Paper number 2001GC000170
ISSN: 1525-2027
Copyright 2001 by the American Geophysical Union
into widely separated vent fields. Hydrothermal
activity in these settings can produce signifi-
cant massive sulfide deposits that may be of
economic significance. Key factors leading to
the development of massive sulfide deposits
include the interaction of seawater with hot
oceanic crust and often the subsequent reaction
of these fluids with organic-rich sediments.
Such ‘‘Kieslager’’ or ‘‘Besshi’’ type massive
sulfide deposits are commonly found in the
geological record and may be extremely large
(e.g., Ducktown, Tennessee; Windy Craggy,
British Columbia). The most prominent mod-
ern Besshi-type deposits are found in the
Guaymas basin [Gieskes et al., 1982], the
Escanaba Trough (Gorda Ridge [Zierenberg et
al., 1993]) and in Middle Valley (Juan de Fuca
Ridge [Davis et al., 1987]).
[3] Middle Valley, a sediment covered axial
valley at the northern end of the Juan de Fuca
Ridge (Figure 1), has been the target of exten-
sive geophysical investigations and was a
major focus of Ocean Drilling Program (ODP)
Legs 139 and 169. The rift valley has been
buried by from 200 m to more than 1000 m of
turbiditic and hemipelagic sediments derived
from the adjacent continental margin during
the Pleistocene sea level low stand. A number
of hydrothermal centers have been documented
in Middle Valley. In this paper, we focus
primarily on the inactive massive sulfide
mound located south of Bent Hill, a 500-m
diameter feature that rises �60 m above thesurrounding seafloor. The Bent Hill massive
sulfide (BHMS) deposits were produced by
high-temperature fluids (�350–4008C [Good-fellow and Peter, 1994; Peter et al., 1994]),
with a significant volume of the deposit likely
already formed 200–140 kyr ago [Mottl et al.,
1994]. Despite its proximity to Bent Hill, a
variety of lines of evidence indicate that the
BHMS deposit is not genetically related to the
intrusions that resulted in the uplift of Bent Hill
[Mottl et al., 1994].
[4] We collected near-bottom magnetic anom-
aly data over the BHMS in order to corrobo-
rate the overall dimensions and hemispherical
shape of the deposit inferred from drill core
data [Fouqet et al., 1998] and to provide
additional constraints on the integrated proper-
ties of this ore deposit. The reduced to the
pole magnetic anomaly is nearly circular, with
a width of 150 m EW and 190 m NS at half
peak value. Forward magnetic models approx-
imating the hemispherical source and with an
induced magnetization of �5 A/m (suscepti-bility 0.11 SI) match the observed anomaly.
Moreover, this same source model with a
density contrast of 2300 kg/m3 yields a grav-
ity signal of 3.5 mgal, close to estimates from
a previous on bottom gravity survey [Ballu et
al., 1998]. Comparison of our near-bottom
magnetic anomaly data with a previous sub-
mersible survey a few meters above the sul-
fide mound [Tivey, 1994a] indicates that the
upper portion of the sulfide mound must have
a significantly higher magnetization than at
deeper levels.
[5] Near-bottom magnetic anomaly variations
away from Bent Hill also provide insights
into the effects of hydrothermal alteration on
the magnetization of the seafloor basalts.
Analyses of sea surface magnetic surveys
suggest near-zero magnetization within Mid-
dle Valley [Currie and Davis, 1994; Davis
and Lister, 1977a]. The low magnetizations
in this and other sedimented ridge settings
have been attributed to either the destruction
of remanence-carrying phases during exten-
sive hydrothermal alteration or to the demag-
netizing effect of elevated temperatures
[Larson et al., 1972; Levi and Riddihough,
1986; Currie and Davis, 1994]. Our deep
tow magnetometer results indicate finite but
small magnetization throughout much of the
valley. Together with other geophysical data,
these results allow us to investigate the
relative importance of temperature and alter-
GeochemistryGeophysicsGeosystems G3G3 gee et al.: deep tow magnetic survey 2001GC000170
ation in affecting magnetization in the
region.
2. Middle Valley
[6] Middle Valley is the northernmost segment
of the Juan de Fuca Ridge (Figure 1). It is
terminated by the Sovanco transform fault to
the north, with the NE-trending Nootka fault
forming a tectonically complicated triple junc-
tion between the Pacific, Explorer, and Juan de
Fuca plates [Davis and Currie, 1993]. The full
spreading rate of the Juan de Fuca Ridge is 6
cm/yr, but most of the recent extension at the
northern end of the Juan de Fuca Ridge is
focused in West Valley, just west of Middle
Valley as a result of a recent westward ridge
jump. Normal faults mark the valley boundaries
to the east and west. An additional west-facing
normal fault in the center of the valley sepa-
rates the central rift graben from a shallow
basement bench to the east [Davis and Villin-
ger, 1992]. The thickness of the turbidite sedi-
ments ranges from over 1 km to the north to
less than a hundred meters to the south with
thinner sediments found to the east of the
central fault [Davis and Villinger, 1992].
0 5km
SOVANCO F Z
STUDY AREA
WESTVALLEY
ENDEAVOURSEGMENT
(Bro
ad Z
one)
(Broad Zone)
NOOTKA
FAULT
50o
130o 127o
48o
Cascadia
Subduction
Zone
MIDDLEVALLEY
PacificPlate
Juan de FucaPlate
ExplorerPlate
48o25'N
48o30'N
128o55'W 128o50'W 128o45'W 128o40'W 128o35'W
-240
0
-240
0
-240
0
-240
0
-240
0
-240
0
-240
0
-230
0
-240
0
-2500
-250
0
-230
0
-240
0
Dead Dog858/1036
Bent Hill856/1035
857855
Figure 1. Bathymetry and location of ODP drill sites (circles) in Middle Valley. Box indicates the areacontaining the near bottom survey lines. Inset shows simplified tectonic map with location of Middle Valley.
GeochemistryGeophysicsGeosystems G3G3 gee et al.: deep tow magnetic survey 2001GC000170
[7] During ODP Legs 139 and 169, four sites
were drilled in the eastern part of the valley
to study various aspects of the hydrothermal
system beneath the sediments [Davis et al.,
1992; Fouqet et al., 1998]. Site 855, located
near the eastern boundary fault, recovered
basaltic flows with minor alteration that char-
acterize the low-temperature fluid recharge
for the hydrothermal circulation. Sites 856
and 1035 were dedicated primarily to sam-
pling the relict high temperature (350–
4008C) BHMS deposit. Results from two
sampling transects revealed an �100 m thick,hemispherical shaped massive sulfide deposit
underlain by a copper rich feeder zone.
Mineralogical, textural, and geochemical
studies indicate that the BHMS deposit
formed by mineral precipitation near or
above the seafloor, rather than by replace-
ment of sediments.
[8] Site 858 is within the Dead Dog Vent field,
a large active vent field discharging fluids as
hot as 2768C from at least 20 active vents[Ames et al., 1993]. The vent field overlies a
small basement high that acts to focus hydro-
thermal fluids to the Dead Dog vent field
[Davis and Fisher, 1994]. Site 857, located
�1.5 km south of the Dead Dog vent field,was designed to reach the permeable hydro-
logical basement where high-temperature cir-
culation was inferred to occur beneath an intact
sediment cover. Drilling revealed that the per-
ceived basement-sediment interface in seismic
refraction records [Rohr and Schmidt, 1994]
was the top of an extensive sequence of inter-
bedded sills and turbidite sediments that must
underlie much of the valley [Langseth and
Becker, 1994].
[9] The location of active high-temperature
vents is evident in a map of heat flow in the
valley (Figure 2). Although no active venting
was found at the BHMS deposit, the smaller
ODP and Lone Star vents 300–400 m south
were actively venting high-temperature fluids
(�2658C). At this site and at the Dead Dogvent field, a shallow subseafloor silicification
zone forms a caprock for the present hydro-
thermal reservoir. This horizon divides the
hydrothermal system into two nearly independ-
ent systems above and below the caprock [Stein
et al., 1998; Stein and Fisher, 2001]. As noted
in previous regional studies [e.g., Davis and
Lister, 1977a], heat flow in the valley is
approximately inversely proportional to sedi-
ment thickness, suggesting that temperature
variations at the basement-sediment interface
are likely to be subdued as a result of circu-
lation within the relatively permeable basement
[Davis and Villinger, 1992; Bessler et al.,
1994].
.4
W / m2.20 .4 .6 .8 1 2 448o30'N
128o40'W128o45'W
4.9
855
1.4
.6.6
858/1036
856/1035
857
.8
.8
1.0
.6.6
.4.2
.6
24
6.6
.8.8
.6
.4
48o28'N
48o26'N
48o24'N
Figure 2. Heat flow data from Middle Valley.Active high temperature vents in the Dead Dog andBent Hill areas are accompanied by high heat flow.Location of ODP drill sites indicated by circles.Contour interval 0.2 for values <1.0 W/m2 and 2 forvalues above 2.0 W/m2. Modified from Davis andVillinger [1992].
GeochemistryGeophysicsGeosystems G3G3 gee et al.: deep tow magnetic survey 2001GC000170
[10] A recent ocean bottom seismometer study
tracked the seismicity (due to thermal contrac-
tion) associated with the mining of heat beneath
the Dead Dog vent field to investigate the
source region of the vent fluids (C. E. Golden,
S. C. Webb, and R. A. Sohn, Hydrothermal
microearthquake swarms beneath active vents
at Middle Valley, Juan de Fuca ridge, submitted
to Journal of Geophysical Research, 2001).
Earthquakes occurred primarily in large, tightly
grouped swarms, but the swarms extended
many kilometers from the vent field forming
a ramp-like structure that deepened to the north
to depths of 2.5 km (below seafloor) away from
the vent field. These results suggest a large
volume of rock provides heat for the fluids at
the vent field.
3. Deep Tow Survey
[11] The gravity or magnetic signal from a two-
dimensional feature with a wavelength of L is
attenuated by exp(�2pz/L), where z is the
height of the observer above the source. Thus
narrow geologic features with wavelengths
shorter than the ocean depth will be so strongly
attenuated as to be virtually undetectable from
surface surveys. To examine the structure of the
narrow (�100 m) BHMS deposit, we thereforetowed an instrument package with both a
magnetometer and gravimeter near the seafloor.
The towed deep-ocean gravimeter was devel-
oped to enable high-resolution surveys of sea-
floor geological features [Zumberge et al.,
1997]. The meter is towed 50–100 meters
above the seafloor at a speed of 1.5 knots
(0.77 m/s). The gravity results will be reported
elsewhere. Here we report results obtained
from a magnetometer towed above the grav-
imeter package.
[12] Both the magnetometer and gravimeter
were deployed from the ship’s 1.73 cm
(0.680 inches) electrically conducting armored
cable, via a telemetry interface (called the
Deep Sea Instrument Interface, or DSII)
(Figure 3). Control communications, the grav-
ity and magnetic data, and acoustic naviga-
tion signals were multiplexed onto this cable
along with power to the instruments. The
DSII navigation package carried multiplexers
and power separation electronics along with
an acoustic interrogation transponder (10–12
KHz bandwidth) that was used to navigate
the towed sensors. Acoustic transponders
placed on the ocean bottom in the survey
area exchanged signals with a transducer
on the DSII to provide precise navigation
of the package. The package also carried
a CTD package. The height of the instrument
over the seafloor was measured with a 3.5
kHz down-looking echo sounder on the DSII.
The vertical position of the instrument was
determined by recording pressure and convert-
ing to a depth assuming a seawater density
profile.
[13] The deep tow magnetometer system was
based on a low cost miniature angular orienta-
tion sensor built by Applied Physics Systems
(model 544). The system consists of a three
axis fluxgate magnetometer and a three axis
accelerometer and was designed to provide roll,
pitch, and azimuth from combinations of the
accelerometer and magnetometer data. The
specified fluxgate noise level and alignment
are 0.4 nT and ±0.28. The direct magnetometeroutput (sampled at 2–3 times per second)
proved too noisy for direct detection of mag-
netic anomalies because of excessive motion of
the magnetometer package during towing. To
improve resolution of the magnetic anomalies
the data were smoothed by applying a 100 s
(full width, 6 sigma) Gaussian filter to the root
mean square total field calculated from the raw
data.
[14] The small size of the Bent Hill sulfide body
and of other targets within Middle Valley made
GeochemistryGeophysicsGeosystems G3G3 gee et al.: deep tow magnetic survey 2001GC000170
accurate navigation of the tow body critical to
the interpretation of the data. We deployed
acoustic transponders on short moorings (100
m) at four sites within Middle Valley at the
beginning of the cruise. The mooring locations
were determined to within an absolute accuracy
of a few meters by combining P-Code GPS
ship locations and acoustic ranges between the
ship and moorings obtained during an initial
survey of the transponders. Throughout the
cruise we navigated the ship and the tow fish
by acoustic ranging between the DSII deep tow
navigation package, the moorings, and a trans-
ducer mounted on the ship’s hull. Ship and
deep tow package positions were determined
roughly every 1 min with an accuracy of a few
meters during each tow.
[15] The magnetic data in this paper were
obtained from three lowerings of the tow
vehicle during a cruise on the R/V Thompson
in July 1998. A constant depth of observation
of 2420 m for the DSII was chosen for the bulk
of the survey, with the intent that the gravimeter
would never be closer than �20 m to theseafloor. The magnetometer was flown �30m above the DSII (Figure 3), so its tow depth
was mostly near 2390 m. A watch stander
monitored the depth of the DSII and controlled
the ship’s 1.73 cm traction winch to keep the
magnetometer on tether with float
0.680" electro-mechanical cable
Interface Package Down-looking sonar Transponder Interrogator Pressure Gauge
gravimeter
nylon tetherTransponder
Figure 3. Tow configuration showing relative positions of the towed magnetometer, gravimeter, interfacepackage (DSII).
GeochemistryGeophysicsGeosystems G3G3 gee et al.: deep tow magnetic survey 2001GC000170
DSII depth constant as changes in the ship’s
speed caused slow fluctuations in wire angle.
This was generally successful within a window
of less than 10 m except during turns when the
excursion in depth could exceed a few hundred
meters. These depth excursions are not a sig-
nificant problem for interpretation of the mag-
netometer data.
[16] We discovered through an analysis of cross
over errors in the resulting smoothed data
stream that there was an apparent dependence
of the total field value on the azimuth (u) of themagnetometer. This heading dependence can
be attributed to small errors (<0.2%) in the gain
and orientation (<0.058) of the individual mag-netometer components. In the data shown in
this paper the azimuthal error has been
removed by fitting sine and cosine components
in azimuth to the data at the track crossing
points, using only those crossovers where the
difference in instrument elevation was <50 m.
A small second harmonic in azimuth depend-
ence was also observed in the cross over point
data and terms in (cos(2u), sin(2u)) were fit andremoved from the data. RMS cross over errors
were reduced from 113 to 32 nT for tow 3 and
from 197 to 48 nT for the first two tows. The
first and second tows also exhibited an offset
from tow 3 of 103 nT. Fortunately, the large
number of track line crossings in this data set at
a wide variety of headings (Figure 4) allows the
azimuthal error to be accurately determined.
The procedure is essentially a mathematical
system of ‘‘boxing’’ the compass and is a
useful technique to apply to any towed mag-
netic survey data. The IGRF field value at Bent
Hill (48.458N, 128.688W) for 1998.6 of 54500nT was subtracted from all total field data to
calculate the magnetic anomalies.
4. Results Near Bent Hill
[17] The observed magnetic anomaly data from
the combination of the three track lines is
shown in Figure 4. The ‘‘race track’’ character
of the track lines is the result of trying to fly
closely spaced track lines repeatedly over the
vicinity of the Bent Hill massive sulfide
(BHMS) deposit while holding long intervals
of constant heading to obtain stable Eotvos
corrections for the gravimeter. The magnetic
anomaly is negative to the west of the eastern
bounding faults, implying low magnetization
beneath the sediments. The color scale in this
figure has been chosen to emphasize anomaly
variations within the sediment-filled valley and
saturates above the strongly magnetized rocks
outside of the graben that are associated with
positive anomalies of several thousand nT. The
Bent Hill sulfide mound is visible as a small
local anomaly maximum near 1288410W,408260N.
[18] The magnetic anomaly associated with the
BHMS deposit is more easily seen in a higher
resolution figure of the magnetic anomaly
(Figure 5). In this figure the field has been
reduced to the pole, assuming the magnetiza-
tion is entirely induced (i.e., ambient field and
induced magnetization both have an azimuth of
0208 and an inclination of 698). The anomalyhas been smoothed (gridded from continuous
curvature splines in tension [Smith and Wessel,
1990]) and interpolated to 10 m spacing. The
amplitude (maximum �525 nT) and shape ofthe BHMS anomaly is well resolved by the ten
track lines crossing the body. The outlines of
the BHMS, ODP, and Bent Hill have been
drawn on Figure 5. The reduction to the pole
realigns the magnetic anomaly to lie directly
over the top (bathymetric expression) of the
BHMS mound. The magnetic field profiles
observed along north-south and east-west slices
through the anomaly map show the Bent Hill
anomaly is slightly wider in the north-south
direction (190 m width at half amplitude) then
in the east-west direction (150 m (Figure 6)).
The smaller ODP deposit �300 m south ofBent Hill is also seen in the figure as a small
GeochemistryGeophysicsGeosystems G3G3 gee et al.: deep tow magnetic survey 2001GC000170
local magnetic anomaly high. The strong neg-
ative anomaly in a broad region around these
small positive anomalies will be discussed in
section 5. Bent Hill, a region of raised sedi-
ments above an instrusion, is not directly asso-
ciated with a significant magnetic anomaly.
[19] The magnetic data constrain the width of
Bent Hill sulfide body. With additional con-
straints on magnetization it would also be
possible to constrain the volume, and more
weakly, the depth of the body. Previous model-
ing has used this approach, incorporating
limited drill core data from ODP Leg139 and
assuming the sulfide mound had the simple
geometry of a sphere or rod [Tivey, 1994a].
Additional drilling during ODP Leg 169 pro-
vides a more complete picture of the sulfide
mound and allows us to more directly model
the magnetic anomaly. The Bent Hill deposit is
a large mound with steep flanks. The base of
the massive sulfides is nearly horizontal at a
depth of �100 m (Figure 6) as determined fromcores recovered from a total of 12 drill holes
||
||
||
||
||
||
||
||
||
||
||
||
||
||
||
||
||
||
||
||
||
||
||
||
||
||
||
||
|
||
||
||
||
|
||
||
200
100
200 300
400
300
100
128o50'W 128o40'W
48o28'N
48o22'N
0 5 km-800 -600 -400 -200 0 200 >350 Anomaly (nT)
Figure 4. Near-bottom magnetic anomaly data and basement temperatures estimated from heat flow data.Color scale for anomaly data chosen to highlight variations within Middle Valley; note that anomalies abovebasement exposures in the east reach several thousand nT. Basement temperature contours from Davis andVillinger [1992].
GeochemistryGeophysicsGeosystems G3G3 gee et al.: deep tow magnetic survey 2001GC000170
along east-west and north-south transects
through the deposit from ODP Legs 139 and
169 [Fouqet et al., 1998; Zierenberg and
Miller, 2000].
[20] We begin by modeling the Bent Hill mas-
sive sulfide deposit as a stack of flat lying,
uniformly magnetized circular disks of varying
diameter that approximate the cross section
revealed by drilling (Figure 6). The top disk
lies at the seafloor, and the bottom of the stack
lies 100 m below seafloor. Model results are
presented assuming an average magnetometer
tow height of 70 m. With this geologically
constrained model, a uniform magnetization
of 5 A/m matches the observed amplitude of
525 nT. The results slightly overpredict the
width of the anomaly in the east-west direction
128o42'W 128o40'W128o41'W
48o26'N
48o25'N
-400 -300 -200 -100 0 100 200 300 400 Anomaly (nT)
-100
-100
-100
-100
2 00
1000
0
0
-200
-300
BHMS
Bent Hill
ODP
LS
Figure 5. High-resolution near-bottom magnetic anomaly data near Bent Hill area. Anomaly data havebeen reduced to the pole [Blakely, 1995]. Outline of Bent Hill, BHMS and smaller Lone Star (LS) andODP mound vents to the south are shown in red (locations from Fouqet et al. [1998]). Note that outlinesof these features have been shifted eastward by �50 m to account for an offset between the transpondernavigated tracks and the locations of features determined by the drilling ship. Near-bottom track linesshown as fine lines.
GeochemistryGeophysicsGeosystems G3G3 gee et al.: deep tow magnetic survey 2001GC000170
300
0
100
200
N
S
E
W
observedE - W Line
observedN - S LineAnomaly at 70 m
(reduced to pole)
depth varying model
uniform magnetizationmodel
Anomaly at 5 m(ambient field direction)
uniform magnetizationmodel
depth varyingmodel
-300 -200 -100 0 100 200 300
Distance (m)
MagnetometerTrack
Dep
th (
m)
2350
2450
2550
70 m
outline of circularcylinders
W E
massivesulfidemound
deep copper zone
sulfidefeederzone
Dep
th b
elow
sea
floor
(m
)
100
80
60
40
20
0
0 5 10 15
Magnetization A/m
Ano
mal
y (n
T)
6000
4000
2000
0
Ano
mal
y (n
T)
Figure 6. Geometry of BHMS deposit and forward models of magnetic anomaly at two different altitudes.(bottom) Geometry of the sulfide deposit along a W-E projection derived from drill core data [Fouqet et al.,1998]. The source for the magnetic models is a series of cylindrical disks (outline shown as dotted line) thatapproximate the shape of the BHMS deposit determined from drill core data. Inset shows magnetization fordepth varying model. (middle) Comparison of observed near bottom anomaly (altitude of �70 m) withforward models generated using the cylindrical disk source. Observed N-S and E-W profiles are slicesthrough the reduced to pole anomaly map (Figure 5). (top) Forward model predictions for anomaly at 5 mabove the seafloor using the same source.
GeochemistryGeophysicsGeosystems G3G3 gee et al.: deep tow magnetic survey 2001GC000170
and underpredict the width in the north-south
direction. The predicted magnetic anomaly is
necessarily circular since we began with a stack
of circular disks. If we had started with a
slightly non-radially symmetric body we could
fit both axes more closely. The fit in the north-
south direction would also be improved by
including the long wavelength field component
associated with the surrounding and deeper
structure.
[21] A magnetization of 5 A/m implies an
average volume susceptibility of 0.11 SI units
assuming an entirely induced magnetization in
the ambient field of 54,500 nT. This value
compares well with the average susceptibility
measured on sulfide samples from ODP drill
cores (arithmetic mean = 0.17 ± 0.15 SI; geo-
metric mean = 0.08 SI [Korner, 1994]). All
samples from the sulfide mound contain pre-
dominantly pyrite (90%), with 2–10% magnet-
ite and smaller amounts of hematite, pyrrhotite
and other sulfides [Korner, 1994]. Because the
susceptibility of both magnetite and pyrrhotite
exceeds that of pyrite or hematite by �3 ordersof magnitude [Hunt et al., 1995], the suscept-
ibility of these samples should predominantly
be related to the proportion of magnetite and
pyrrhotite. Moreover, hysteresis parameters
suggest that the magnetite within the sulfide
mound occurs primarily as coarse, multidomain
grains (saturation remanence/saturation mag-
netization = 0.01–0.15 [Korner, 1994]).
Although a significant remanent contribution
to the anomaly cannot be excluded, these
results suggest that the magnetization may be
adequately modeled as entirely induced. The
susceptibility of 0.11 S.I. suggests that the
sulfide mound contains an average of �3.5%magnetite + pyrrhotite by volume.
[22] The magnetic anomaly (Figure 6) predicted
by the uniform susceptibility model provides a
close match, both in amplitude and shape, to
the anomaly measured �70 m above the sea-
floor by the towed magnetometer. However,
this model greatly underpredicts the anomaly
(6000 nT) measured previously over the
BHMS mound at a height of �5 m by amagnetometer on Alvin in 1990 [Tivey,
1994a]. Two additional deep tow profiles col-
lected in 1988 reveal amplitudes of �600 nT atan elevation of 50 m and 1250 nT at 25 m
above bottom [Tivey, 1994a]. These values are
higher than would be predicted by our uni-
formly magnetized model for these elevations,
even though neither near-bottom profile
crossed the center of the sulfide mound. The
magnetic anomaly data at 70 m and those
collected by submersible much nearer the sea-
floor may be reconciled only if the upper,
narrower portion of the sulfide mound has
significantly higher magnetization than the
deeper layers. This is consistent with the ODP
results, which show a significant gradient in
susceptibility with depth, with susceptibilities
as high as 0.47 S.I. at the top of the deposit and
much lower values deeper in the deposit
[Korner, 1994].
[23] Both the deep tow and the Alvin data may
be fit using a linear gradient in magnetization
from a value of 12 A/m at the top of the deposit
tapering to 3 A/m at 30 m depth and with
constant magnetization (3 A/m) below. The
measurements from nearest the seafloor are
controlled by the high magnetization shallow
in the deposit, whereas the measurements from
higher above the seafloor depend more on the
average magnetization of the body (Figure 6).
Any model with values near 12 A/m shallow in
the deposit will fit the submersible data, and
models with average magnetizations near 5 A/
m over the bulk of the deposit will fit the data
from the 70 m tow. The magnetization of the
deepest layers is only very weakly constrained
by the data since these layers are farther from
the magnetometer. The higher magnetization
shallow in the deposit corresponds to a much
higher fraction of magnetite and pyrrhotite
GeochemistryGeophysicsGeosystems G3G3 gee et al.: deep tow magnetic survey 2001GC000170
(�10% by volume), but this value is still withinthe range of values determined from the core
material. Extensive low-temperature oxidation
of pyrrhotite to an assemblage of pyrite +
magnetite was noted in cores from the BHMS
[Duckworth et al., 1994]. Although no system-
atic variation in the degree of oxidation with
depth was evident, we speculate that enhanced
seawater circulation in the upper portions of the
deposit may have enriched magnetite relative to
pyrrhotite. Because magnetite has a higher
spontaneous magnetization and Curie temper-
ature than pyrrhotite, oxidation to magnetite
might be responsible for the enhanced magnet-
ization of the upper portion of the deposit.
5. Hydrothermal Effects on
Magnetization in Middle Valley
[24] Sea surface magnetic anomaly data reveal a
broad magnetic low over Middle Valley, with
higher values outside of the valley (Figure 7).
Currie and Davis [1994] have reviewed the
possible origins of this negative anomaly in
Middle Valley. The Bruhnes-Matuyama boun-
dary lies some 5 km beyond the valley boun-
dary faults, and so this low is not the result of
reversely magnetized crust. Rather, they asso-
ciate the anomaly with very low magnetization
of the crust below Middle Valley compared to
typical oceanic ridge basalt. A model fit to the
sea surface anomaly data places zero magnet-
ization over an �13 km wide region beneaththe valley with values near +14 A/m outside of
the valley [Currie and Davis, 1994].
[25] Two primary explanations have been
advanced for low magnetization values, and
correspondingly subdued anomaly amplitudes,
in sedimented ridge environments (see review
by Levi and Riddihough [1986]). Curie temper-
atures as low as 1008–2008C are observed forunoxidized titanomagnetites in spreading cen-
ter basalts [e.g., Johnson and Atwater, 1977;
Marshall, 1978]. Although more recent studies
have documented a broader range in titanomag-
netite compositions and Curie temperatures in
mid-ocean ridge basalts [Gee and Kent, 1997;
Zhou et al., 2000], Ti-rich titanomagnetites
with relatively low Curie temperatures are
likely to be the volumetrically dominant mag-
netic phase in most unaltered seafloor lavas.
Temperatures beneath Middle Valley are high
enough to significantly reduce the spontaneous
magnetization, and hence the remanence, of
any minerals with such low Curie temperatures.
However, as noted by Levi and Riddihough
[1986], the existence of very low amplitude
anomalies over crust as old as 3 Ma in sedi-
mented areas such as the Gorda Ridge/Esca-
naba Trough and the Gulf of California argues
against such thermal demagnetization as a
general explanation. Moreover, low tempera-
ture (<1008C) alteration during the initialstages of burial [Davis and Wang, 1994] should
have resulted in the transformation to cation-
deficient titanomaghemites with higher Curie
temperatures that would be less susceptible to
such thermal effects [Levi and Riddihough,
1986].
[26] A more likely cause for the absence of
lineated magnetic anomalies over sedimented
ridges is pervasive hydrothermal alteration
under the thick sediment blanket [Levi and
Riddihough, 1986]. Hydrothermal alteration
results in the replacement of titanomagnetites
by nonmagnetic phases (dominant volumetri-
cally) and a smaller amount of magnetite [Ade-
Hall et al., 1971; Pariso and Johnson, 1991;
Fujimoto and Kikawa, 1989]. As a result of this
hydrothermal alteration, remanence and satura-
tion magnetization (a direct measure of the
proportion of magnetite) can be reduced by
an order of magnitude or more [Woolridge et
al., 1990; Pariso and Johnson, 1991] with the
remaining magnetic phase being near end-
member magnetite [Shau et al., 1993]. Com-
parison of magnetizations of basalt samples
GeochemistryGeophysicsGeosystems G3G3 gee et al.: deep tow magnetic survey 2001GC000170
recovered from Site 855 (located on the eastern
boundary fault) and Sites 857 and 858 (near the
Dead Dog vent field and beneath the BHMS,
respectively) provide direct confirmation of the
effect of hydrothermal alteration on Middle
Valley basalts. A small number of samples
from the eastern site are characterized by high
magnetization (�10 A/m) and cation-deficienttitanomaghemites similar to basalts from non-
sedimented ridges [Fukuma et al., 1994]. In
129o30'W 129o00'W 128o30'W 128o00'W48o00'N
48o30'N
49o00'N
-800 -600 -400 -200 0 200 400 600 800
Anomaly (nT)
Sovanco FZ
Nootk
a Fa
ult
HeckelSeamount
Chain
Jara
mill
o
Old
uvai
Bru
nhes
Figure 7. Sea surface magnetic anomaly data in the vicinity of Middle Valley. Box indicates the areacontaining the near bottom survey lines. Locations of selected near ridge faults shown for reference. Note thatBrunhes/Matuyama boundary occurs just east of the eastern bounding faults.
GeochemistryGeophysicsGeosystems G3G3 gee et al.: deep tow magnetic survey 2001GC000170
contrast, samples from Sites 857 and 858 have
substantially lower magnetizations (�0.1 A/m)carried predominantly by pure magnetite
[Fukuma et al., 1994; Fouqet et al., 1998].
[27] The degree of alteration of basement
rocks might be expected to vary greatly
because there are large spatial variations in
heat flow within Middle valley (Figure 2).
A direct extrapolation from surface heat flow
measurements to basement depths predicts
temperatures varying by more than 2008Cat the sediment basement interface and
temperatures exceeding 4008C at the base-ment beneath both Dead Dog and Bent Hill
(Figure 4) [Davis and Villinger, 1992].
Elsewhere beneath Middle Valley, these
extrapolated basement temperatures range
mostly between 1008 and 2008C, with values<1008C more typical for the shallow base-ment to the east of the eastern bounding faults
(variations in basement depth are shown
in Figure 8a). This method yields accurate
basement temperatures only if vertical con-
duction is the dominant mechanism of heat
loss and if the physical properties of the
sediments and the thickness of the sediments
are well known. In areas with high heat flow
(>1 W/m2) where hydrothermal discharge
occurs, both nonvertical heat conduction and
advective transport will modify the temper-
ature field, and basement temperature esti-
mates are likely to be erroneous [Davis and
Villinger, 1992].
[28] Although the pattern of temperatures at the
basement-sediment interface shown in Figure 4
may be qualitatively correct, subsequent results
suggest that present-day maximum tempera-
tures in the hydrothermal reservoir beneath
both Dead Dog vent field and the ODP mound
south of Bent Hill are somewhat lower. The
temperature of fluids currently venting at these
two sites are �2758C [Ames et al., 1993;
Zierenberg and Miller, 2000]. This value is
close to that inferred for the reaction zone in
the sill/sediment complex (250–3008C [Stakesand Schiffman, 1999; Peter et al., 1994]).
Davis and Wang [1994] estimate a temperature
at top of the shallowest sill to be �2808C fromextrapolation of the gradient seen in the top of
Hole 857D south of Dead Dog. Basement
temperatures comparable to that of the active
vents (250–2808C) have also been estimatedfor Hole 857C using the observed heat flow
and grain conductivities of 2.6–3.2 W/m8K[Villinger et al., 1994]. The consistency of
these temperature estimates suggests that base-
ment temperatures in the immediate vicinity of
both Dead Dog and Bent Hill are presently
250–2808C.
[29] Lateral variations in basement temperatures
are also likely to be less than inferred from the
extrapolated heat flow data. The general
inverse correlation between heat flow and sedi-
ment thickness is thought to reflect a variable
thickness, low permeability sediment blanket
overlying a highly permeable and approxi-
mately isothermal basement [Davis and Lister,
1977b]. This model is supported by the small
variations in basement temperature (108–208C)inferred for the eastern flank of the Juan de
Fuca Ridge, where heat flow is well correlated
with basement topography [Davis et al., 1989].
The more pronounced basement relief in Mid-
dle Valley (Figure 8a) is likely to result in
somewhat more heterogeneous basement tem-
peratures (variation of 308–608C [Bessler et
al., 1994]). On the basis of two-dimensional
thermal modeling, these authors conclude that
basement permeabilities greater than 10�13 to10�12 m2 would be required to maintain base-ment temperatures within a 208C range. How-ever, measured permeabilities in Middle Valley
are typically 1–2 orders of magnitude lower
(10�14 to 10�13 m2) although permeabilities infault zones are locally as high as 10�10 m2
[Becker et al., 1994]. We therefore suggest that
the temperature of basement in Middle Valley
GeochemistryGeophysicsGeosystems G3G3 gee et al.: deep tow magnetic survey 2001GC000170
may vary considerably and that the isotherms in
Figure 4 provide a reasonable qualitative
description of this temperature distribution.
[30] Because higher temperatures are likely to
be associated with more pervasive hydrother-
mal alteration and will also facilitate thermal
demagnetization, one might expect that varia-
tions in basement temperatures would also be
reflected in the pattern of basement magnet-
ization. However, previous sea surface mag-
netic anomaly surveys have insufficient
resolution to discern whether significant varia-
tions in magnetization might be associated with
variable degrees of alteration within the valley.
The coarse spacing of survey lines (�10 km) inthis region, together with the inherent loss of
short wavelength information in sea surface
anomaly data, allows only broad scale features
of the crustal magnetization to be determined.
128o45'W 128o14'W 128o37'W
1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 11.0 >12.0
48o23'N
48o24'N
48o25'N
48o26'N
48o27'N
48o28'N
48o29'N
48o30'N
Mag. (A/m)Basement Depth (m)
128o45'W 128o14'W
3900 3700 3500 3300 3100 2900 2700 2500 2300
48o23'N
48o24'N
48o25'N
48o26'N
48o27'N
48o28'N
48o29'N
48o30'N
A B
Dead DogVent Field
Bent HillMassiveSulfide
858/1036
856/1035
857855
858/1036
856/1035
857855
2400
3100
3500
2600
2900
2800
2700
68
410
6
6
4
2
6
8
14
Figure 8. Depth to basement and magnetization solution from inversion of near-bottom anomaly data inMiddle Valley. (a) Contour map of sediment-basement interface used for magnetic inversion (after Davis andVillinger [1992]). (b) Magnetic anomaly data in Figure 4 were smoothed using a 100 m block median filterand gridded using continuous curvature splines under tension [Smith and Wessel, 1990]. The magnetic sourcewas a constant thickness (0.5 km) layer draped beneath the sediment-basement interface. The inversion wascarried out assuming an ambient field direction from the IGRF at the site (006/69) and a remanentmagnetization expected from an axial geocentric dipole (000/66). The passband for this analysis was 1–20km (wavelength) with tapers between 20 and 10 km and between 2 and 1 km. The original data are reflectedinto adjacent quadrants to reduce wrap-around effects at the edges. The magnetometer tows were sufficientlylevel that the tow depth could be assumed to be constant in the inversion without significant error. Sufficientmagnetic annihilator has been added to the solution to make the magnetizations positive everywhere.Locations of ODP drill sites (circles) and near bottom survey tracks are shown for reference.
GeochemistryGeophysicsGeosystems G3G3 gee et al.: deep tow magnetic survey 2001GC000170
On the basis of the sea surface anomaly data,
Currie and Davis [1994] have modeled the
Middle Valley anomaly low as a broad zone
of essentially nonmagnetic crust. Our near-
bottom magnetic anomaly data allow us to
evaluate whether the hydrothermal alteration
in this area has uniformly reduced the magnet-
ization.
[31] The deep tow magnetic field data (Figure
4) show large variations over Middle Valley
with some of the lowest field values associated
with the Dead Dog vent field and with the
region around the BHMS deposit. These
anomaly variations are the result of both
varying seafloor magnetization as well as
differences in the depth to the magnetized
basement. There is more than 1 km of top-
ography on the basement-sediment interface
associated with the large normal faults within
the valley (Figure 8a). To account for this
variable altitude above the magnetic source,
we have inverted for the magnetization within
a uniform thickness layer (500 m) draped from
the sediment-basement interface [Parker and
Huestis, 1974]. We have calculated the anni-
hilator [Parker and Huestis, 1974; Blakely,
1995] for this basement topography and added
a sufficient multiple of it to the inversion so
that magnetization is everywhere positive,
consistent with the location of Middle Valley
entirely within the Brunhes. Adding the mini-
mum annihilator for positivity also results in
magnetizations as high as �14 A/m east of theboundary faults. This value is well within the
range of magnetization determined from com-
parable age basalts samples from the Juan de
Fuca [e.g., Johnson and Holmes, 1989;
Fukuma et al., 1994] and is consistent with
magnetizations inferred from regional mag-
netic inversions [Tivey, 1994b].
[32] Because our data coverage is sparse, some
care should be exercised in interpreting the
resulting magnetization solution (Figure 8b).
The inversion depends on Fourier methods,
and there are obvious problems with Gibbs
phenomena (ringing and overshoot) particu-
larly in regions without data. The map pro-
duced by the inversion procedure can only be
usefully interpreted in regions near deep tow
lines. We apply spatial filtering to accommo-
date the very nonuniform distribution of data in
space, but the results depend on the passbands
of the spatial filter used in the analysis. Even
with this filtering, the large field gradient over
the eastern faults generates large oscillations
(ringing) in the eastern edge of the inversion.
The labels for the Dead Dog and Bent Hill vent
fields obscure two sections of the inversion,
which are completely unconstrained by data.
Despite these problems, we believe the inver-
sion is useful in the densely sampled center
section of the model between 488250N and488280N.
[33] Our inversion solution (Figure 8b) sug-
gests that magnetizations within Middle Valley
are not uniformly near zero, as had been
inferred from analysis of the sea surface
anomaly pattern [Currie and Davis, 1994].
Some ridge-parallel patterns in the magnet-
ization solution (e.g., near the eastern boun-
dary fault) are closely related to basement
structures, and variations in geochemistry,
paleofield, or source layer thickness may also
contribute to the complex magnetization pat-
tern. However, we interpret this magnetic
heterogeneity as reflecting primarily differen-
ces in the degree of hydrothermal alteration.
For example, a large region around the Bent
Hill sulfide mound is characterized by near
zero magnetization (<3 A/m), suggesting that
this area was extensively altered in the process
that formed the deposit. Similarly, low mag-
netizations inferred from a narrow magnetic
anomaly low near the TAG hydrothermal field
have also been interpreted in terms of perva-
sive hydrothermal alteration [Tivey et al.,
1993]. The magnetization beneath Dead Dog
GeochemistryGeophysicsGeosystems G3G3 gee et al.: deep tow magnetic survey 2001GC000170
vent field is also very low, although perhaps
slightly higher than that seen under Bent Hill,
with values between 3 and 4 A/m reflecting a
somewhat lower integrated degree of alteration
of the basement rocks. The remainder of the
region between the central graben fault and
the eastern boundary faults has low but finite
magnetization of between 4 and 6 A/m.
[34] Substantially higher magnetizations (�10A/m) are present west of the central graben
fault. Although the extent of track lines is
limited, all deep tow profiles in this region
show little variation across and to the west of
the Dead Dog vent field (Figure 4). These
higher magnetization values are primarily the
result of the much greater depth to basement
west of the central fault in Middle Valley
(sediment thickness rapidly increases to more
than 1 km west of this fault). The higher
magnetization in the down dropped block west
of the central fault suggests these rocks may
have experienced lower temperature or shorter
duration hydrothermal alteration and conse-
quently less reduction in magnetization.
[35] Although the difference in magnetization
inferred for Bent Hill and Dead Dog is small,
we suggest that the lower magnetization sur-
rounding Bent Hill is significant and that it
reflects more intense hydrothermal alteration
resulting from the higher paleotemperatures
associated with this sulfide deposit. Present-
day basement temperatures are similar at both
Bent Hill and Dead Dog, and so it is unlikely
that the difference in magnetization is due to
differences in the degree of thermal demagnet-
ization. Alteration mineral assemblages, oxy-
gen isotopic data, and fluid inclusion studies all
yield a remarkably consistent estimate of max-
imum alteration temperatures (�2758C) forsamples from Sites 857 and 858 [Stakes and
Schiffman, 1999]. Together with the absence of
significant sulfide accumulations, these data
indicate that basement temperatures surround-
ing the Dead Dog vent field were never much
above 2758C [Davis and Fisher, 1994]. In
contrast, the fluids responsible for the massive
sulfide accumulation at Bent Hill had substan-
tially higher temperatures (350–4008C [Good-fellow and Peter, 1994; Peter et al., 1994]). We
suggest that the higher fluid temperatures asso-
ciated with the Bent Hill sulfide deposit and/or
more protracted exposure to hydrothermal flu-
ids have resulted in more intense hydrothermal
alteration and hence a lower average magnet-
ization than evident near the Dead Dog vent
field.
[36] The magnetization of basement beneath the
Dead Dog vent field (Site 858) is nominally
lower than that near Site 857. This contrast is
surprising because these two areas are presently
hydrologically connected and have experienced
similar alteration conditions [Davis and Fisher,
1994]. Moreover, sills at Site 857 are more
pervasively altered than are corresponding
basement samples from Site 858 [Stakes and
Schiffman, 1999]. From these observations, one
might expect that magnetizations near Site 858
would be higher than near Site 857. Despite the
similarity of present alteration conditions at the
two sites, we suggest that the lower magnet-
ization at Site 858 may reflect a higher inte-
grated degree of alteration.
[37] There are a few additional features in the
inversion solution that are worth noting. The
very low magnetization around Bent Hill
extends to the ESE to a point directly adjacent
to the normal fault bounding the valley. No
significant heat flow anomaly is currently seen
above the site, but the magnetic anomaly low is
clearly seen on all three deep tow tracks cross-
ing overhead (Figure 4), although Fourier edge
effects may result in an overprediction of this
magnetization low (make it more negative than
it is) in the inversion solution. We suggest that
this region must have experienced an episode
of high-temperature alteration and that some
GeochemistryGeophysicsGeosystems G3G3 gee et al.: deep tow magnetic survey 2001GC000170
sulfide mineralization could be present within
the sediments beneath the surface. A small area
of low magnetization is also seen near
48827.30N, 128838.30E, which does seem tobe associated with a current day heat flow
anomaly of 6 W/m2.
6. Conclusions
[38] A deep tow magnetometer survey of Mid-
dle Valley, Juan de Fuca Ridge has been used to
develop a high-resolution map of magnetic
field anomaly in the region around the Bent
Hill Massive sulfide deposit. The shape and
extent of the BHMS mound is well resolved by
the deep tow magnetometer survey. Using a
magnetic source whose geometry is constrained
by the drilling data, the amplitude of the
anomaly can be fit assuming an average mag-
netization of 5 A/m. This magnetization is
consistent with average susceptibilities meas-
ured from ODP drill cores from the sulfide
mound and a combined fraction of magnetite
and pyrrhotite near 3.5%. Fitting the much
larger magnetic field anomaly detected previ-
ously in observations from 5 m above the
seafloor from Alvin [Tivey, 1994a] apparently
requires a strong gradient in the magnetization
with depth with surface values near 12 A/m.
However, this is also consistent with ODP core
results that show higher susceptibility at the top
of the deposit associated with a higher fraction
of magnetic minerals (10% magnetite + pyr-
rhotite). These results highlight the utility of
magnetic methods, particularly when anomaly
data are available at multiple levels, as an
exploration tool for mineral deposits in the
marine environment.
[39] The deep tow magnetometer survey at a
larger scale shows low but finite magnetization
in the basement rocks over most of the eastern
part of the Middle Valley graben. We interpret
the variable magnetization as reflecting hetero-
geneous hydrothermal alteration of the base-
ment. The lowest magnetization is found in the
region adjacent to the Bent Hill sulfide deposit
where the magnetization must be essentially
zero. The Dead Dog vent field is also associ-
ated with weak magnetizations, although
slightly higher values than observed near the
BHMS deposit. The slightly lower magnetiza-
tions associated with the BHMS deposit likely
result from the higher fluid temperatures
responsible for the sulfide mineralization and/
or a longer exposure to hydrothermal fluids.
Magnetization variations in this sedimented
ridge environment provide an indication of
the integrated hydrothermal alteration of the
basement and may prove valuable in locating
regions of past alteration that would otherwise
be difficult to detect.
Acknowledgments
[40] Partial support for this study was provided through
NSF grants OCE97-12027 (J.G.), OCE98-19779 (S.W.),
and OCE96-18325 (M.Z. and H.S.). We thank two
anonymous reviewers for helpful comments that im-
proved the manuscript.
References
Ade-Hall, J. M., H. C. Palmer, and T. P. Hubbard, The
magnetic and opaque petrological response of basalts to
regional hydrothermal alteration, Geophys. J.R. Astron.
Soc., 24, 137–174, 1971.
Ames, D. E., J. M. Franklin, and M. H. Hannington,
Mineralogy and geochemistry of active and inactive
vent chimmeys and massive sulfide, Middle Valley,
northern Juan de Fuca ridge: An evolving hydrothermal
system, Can. Mineral., 31, 997–1024, 1993.
Ballu, V. S., J. H. Hildebrand, and S. C. Webb, Seafloor
gravity evidence for hydrothermal alteration of the sedi-
ments in Middle Valley, Juan de Fuca Ridge,Mar. Geol.,
150, 99–111, 1998.
Becker, K., R. H. Morin, and E. E. Davis, Permeabilities
in the Middle Valley hydrothermal system measured
with packer and flowmeter experiments, Proc. Ocean
Drill. Program, Sci. Results, 139, 613–626, 1994.
Bessler, J. R., L. Smith, and E. E. Davis, Hydrological and
thermal conditions at a sediment/basalt interface: Inter-
pretation of field measurements at Middle Valley, Proc.
Ocean Drill. Program, Sci. Res, 139, 667–675, 1994.
GeochemistryGeophysicsGeosystems G3G3 gee et al.: deep tow magnetic survey 2001GC000170
Blakely, R. J., Potential Theory In Gravity and Magnetic
Applications, 441 pp., Cambridge Univ. Press, Cam-
bridge, Mass., 1995.
Currie, R. G., and E. E. Davis, Low crustal magnetization
of the Middle Valley sedimented rift inferred from sea-
surface magnetic anomalies, Proc. Ocean Drill. Pri-
gran, Sci. Results, 139, 19–27, 1994.
Davis, E. E., and R. G. Currie, Geophysical observations
of the northern Juan de Fuca ridge system: Lessons in
sea-floor spreading, Can. J. Earth Sci., 30, 278–300,
1993.
Davis, E. E., and A. T. Fisher, On the nature and conse-
quences of hydrothermal circulation in the Middle Val-
ley sedimented rift: Intferences from geophysical and
geochemical observations, Leg 139, Proc. Ocean Drill.
Program, Sci. Results, 139, 207–289, 1994.
Davis, E. E., and C. R. B. Lister, Tectonic structures on
the Juan de Fuca ridge, Geol. Soc. Am. Bull., 88, 346–
363, 1977a.
Davis, E. E., and C. R. B. Lister, Heat flow measured
over the Juan de Fuca ridge: Evidence for widespread
hydrothermal circulation in a highly heat transportive
crust, J. Geophys. Res., 82, 4845–4860, 1977b.
Davis, E. E., and H. Villinger, Tectonic and thermal struc-
ture of the Middle Valley sedimented rift, Juan de Fuca
ridge, Proc. Ocean Drill. Program, Inital Rep., 139, 9–
41, 1992.
Davis, E. E., and K. Wang, Present and past temperatures
of sediments at site 857, Middle Valley, Northern Juan
de Fuca ridge, Proc. Ocean Drill. Program, Sci. Results,
139, 565–570, 1994.
Davis, E. E., W. D. Goodfellow, B. D. Bornhold, J. Ads-
head, B. Blaise, H. Villinger, and G. M. Le Cheminant,
Massive sulfides in a sedimented rift valley, northern
Juan de Fuca Ridge, Earth Planet. Sci. Lett., 82, 49–
61, 1987.
Davis, E. E., D. S. Chapman, C. B. Forster, and H. Vil-
linger, Heat-flow variations correlated with buried base-
ment topography on the Juan de Fuca Ridge flank,
Nature, 342, 533–537, 1989.
Davis, E. E., et al., Proceedings of the Ocean Drilling
Program, Inital Reports, vol. 139, Ocean Drilling Pro-
gram, College Station, Tex., 1992.
Duckworth, R. C., A. E. Fallick, and D. Rickard, Miner-
alogy and sulfur isotopic composition of the Middle
Valley massive sulfide deposit, northern Juan de Fuca
ridge, Proc. Ocean Drill. Program, Sci. Results, 139,
373–385, 1994.
Fouqet, Y., R. A. Zierenber, D. J. Miller et al., Proceed-
ings of the Ocean Drilling Program, Inital Reports, vol.
169, Ocean Drilling Program, College Station, Tex.,
1998.
Fujimoto, K., and E. Kikawa, Alteration of titanomagne-
tites and its related magnetic properties in the Noya
Geothermal Area, Central Kyushu, Japan, J. Geomag.
Geoelectr., 41, 39–64, 1989.
Fukuma, K., U. Korner, and H. Oda, Rock magnetic prop-
erties of mafic rocks from Middle Valley sedimented
rift, Proc. Ocean Drill. Prog., Sci. Results, 139, 519–
533, 1994.
Gee, J., and D. V. Kent, Magnetization of axial lavas from
the southern East Pacific Rise (148–238S): Geochemicalcontrols on magnetic properties, J. Geophys. Res., 102,
24,873–24,886, 1997.
Gieskes, J. M., M. Kastner, G. Einsele, K. Kelts, and J.
Niemitz, Hydrothermal activity in the Guaymas basin,
Gulf of California: A synthesis, Initial Rep. Deep Sea
Drill. Proj., 64, 1159–1167, 1982.
Goodfellow, W. D., and J. M. Peter, Geochemistry of
hydrothermally altered sediment, Middle Valley, north-
ern Juan de Fuca Ridge, Proc. Ocean Drill Prog., Sci.
Results, 139, 207–289, 1994.
Hunt, C. P., B. M. Moskowitz, and S. K. Banerjee, Mag-
netic properties of rocks and minerals, in Rock Physics
and Phase Relations: A Handbook of Physical Con-
stants, AGU Reference Shelf Ser., vol. 3, edited b T. J.
Ahrens, pp. 189–204, AGU, Washington, D.C., 1995.
Johnson, H. P., and T. Atwater, Magnetic study of basalts
from the Mid-Atlantic ridge, lat 378N, Geol. Soc. Am.Bull., 88, 637–647, 1977.
Johnson, H. P., and M. L. Holmes, Evolution in plate
tectonics: The Juan de Fuca Ridge, in The Geology of
North America, The Eastern Pacific Ocean and Hawaii,
vol. N, edited by E. L. Winterer, D. M. Hussong, and
R.W. Decker, pp. 73–91, Geol. Soc. Am., Boulder,
Colo., 1989.
Korner, U., Rock magnetic properties of hydrothermally
formed iron sulfides from Middle Valley, Juan de Fuca
ridge, Proc. Ocean Drill. Program, Sci. Results, 139,
535–542, 1994.
Langseth, M. G., and K. Becker, Structure of the igneous
basement at sites 857 and 858 based on leg 139 down-
hole logging, Proc. Ocean Drill. Program, Sci. Results,
139, 207–289, 1994.
Larson, P. A., J. D. Mudie, and R. L. Larson, Magnetic
anomalies and fracture zone trends in the Gulf of Cali-
fornia, Geol. Soc. Am. Bull., 83, 3361–3368, 1972.
Levi, S., and R. Riddihough, Why are marine magnetic
anomalies suppressed over sedimented spreading cen-
ters?, Geology, 14, 651–654, 1986.
Marshall, M., The magnetic properties of some DSDP
basalts from the North Pacific and inferences for Pacific
plate tectonics, J. Geophys. Res., 83, 289–308, 1978.
Mottl, M. J., C. G. Wheat, and J. Boulegue, Timing of ore
deposition and sill intrusion at Site 856: Evidence from
stratigraphy, alteration, and sediment pore-water compo-
GeochemistryGeophysicsGeosystems G3G3 gee et al.: deep tow magnetic survey 2001GC000170
sition, Proc. Ocean Drill. Program, Sci. Results, 139,
679–693, 1994.
Pariso, J. E., and H. P. Johnson, Alteration processes at
Deep Sea Drilling Project/Ocean Drilling Program Hole
504B at the Costa Rica rift: implications for magnetiza-
tion of oceanic crust, J. Geophys. Res., 96, 11,703–
11,722, 1991.
Parker, R. L., and S. P. Huestis, The inversion of magnetic
anomalies in the presence of topography, J. Geophys.
Res., 79, 1587–1593, 1974.
Peter, J. M., W. D. Goodfellow, and M. I. Leybourne,
Fluid inclusion petrography and microthermometry of
the Middle Valley hydrothermal system, northern Juan
de Fuca Ridge, Proc. Ocean Drill. Program, Sci. Re-
sults, 139, 411–428, 1994.
Rohr, K. M. M., and U. Schmidt, Seismic structure of
Middle Valley near sites 855–858, Leg 139, Juan de
Fuca Ridge, Proc. Ocean Drill. Program, Sci. Results,
139, 3–17, 1994.
Shau, Y.-H., D. R. Peacor, and E. J. Essene, Formation of
magnetic single-domain magnetite in ocean ridge basalts
with implications for sea-floor magnetism, Science, 261,
343–345, 1993.
Smith, W. H. F., and P. Wessel, Gridding with continuous
curvature splines in tension, Geophysics, 55, 293–305,
1990.
Stakes, D. S., and P. Schiffman, Hydrothermal alteration
within the basement of the sedimented ridge environ-
ment of Middle Valley, northern Juan de Fuca Ridge,
Geol. Soc. Am. Bull., 111, 1294–1314, 1999.
Stein, J. S., and A. T. Fisher, Multiple scales of hydro-
thermal circulation in Middle Valley, northern Juan de
Fuca Ridge: Physical constraints and geologic models,
J. Geophys. Res., 106, 8563–8580, 2001.
Stein, J. S., A. T. Fisher, M. Langseth, W. Jin, G. Iturrino,
and E. Davis, Fine-scale heat flow, shallow heat sources,
and decoupled circulation systems at two sea-floor hy-
drothermal sites, Middle Valley, northern Juan de Fuca
Ridge, Geology, 26, 1115–1118, 1998.
Tivey, M. A., High-resolution magnetic surveys over the
Middle Valley mounds, northern Juan de Fuca Ridge,
Proc. Ocean Drill. Program, Sci. Results, 139, 29–35,
1994a.
Tivey, M. A., Fine-scale magnetic anomaly field over the
southern Juan de Fuca Ridge: Axial magnetization low
and implications for crustal structure, J. Geophys. Res.,
99, 4833–4855, 1994b.
Tivey, M. A., P. A. Rona, and H. Schouten, Reduced
crustal magnetization beneath the active sulfide mound,
TAG hydrothermal field, Mid-Atlantic ridge at 268N,Earth Planet. Sci. Lett., 115, 101–115, 1993.
Villinger, H. W., M. G. Langseth, H. M. Groschel-Becker,
and A. T. Fisher, Estimating in-situ thermal conductivity
from log data, Proc. Ocean Drill. Program, Sci. Results,
139, 545–552, 1994.
Woolridge, A. L., S. E. Haggerty, P. A. Rona, and C. G.
A. Harrison, Magnetic properties and opaque mineral-
ogy of rocks from selected seafloor hydrothermal sites at
oceanic ridges, J. Geophys. Res., 95, 12,351–12,374,
1990.
Zhou, W., R. Van der Voo, D. R. Peacor, and Y. Zhang,
Variable Ti-content and grain size of titanomagnetite as
a function of cooling rate in very young MORB, Earth
Planet. Sci. Lett., 179, 9–20, 2000.
Zierenberg, R. A., and D. J. Miller, Overview of the
Ocean Drilling Project Program Leg 169: Sedimented
ridges, Proc. Ocean Drill. Program, Sci. Results, 169,
1–39, 2000.
Zierenberg, R. A., J. L. Morton, R. A. Koski, and S. L.
Ross, Geologic setting of massive sulfide mineralization
in the Escanaba Trough, in Geological, Hydrothermal,
and Biological Studies at Escanaba Trough, Gorda
Ridge, Offshore Northern California, edited by J. L.
Morton, R. A. Zierenberg, and C. A. Reiss, U. S. Geol.
Surv. Bull., 2022, 171–197, 1993.
Zumberge, M. A., J. R. Ridgway, and J. A. Hildebrand, A
towed marine gravity meter for near-bottom surveys,
Geophysics, 62, 1386–1393, 1997.
GeochemistryGeophysicsGeosystems G3G3 gee et al.: deep tow magnetic survey 2001GC000170