1
Locating hydrothermal vents by detecting and modeling buoyant, advected plumes
Scott R. Veirs, Russell E. McDuff, Marvin D. Lilley, and John R. Delaney
School of Oceanography, Box 357940, University of Washington, Seattle, WA 98195-7940
Please correspond to Scott Veirs:
E-mail: scottv@ocean.washington.edu
Prior to June, 1999:
Mailing address: 3418 Dover Road, Redwood City, CA 94061
Phone: (650) 261-9162
Fax: (650) 365-0678
2
Abstract
An improved method of detecting buoyant hydrothermal plumes and locating the vents
that emit them is introduced. Fluid stability is computed from conductivity, temperature, and
depth measurements acquired during navigated, towed, vertically oscillating casts over the
Endeavour Segment of the Juan de Fuca Ridge. The altitude of each instability and the
corresponding current record are used to constrain a model of plume rise in an ambient cross
flow, enabling the estimation of a hydrothermal source location. The method reliably locates
sources where they are known to exist: in all of the fields of focussed vents that have been
mapped by submersible, and in the only known, isolated, diffuse flow sites. The regional
distribution of instabilities reveals evidence of many undiscovered hydrothermal sources. Within
the axial valley, convective up flow zones appear to be more closely spaced, more elongate, and
present further south than previously mapped. Beyond the axial valley, the existence of
instabilities suggests that hydrothermal up flow penetrates the relatively unfissured outer slopes
of the central ridge.
3
Introduction
Our understanding of hydrothermal circulation in the young oceanic crust and its impact
on the oceans is hindered by a technological inability to map convective up flow efficiently,
comprehensively, and precisely. The distribution of hydrothermal vents helps to test models of
fluid circulation in the crust [Brikowski and Norton, 1989; Fehn et al., 1983; Rosenberg et al.,
1993; Travis et al., 1991; Wilcock and McNabb, 1996], provides spatial context for interpreting
geochemical differences between individual vents and vent fields [e.g. Butterfield, 1994], and
guides the design of experiments that address an enduring problem in ocean ridge studies:
quantification of fluxes of heat, water, and chemicals through both the volcanic axis and flanks
[Elderfield and Schultz, 1996; Stein and Stein, 1994].
The fundamental motivation for using stability to estimate proximity to a vent is that the
equilibration time of turbulent hydrothermal plumes in the Northeast Pacific is small,
considerably less than the characteristic periods of currents, and approximately independent of
source strength and geometry [Middleton, 1979; Middleton and Thomson, 1986]. Over areas
where potential temperature or light transmissivity anomalies cannot discriminate between
advected and local hydrothermal signals, instabilities can locate convective up flow precisely and
unambiguously.
The goal of this paper is to show how a stability survey can improve maps of sea floor
hydrothermal activity, and thereby help to constrain what factors govern the convection of water
in an intermediate spreading rate ridge crest. We begin by explaining how stability can be
estimated from hydrographic data collected over hydrothermal vent systems.
We then
demonstrate how plume theory can be used to estimate a vent location, given the altitude each an
instability and a record of deep sea currents velocity and direction. After verifying the method
4
over known vent fields, we present the regional distribution of instabilities and inferred vent
locations. Finally, we discuss how the distribution of vents changes present conceptualizations
of the Endeavour Segment hydrothermal system.
Observations
The Mixing Zephyrs field program involved the acquisition of a navigated hydrographic
survey of the Endeavour Segment of the Juan de Fuca Ridge during two cruises of the R/V
Atlantis II (May 25-June 19, and September 15-28, 1995). The survey coverage (Figure 1) was
designed to extend the data set of Thomson et al. [1992] and provide a regional hydrothermal
context for interpreting chemical data collected by submersible from two adjacent, well-mapped
vent areas: Main Endeavour Field (MEF) [Delaney et al., 1992] and High-Rise Field (HRF)
[Robigou et al., 1993]. An underlying assumption of the hydrographic program (based in
focussed plume theory and field experience) was that the distribution of active, high-temperature
sites could be inferred from the locations of the plumes rising from and equilibrating above them.
On 48 occasions during Mixing Zephyrs, a CTDT (package of instruments measuring
conductivity, temperature, depth, and light transmissivity) was raised and lowered while the ship
was either stationary, or under way at ~0.5 m/s (~1 knot). The latter “tow-yo” technique [Baker
et al., 1985] is a vertically oscillating tow (VOT) that results in measurements (along a saw-tooth
path) which, when contoured, display water conditions throughout a vertical plane between ~10300 m above the variable bathymetry of the ridge (Figure 2). Attention centered on the effluent
layer [Lupton, 1995], a depth range where hydrothermal plumes from focussed, intense sources
on the Endeavour Segment are expected to equilibrate in the stratified Northeast Pacific. The
CTDT data were collected using a Sea-Bird Electronics 25 Sealogger CTD and a SeaTech (25
cm path length) light transmissometer, typically during night time tows lasting from 2 to 5 hours.
5
An acoustic transponder attached to the CTDT enabled measurements to be located with
a precision of about 10 m and an accuracy that depended on position within the sea floor
transponder network. Near the MEF and HRF, at the center of the network, accuracy was ~10
m; beyond this region, accuracy was ~100 m. During each cycle of a VOT, the ascent and
descent paths are typically 30-55o from vertical, and consecutive near-bottom approaches are
about 500-800 m apart horizontally.
The utility of stability
The presence of oscillatory, rotating currents in the deep ocean complicates the spatial
relationship between a hydrothermal plume and its sea floor source. Locating a vent from a
tracer in its advected plume depends on the nature of the tracer, its rate of decay in the water
column, and the currents [Baker et al., 1995; Lilley et al., 1995a; Lupton, 1995]. A conservative
tracer, by definition without sources or sinks within the ocean interior, cannot be used to
discriminate unambiguously between signals from a potent source in the distance and a weak
source directly below. Non-conservative tracers, on the other hand, have widely varying
residence times in the water column [Kadko et al., 1990; Lupton, 1995]. If the initial
concentration of a non-conservative tracer decreases (through chemical reactions, biological
interactions, or mixing processes) to background levels before being advected over additional
sources, then that tracer enables more precise specification of its source location and does not
mask signals from other sources.
Additionally, if a non-conservative tracer signal decays faster
than the characteristic period of an oscillatory current, then it cannot be swept over its own
source and lead to uncertainty about the source intensity.
6
Limitations of standard plume tracers
The most spectacular temperature anomalies in the ocean occur where high temperature
vent fluids discharge, but they are extremely short-lived. The initial difference between vent and
ambient fluid temperatures is diluted by a factor of ~103-104 [Lupton et al., 1985; McDuff, 1995]
so rapidly through turbulent mixing with cold, ambient sea water that the spatial extent of strong
thermal anomalies is too limited to be effectively mapped by available hydrographic techniques.
Temperature anomalies decrease from ~350 K to 5 K before rising more than ~10 m [Speer and
Rona, 1989] and in the order of 30 s after release from a focussed vent [McDuff, 1995].
Upon equilibration, which takes ~0.5 hr in the Northeast Pacific [Middleton and
Thomson, 1986], potential temperature anomalies in the Pacific are typically only 0.01-0.1 K and
subsequently decrease slowly through diffusion. Salinity (S) and potential temperature () are
conservative tracers of hydrothermal effluent in the Northeast Pacific [Baker and Massoth, 1987;
Cannon et al., 1993; Lupton et al., 1985; Speer and Rona, 1989]. Both remain detectable as
anomalies even when advected far from their source. Potential temperature anomalies (Figure 2,
referenced to salinity [McDougall, 1990] rather than to potential density in this study) may
remain detectable for weeks to years within sub-mesoscale coherent vortices [D'Asaro et al.,
1994].
During the Mixing Zephyr cruises, the most intense  recorded was 0.65 K, detected
within a buoyant plume ~60 m above vent orifices emitting 350-380 oC effluent. In contrast,
~0.08 K was commonly encountered over the axis and was intermittently observed in the
effluent layer up to 2.2 km to the west and 2.7 km to the east of the axis. Hydrothermal vents
with different source fluid temperatures and salinities can produce thermal and salinity anomalies
that have similar magnitudes upon equilibration. Because the initial signal is reduced so quickly,
7
 and even potential temperature gradient cannot be used to reliably differentiate between heat
from a weak, local source and heat advected from a strong, distant source. Without knowledge
of initial effluent characteristics and consideration of mixing models, neither parameter can
constrain the proximity of anomalous fluid to its hydrothermal source, even if the advective field
is well-characterized. A potential temperature anomaly or inversion encountered near (<100 m
above) the sea floor, for example, could be interpreted as evidence of nearby venting.
Alternatively, the warm fluid could have equilibrated high above an axial focussed source and
then been advected on an isopycnal surface that intersects or closely (<100 m) overlies a
hydrothermally inactive portion of the ridge crest (Figure 2).
Measurements of percentage light transmissivity (c) also fail to reliably indicate source
proximity, for two reasons. First, the residence time of the occlusive particles appears to be
relatively long despite their non-conservative nature [Kadko et al., 1990]. The light scattering
anomaly in an event plume decreased by < 15% over 60 days [Lupton et al., 1998]. Secondly,
the initial particle concentration varies widely between hydrothermal sources. Strong c
anomalies often mark rising plumes above the MEF (Figure 3A), while plumes from high
temperature vents in the Mothra area can exhibit comparable temperature (and density)
anomalies with almost no coincident change in c (Figure 3B) [Lilley et al., 1995b]. Diffuse
hydrothermal effluent is generally transparent [Rona et al., 1997] and is therefore unlikely to
produce a detectable c signal.
Nevertheless, two new vent fields, Salty Dawg (Kelley et al., in preparation) and Mothra
(Kelley et al., in preparation), were discovered beneath strong unequilibrated anomalies of
potential temperature (and in the case of Salty Dawg, light transmissivity) that were located
during Mixing Zephyrs CTDT VOTs. These discoveries demonstrate that  and c anomalies can
8
locate (high temperature, point source) hydrothermal activity, but only if a rising plume is
traversed so close above its source that the source signals are not diluted to the equilibrated
magnitudes typical of the geothermal boundary layer [Thomson et al., 1995]. Surveys of thermal
and anomalies alone can cause important hydrothermal signals to be overlooked or
misinterpreted.
Diagnostic advantages of stability
The main advantage of using stability to estimate proximity to a vent is that the majority
of hydrothermal sources at the Endeavour Segment, despite their diverse physical and chemical
characteristics, produce buoyant plumes that affect the density structure of the overlying ocean.
Effluent rising from most focussed (high temperature effluent flowing at high velocity from a
small diameter orifice) and diffuse (low temperature effluent percolating through an extensive
area) sources can be modeled as a turbulent plume. The transition between the diffuse and
focussed regimes is best defined not by temperature ranges, but by the dimensionless number
noted by McDuff [1995]:
B1/ 2
,
AN 3/ 2
(2)
a function of the cross-sectional area (A) of the hydrothermal source, as well as the buoyancy
flux (B) and buoyancy frequency (N)
B

N (
gQ
(3)
g d 1/ 2
)
 dz
(4)

which characterize the strength of the source and ambient stratification, respectively. (Q is
volume flux – defined as the product of A and effluent vertical velocity, w;  is the density
9
contrast between vent and ambient fluids; g is gravitational acceleration;  is a reference density;
and z is pressure depth) For most known Endeavour hydrothermal sources, the dimensionless
criterion is much greater than one [McDuff, 1995], indicating that a plume model best describes
the hydrodynamic behavior of their effluent. For large diffuse flow sites and the upper surfaces
of the largest sulfide structures at Endeavour, the criterion is ~1, suggesting that a heated surface,
mixed layer model [Little et al., 1988; Trivett, 1994] may be applicable. Field observations of
diffuse flow [Rona and Trivett, 1992; Trivett and Williams, 1994] reinforce this possibility.
Nevertheless, we assume that hydrothermal sources can be located by the plumes they emit
because the plumes equilibrate in a time that is considerably shorter than the characteristic
periods of currents, and approximately independent of source strength and geometry [Middleton,
1979; Middleton and Thomson, 1986].
Hydrothermal plumes weaken the ambient vertical density gradient above the Juan de
Fuca Ridge within the effluent layer [Baker and Massoth, 1987] and create detectable density
inversions as they rise [Lupton et al., 1985]. One measure of the vertical density gradient is
stability (E) [Pond and Pickard, 1983],
E
1  2
,
 z
(1)
a diagnostic parameter that can be estimated from a time series of z and potential density
referenced to 200 bars (~2000 m), 2. We estimate E from data acquired during both vertical
casts and navigated VOTs. First, 1 Hz 2 data (bin averages of the raw 8 Hz samples) are
smoothed by a 32-point running mean box car. Though the 2 data are spatially distributed
along a path that is rarely vertical, typical isopycnal surfaces over the ridge are nearly level
(Figure 2) on the scale of adjacent isobaric measurements along a VOT track (500-800 m
10
horizontal separation). Consequently, we compute an along-path gradient by proceeding through
the smoothed 2 time series until |z|10m. Whenever that depth interval is obtained, the end
point 2 values are differenced and divided by z. Finally, this approximation of the vertical
potential density gradient is divided by a reference (2000 m) density. The resultant estimate of E
is associated with the central value of the depth interval, and the corresponding time and
position.
This incremental gradient calculation circumvents (tractable, but inconvenient) situations
where continuous gradient computation is problematic: when the CTDT is towed at constant
depth (potential division by z=0 ), and at the points when the CTDT reverses vertical direction.
The width of the averaging window and the depth interval were chosen through consideration of
the variability of the raw and smoothed 2 data, the expected spatial extent of large plume
density variations (~10 m), and the results of testing the algorithm over vents of known character
and location. When isopycnal surfaces are near horizontal and the CTDT path is near vertical,
this stability parameter approximates the vertical component of the potential density gradient
along the CTDT path.
Locating a hydrothermal source through plume equilibration theory
Middleton [1979] and Middleton and Thomson [1986] show that the characteristic time
() taken by a plume to equilibrate above its source depends primarily on the buoyancy
frequency (N):
  N 1
(5)
At our site, average values of N from vertical casts over the depth range (1800-2300 m) through
which plumes rise were ~0.002 s-1 at background and other uncontaminated locations, and
~0.0015 s-1 within the geothermal boundary layer over the ridge. The approximate constancy of
11
the equilibration time of effluent from diverse hydrothermal sources suggests a method of
estimating the distance downstream beyond which their plumes should contain no large scale (10
m) instabilities. By assuming that rising plumes in a strong current experience the background
stratification, N=0.002 s-1, the characteristic time that elapses before plumes equilibrate over the
Endeavour Segment is about 1500 s, or ~0.5 hr. Thus, the maximum radial range (X) from the
source to the last major instabilities should be approximately equal to the maximum observed
current velocity (U) multiplied by the characteristic equilibration time ():
X  U
(6)
We measured U with two current meters moored throughout the study period at 47o 57.0’
N, 129o 05.6’ W, ~100 m E of the NE corner of the MEF, in about ~2200 m of water (Figure
5B). Every 10 min the meters recorded speed and direction 25 m and 300 m above the sea floor.
The currents oscillate at both depths with peak powers at about 6, 12, 16.5, 24, and 26 hours –
tidal and inertial periods that are common above the Juan de Fuca Ridge [Cannon and Thomson,
1996]. Integrating the components of current speed throughout the study period shows that mean
velocities of ~5 cm/s to the SSW at ~1900 m depth (300 m altitude) and ~1 cm/s to the N at
~2175 m depth (25 m altitude). Analysis of the velocity magnitude without considering
direction reveals a general increase in speed between the lower (39 cm/s max; 3.5 cm/s mean, 1.3
cm/s median) and upper (67 cm/s max; 14 cm/s mean, 10.5 cm/s median) meter.
A fundamental assumption of our method is that the upper current meter mooring
provides a record of U that is applicable regionally in modeling plume trajectories. We justify
this with the observation of high coherency in horizontal and vertical flow characteristics above
the southern Juan de Fuca Ridge [Baker and Cannon, 1993], and the correlation of known vent
locations with those predicted using our method. However, the legitimacy of this assumption is
12
challenged by evidence (from current meters distributed in the axial valley and on the
southwestern slopes of the central ridge) that there is some horizontal heterogeneity and vertical
shear in the velocity field over the Endeavour Segment [Kadko et al., 1990].
The product of the maximum measured U (0.67 m/s) and  (1500 s) is ~1 km. One way
to improve this estimate of X is to declare U to be the maximum velocity recorded (at the upper
current meter) not during the entire survey period, but during the interval (t-, t), where t is the
time at which an instability is observed. This leads to X<500 m for most instabilities, an estimate
which can be further improved through incorporation of the measured altitude (height above the
local bottom) of each instability into a model of plume trajectory in an ambient cross flow.
Middleton [1986] extends the classic plume theory of Morton et al. [1956] to conditions
of uniform cross-flow (U), and provides the equations needed to derive a plume trajectory. The
height (Z) of the center of a Gaussian plume as a function of the range (X) downstream from the
source is
Z   1 {bo  [bo  3 (UN ) 1 ( M 2  B 2 N 2 )1/ 2 (cos   cos( NXU 1   ))]1/ 3}
3
(7)
in which  is an entrainment coefficient, bo is the radius of the source, M is source momentum
flux (defined as Qw), and  is a measure of the ratio of momentum and buoyancy flux at the
source, made dimensionless with N [Middleton, 1979]:
  tan 1 ( NMB 1 ) .
(8)
Hydrothermal plumes observed to date at the Endeavour Segment have non-zero momentum and
buoyancy fluxes at their source, and therefore have 0 < < /2 [Middleton, 1986]. While the
trajectory of an advecting, bent over plume is affected by source characteristics, the down stream
distance at which the plume reaches equilibrium depends most strongly on U, , and N, not bo,
B, or M (Figure 4).
13
After solving equation 7 algebraically for the downstream distance, X~f(Z,U,M,B,,bo,N),
U
X 
N
3


UN[( Z  bo ) 3  bo ] 
1


   cos  cos  
2
2
 2 1 / 2 
3

(
M

B
N
)



(9)
we use a current velocity (U) and the altitude (Z) of an instability to estimate the range to its
source. As before, we estimate U as the maximum velocity observed (at the upper current meter)
during the interval (t-, t). We assume that the source characteristics are similar to those of a
black smoker, with B=3.7x103 m4/s, M=1.3x103 m4/s2, and bo=0.02 m [McDuff, 1995].
Furthermore, we set =0.33 for these source properties following Middleton and Thomson
[1986] and assume that the plume rises through a stratified environment in which N=0.002 s-1.
We estimate the altitude of an instability by two independent methods. An acoustic
pinger attached to the CTDT cage enabled monitoring of the instrument package altitude during
each transect with a Precision Depth Recorder (PDR) with an accuracy of ~5 m. During data
processing, altitude was calculated as the difference between the depth of the underlying sea
floor (interpolated from a gridded Hydrosweep bathymetric data set) and pressure depth. With
attention to areas where the true depth is certain from myriad submersible visits, intercomparison
of these two methods suggests that the Hydrosweep depths are typically ~25 m greater than the
sum of the PDR altitude and the CTDT pressure depth (z). In the following calculations we
estimate instability altitude as the difference between z and the Hydrosweep sea floor depth
because its likely underestimated value results in a more conservative (over-estimation of the)
source range.
By utilizing the altitude information we estimate X for each detected instability more
precisely than is possible with other mapping methods and tracers. Additionally, we can
estimate the bearing of the source by computing from the current vector time series the sequence
14
of displacements that would have occurred during the time interval (t-,t). Reversing the
displacements, we obtain a trace of the position of the instability as it rose and was advected to
the location where it was detected. The bearing to a source location is determined by the
intersection of this trace and a circle of radius X.
A number of factors can result in the source of an instability having the same bearing, but
a range less than X. N could be less than the assumed background value if nearby or upstream
venting has eroded the local stratification. Z could be overestimated, particularly if the source is
located on a topographic high or a sulfide structure of substantial relief. If the plume trajectory
(modeled on the basis of the assumed ambient stratification and vent characteristics) does not
attain the altitude at which an instability is observed, our computational method reverts to
calculating X according to equation 6, setting U equal to the maximum current velocity observed
in the interval (t-, t).
Results: The distribution of instabilities and their sources
The stability of the water over the Endeavour Segment has been analyzed for every
Mixing Zephyrs CTDT VOT and vertical cast. Coincident depth profiles of  and X (Figure 3),
as well as S and  were used to rate the confidence that each instability was related to a
hydrothermal source. An instability was accepted as an indicator of local hydrothermal activity
only if it met three criteria: (1) the density profile was clearly inverted (above 2 noise); (2) the
coincident parameter profiles were consistent with a hydrothermal source; (3) the instability
altitude was not unusual (300m above the sea floor). The classification was strengthened if
adjacent profiles (same VOT) indicated the density disturbance was local, and if intersecting
and/or nearby transects yielded corroboratory instabilities. Figure 5 displays the locations of
only those instabilities classified as indicative of hydrothermal activity.
15
Verification over known vents
The general resilience of our method is indicated by the observation that despite sparse
survey coverage of the High-Rise and Salty Dawg areas (Figure 1A), strong instabilities were
detected over both of them, as well as above the thoroughly surveyed Mothra and Main
Endeavour fields (Figure 5). All instabilities overlying the well mapped Salty Dawg, High-Rise,
and Main Endeavour fields have bearings which intersect known vents (Figure 5). The
prevalence of such intersections supports the assumption that the upper current meter
measurements can be applied regionally in modeling plume advection. The loftiest inversions
(200-300 m altitude), likely indicating particularly intense hydrothermal sources, are all located
above the axis and clustered near the Main Endeavour, Mothra, and High-Rise fields. Lower
(100-200 m) inversions supplement those upper clusters, and are also present over the Salty
Dawg field.
The best place to ascertain how reliably our method locates hydrothermal sources is
directly above the MEF, a well-mapped field of focussed vents directly adjacent to the current
meter mooring. The MEF consists of 13 known sulfide structures which host ~100 black
smokers and are both surrounded and interspersed with patches of diffuse flow [Delaney et al.,
1992]. Figure 5B shows the MEF vicinity, the distribution of all known vents, and the proximity
of the current meter mooring. In this area (and within ~1 km radius)  maps a complex
assemblage of thermal features – some buoyant effluent and some equilibrated, advected fluid.
In contrast, stability enables the inference of four specific vent locations within the field.
Although the two southwestern instabilities (within the grey box in Figure 5B) were
encountered 114 and 169 m above the southern cluster of MEF vents, the current meter and
plume model indicate the sources are northern vents, likely the black smokers atop the structures
16
known as Lobo and Grotto. The two more northern instabilities (within the grey box) are
associated with extraordinary density inversions and potential temperature anomalies,
compelling evidence that the buoyant plumes from the underlying black smokers were traversed
by the CTDT. Given the westward currents at the time, the plume model predicts sources to be
along eastward bearings and with maximum ranges that extend beyond the eastern edge of the
field. However, the intensity of the instabilities and thermal anomalies suggests that the fluid
was far from equilibrium, younger than ~0.5 hr old, and that the source locations therefore lie on
the same bearing, closer to the instabilities than the modeled maximum range. At an altitude of
30 m, exactly above the 15-18 m tall, ~33 m wide active sulfide structure named Dudley
[Delaney et al., 1992], the northernmost instability is related to a plume of buoyant effluent
about 25 m wide.
An ideal test of whether instabilities due to diffuse flow can be detected with the CTDT is
to survey the water column over a known diffuse vent site that is isolated from focussed vents.
During the few submersible excursions have traversed between known vent fields, at least four
diffuse flow sites segregated from focussed vent fields have been mapped: Clam Bed, Cirque,
Dune, and Quebec (Figure 1).
The altitudes of the instabilities above Clam Bed and Quebec suggest that the diffuse
vents known to be active in the vicinity are not isolated; rather, they are accompanied by nearby,
undiscovered focussed vents. Clam Bed is associated with at least one white smoker [Robigou et
al., 1993], the probable source of a moderate instability (Figure 5A) that was encountered too
high in the water column (~250 m) to be from a diffuse vent. A strong  signal sensed south of
the MEF at an altitude of 110 m was expected to lead to a focussed source, but the subsequent
dive found only an area of diffuse venting which was dubbed Quebec.
17
Cirque and Dune lie ~500 m west of the Salty Dawg field (Figure 5C), beyond the reach
of its unstable effluent. A profile from a nearby tow (Figure 3C) reveals an encounter with
buoyant, moderately warm, clear effluent 40 m above the local topography – a signature
consistent with that expected of diffuse flow [Trivett, 1994]. The associated current record
yields a bearing toward the known sites (Figure 5C). Other instabilities encountered on a
distinct, parallel transect create a cluster which reinforces the possibility that our method may be
capable of detecting plumes from diffuse flow sites. Alternatively, focussed flow may be
emanating from sulfide structures perched on the valley rim above Cirque and Dune. While no
sea floor observations are available to confirm or deny this latter possibility, sulfide talus was
noted along the base of the western wall of the axial valley at that latitude during ALVIN Dive
2931.
Undiscovered vents
If the assumptions inherent in our method are correct, many of the strong instabilities
detected over the Endeavour Segment must have originated beyond the presently accepted
boundaries of the known vent fields (Figure 5). We present the most prominent instabilities from
south to north, in two groups: first those that are in the axial valley, and then those that are
located outside the axial valley.
A concentration of instabilities detected about 5.3 km SSW of the MEF during a single
transect (Figure 5A) evidence an axial, high temperature, moderate c source near 47o 54.3’ N,
129o 07.8’ W. Its latitude is similar to that of a thermal and c anomaly attributed to a southern
source by Thomson et al. [1992], but its longitudinal position is about 3 km to the west, a
discrepancy that cannot be explained by the navigational uncertainty (~1 km) associated with use
of Loran-C in the 1988 field work.
18
The present boundary of the Mothra field (Figure 5A) was first delineated [Kadko et al.,
1990] as a region of unusually high and occasionally exceptional  (Figure 3B). Remotely
operated vehicle investigations (Kelley et al., in preparation) verified the presence of high
temperature sources in the northern half of the projected field area. Many of the instabilities lie
beyond the influence of the mapped vents and cluster along the axis to the south-southwest. The
distribution suggests that the Mothra field extends further S than has been mapped by
submersible reconnaissance.
About 1.2 km along the axis south of the MEF (~47o 56.2’ N, 129o 06.2’ W), an
instability was detected at an altitude of ~100 m. It is associated with a distinct  that was
presumed (during the field program) to have originated in the MEF. However, the c signal is
less than is characteristic of MEF effluent and the elevation is below the typical rise height of
plumes from the MEF. Because the currents were quiescent at that time it is theoretically
implausible that the buoyant fluid originated from either the Mothra or Main Endeavour fields.
Therefore, we infer that a hydrothermal source exists midway between the two known vent
fields, a result that is inconsistent with the concept [Delaney and Crest 91-Flange Research
Teams, 1991] that high temperature up flow zones are spaced ~2 km apart along the axis of the
Endeavour Segment.
Five axial instabilities detected near the MEF (Figure 5A and 5B) offer compelling
evidence that hydrothermal sources exist well beyond its accepted borders. One is associated
with a plume that was noted during the field program based on its thermal anomaly, but
attributed at the time to a Main Endeavour source. It is unlikely the source is in the field because
the rising plume was detected more than 500 m north-northeast of the nearest known vent in the
MEF, the instability was detected only 40 m above the bottom, and currents were from the SSE,
19
not the SW. Two other instabilities were sensed ~500 m south of the nearest known active MEF
vent (Salut) about 110 m above the Quebec diffuse flow site. Particularly strong currents at the
time may explain why no high-temperature vents were discovered in the Quebec area directly
below the instabilities. The hydrogen concentration in a water sample from the same plume is
high (30.3 nM) [McLaughlin, 1998] for an equilibrated plume sample, supporting the proposition
that the plume was buoyant and its source was probably closer than the MEF vents. The sea
floor in the vicinity hosts abundant, inactive sulfide structures, and weathered sulfide talus, but
no focussed, high temperature, hydrothermal activity has been witnessed during the few
submersible transits of the vicinity (J. R. Delaney, personal communication). The instability is
associated with a strong decrease in c, ruling out the possibility that it originated from Quebec,
which vents clear fluid. The final two instabilities imply a source may exist E of the MEF near
the center of the axial valley.
Mid-way between the Salty Dawg and High-Rise fields (Figure 5C), three instabilities
detected at the intersection of two transects cluster 120-220 m above the bottom. Westward
currents suggest that advection from known sources in the adjacent fields is unlikely. The
nearest ALVIN dive (2931) began about 300 m southwest of the cluster, but witnessed no
evidence of nearby hydrothermal venting and sampled only fresh basalt.
The three instabilities detected near the sea floor on the east flank of the central ridge
include a moderate instability (2t1612d, Figure 5A) ~1.5 km ESE of the Mothra area. There, the
unstable fluid is isolated from axial sources bathymetrically and isopycnally. Located only ~10
m (PDR) above the 2250 isobath, the instability is not only beneath the effluent layers of known
fields, but also lower than the depth of most known vents. The vertical extent and magnitude of
the density inversion (Figure 3D) is greater than that seen near Cirque and Dune (Figure 3C),
20
suggesting that this hydrothermal site is more vigorous than Cirque and Dune. The two other
near-bottom instabilities on the east flank were detected E of the southern axial cluster and ESE
of MEF.
Similar, near-bottom instabilities overlie the west flank of the central ridge between the
latitude of HRF and MEF. While current measurements and the  and c signatures evidence
significant contamination from axial sources, the fluid is clearly unstable and too distant from the
axial valley to have originated there.
Discussion
A first-order geological problem in ridge crest science is characterization of the two- and
three-dimensional nature of the circulation of hydrothermal fluid within the crust, and
identification of the processes that control the convective heat and water fluxes [Johnson, 1993].
It is not yet clear how fundamental factors – spreading rate, heat source geometry, or the
interactions of tectonic and volcanic processes – govern the subsurface permeability structure of
the crust, and thereby the subsurface geometry and surface expression of axial and near-axis
convection [Fornari and Embley, 1995]. Focussed, high-temperature, axial up flow has been
observed globally at ~20 sites [Von Damm, 1995], but its distribution varies widely. On the East
Pacific Rise (9o09’-54’ N), 19 out of 20 high-temperature vents are located with 20 m of the
<200 m wide axial summit caldera [Haymon et al., 1991], but on the Mid-Atlantic Ridge the
high temperature vents of the TAG area are ~1.5 km from the bathymetric rift valley axis [Rona
et al., 1986].
The Endeavour Segment presents an unusual opportunity to juxtapose extant models of
hydrothermal circulation with a suite of observations gathered at the axis, on the distant (>20 km
off axis) flanks, and to a limited extent in between. The distribution of instabilities maps the
21
surface expression of convective up flow, providing additional insight into the three-dimensional
organization of the subsurface circulation and helping to discern which models best explain the
hydrothermal system at Endeavour.
The cooling of young oceanic crust as it moves off axis is accomplished primarily
through hydrothermal circulation [e.g., Wilcock, 1998], and is expected to occur in two
convective modes: first active and then passive [Fehn et al., 1983; Lister, 1982]. Near the
Endeavour axis, the observed magnitude of the heat flux through individual vent fields is
difficult to reconcile with conductive heat transfer from a persistent magma chamber to a
hydrothermal fluid [Wilcock and Delaney, 1996]. An alternative mechanism capable of
providing the measured heat flux is active convection, in which cold sea water cools hot crustal
rock until it cracks, allowing deeper circulation and further cooling [Lister, 1982]. Although
conductive off-axis heat flow surveys [e.g., Stein, 1994] show that convection continues long
after the crust is capped with relatively impermeable sediments, it is uncertain at what age (or
distance from the axis) cracking ceases and convection becomes passive [Wilcock and Delaney,
1996].
All models of hydrothermal circulation in the oceanic crust depend on assumptions about
the formation and evolution of permeability [Fisher, 1998]. At one extreme is the idea that the
upper oceanic crust can be modeled as a infinite horizontal porous medium with a homogenous
isotropic (bulk) permeability in which fluids circulate in uniformly spaced convection cells of
equal aspect ratio. At the other is the possibility that the permeability structure of the crust is
highly heterogeneous and hydrothermal circulation is correspondingly complex, organized by the
primary stratigraphy and topography of the crust.
22
Past hydrographic surveys of the Endeavour Segment [Kadko et al., 1990; Lilley et al.,
1995b; Lupton et al., 1985; Thomson et al., 1992], side scan sonar maps (Bhat et al., in
preparation), and dive programs (Kelley et al., in preparation) revealed prominent and potential
vent fields. Their approximately uniform ~2 km spacing along the axis suggests that convection
is organized in cells with similar axial width [Delaney and Crest 91-Flange Research Teams,
1991; Wilcock, 1998]. The depth to which fluids circulate in axis-parallel cells is constrained
through a recent microseismic survey [Archer et al., 1996] which infers that a heat source ~2 km
wide underlies the axial valley floor at a depth of ~3-4 km.
Geophysical and geochemical variations across the Endeavour Segment indicate that
circulation may be organized in axis-perpendicular cells. Sea water may enter the crust off axis
along inward-facing faults on the outer margins of the inter-ridge basins [Johnson et al., 1993].
Based on deep-towed magnetic transects of the Endeavour axis, Tivey and Johnson [1987]
propose that ridge-perpendicular hydrothermal down flow of oxygenated sea water into
extensively fissured crust could induce low temperature oxidation of magnetic minerals, causing
local magnetic lows (in the axial valley and inter-ridge basins) within the central anomaly
magnetic high. Across-axis gravimeter surveys reveal that the axial ridge is less dense (more
porous) than adjacent ridges, and narrow regions of low crustal density are also evident in the
outer margins of the inter-ridge valleys [Holmes and Johnson, 1993]. While it is possible that
off axis down flow supplies only off axis up flow (in along strike circulation cells), isotopic
geochemical differences between axial vent fluids and available surface basalt samples suggest
that axis-perpendicular circulation may incorporate fluids which have reacted with off axis
basalts into axial up flow (K. Bain, in preparation).
23
The instability map (Figure 5) shows that near-axis up flow is located on the outer slopes
of the central ridge near the 2200 m isobath, ~1.5 km from the apparent hydrothermal (as
opposed to bathymetric) axis. The vent locations inferred from axial instabilities are rarely
separated along strike by more than ~1.5 km and intermittently group into axially elongated
clusters. The similarity of this separation of axial and near axis up flow with the along-axis
separation of up flow implies a simple three-dimensional conceptualization of the entire
Endeavour Segment ridge crest hydrothermal system is offered by porous medium models which
assume either (1) that the upper ~0.75 km of ocean crust is an approximately isotropic crack
network and convection cells have an aspect ratio of ~1, or (2) that the upper 3 km of ocean crust
is an approximately isotropic crack network and convection cells have an aspect ratio of ~0.25.
The revised geometry is more consistent with the aspect ratio predicted to maximize heat
transport in a horizontally infinite medium at moderate to high Rayleigh numbers [Wilcock,
1998]. It is interesting to juxtapose this idea with the observation that heat flow variations ~30
km off axis evidence convection cells in the crust spaced ~1.65 km apart (aspect ratio of 1-1.25
in a layer ~600 m thick), a separation that was successfully simulated with a uniform
permeability circulation model [Davis et al., 1996].
In contrast, other permeability models of near axis convection can generate up flow zones
spaced 6-7 km apart [Fehn et al., 1983; Travis et al., 1991]. Conductive heat flow measurements
made along an intermittent transect of the western flank of the Endeavour (2.5-25 km) reveal
maxima spaced ~6 km apart atop near-axis ridges, but the high heat flow zones were attributed
not to uniform permeability and cell spacing, but to the importance of topography in focussing
up flow and high permeability faults in channeling down flow [Johnson et al., 1993]. On older
(~1-4 Ma) Endeavour crust, ~20-120 km off axis, hydrothermal effluent exits through basement
24
outcrops on otherwise sedimented ridge flanks and pore water analysis indicates that crustal
topography focuses the hydrothermal up flow [Wheat and Mottl, 1994].
The presence of up flow on the outer slopes is surprising because the geological
conditions there are profoundly different from those in the axial valley where high temperature
axial venting occurs. The volcanic crest beyond the axial valley and the outer slopes have been
mapped in side scan sonar surveys as conspicuously unfractured, overlapping flows [Kappel and
Ryan, 1986]. While the crest of the ridge exhibits a smooth acoustic backscatter signature
indicative of lobate and sheet flows, the steeper outer slopes have a bulbous texture indicative of
pillow flows (Bhat et al., in preparation). The outer slopes have been characterized geologically
(based on visual observations of the sea floor only WNW of Salty Dawg) as a single, completely
unfissured, lobate pillow flow [Tivey and Johnson, 1987]. Could fluid pass through strata that
are unbroken at the surface and emerge with up flow spaced along- and across-axis in
accordance with a uniform permeability model of hydrothermal circulation in the upper crust?
A striking number of the vent locations inferred in this study are approximately aligned
with the 2200 m isobath from HRF south along the western wall of the axial valley. This is
compelling evidence that high permeability normal faults along the axial summit caldera margins
channel hot fluids to the sea floor, a concept that also explains hydrothermal distributions at slow
[Rona et al., 1986] and fast spreading ridges [Haymon et al., 1991]. The apparent elongation of
the Main Endeavour and Mothra vent fields reinforces the idea that permeability beneath major
vent fields is probably anisotropic, ~10 times more permeable along- than across-axis [Wilcock
and McNabb, 1996] due to the prevalence of axis-parallel fractures. Future analysis of
Endeavour segment surface geology may strengthen the correlation of vent location with the
25
oblique intersection of large fractures and faults [Delaney et al., 1992], or with fissure density
[Wright et al., 1995].
Conclusions
This study significantly changes our understanding of the regional pattern of
hydrothermal venting on the Endeavour Segment of the Juan de Fuca Ridge. In particular, the
observation of unstable fluid beyond the perimeters of the known fields, in the extreme south,
and outside the axial valley demonstrates that hydrothermal activity is more pervasive than
previously discerned. The distribution of up flow appears to reflect a subsurface circulation that
is more complex than up flow zones spaced ~2 km apart along the axis. The instabilities
encountered midway between known fields imply that the aspect ratio of circulation cells in the
axial plane may be higher than previously envisioned.
Off axis thermal or light transmission features can be mistakenly attributed to advection
from axial sources. Some instabilities discerned in this work indicate that particular off axis
thermal features may instead come from moderate to low temperature vents located on the valley
rims and the outer slopes of the central ridge. Their existence implies either that active
hydrothermal circulation occurs over a heat source which extends ~1.5 km off axis, that
convective up flow centered on the axis is diverted in the subsurface beyond the axial valley, or
that passive up flow is established within ~1.5 km of the axis.
The theoretical residence time of instabilities in the water column combined with plume
trajectory modeling enables more precise hydrothermal prospecting than is possible through sole
dependence on potential temperature and light transmission anomalies. This is the utility of
stability, but the method requires an exhaustive spatial survey that can be focussed only through
26
measuring the distributions of more-conservative tracers while modeling or measuring current
velocity fields, a daunting operational task. However, stability surveys may help to locate vents
more efficiently in combination with new acoustic methods [Rona et al., 1997], the proven
diagnostic power of towed camera surveys [Haymon et al., 1991], seismic detection of lithified
sediments [Rohr and Schmidt, 1993], and in situ chemical samplers (G. Massoth, personal
communication).
Inter-comparison of the instability distributions with both geologic maps of the
Endeavour Segment derived from side scan sonar data, and seismic constraints on heat sources
and subsurface fault zones promises to clarify which factors govern circulation within the young
oceanic crust at intermediate spreading rates. Physical, chemical, and microbiological
investigation of the inferred up flow sites, in particular those located off-axis, may elucidate the
nature of a hydrothermal system that appears to be more complicated than previously recognized.
27
Acknowledgements: This paper is dedicated to Mary Landsteiner. Especially during long
nights at sea, her reserves of humor, energy, optimism, and scientific skills seemed boundless –
somehow unfettered by her cancer. With them, she ultimately enabled the success of much
oceanographic research, and the Mixing Zephyrs CTDT program in particular. Rick Thomson
kindly provided the current meter data. We thank Deborah Kelly, Val Veirs, and Venugopal
Bhat for helpful reviews. We gratefully acknowledge Paul Johnson and Rick Thomson for
helpful discussions, and Geoff Lebon for his excellent lessons in CTDT deck logistics. Abdellah
Cherkaoui processed the navigational data and worked the night shifts along with CTDT team
members: Christian Parker, Deborah Kelley, Venugopal Bhat, Katherine Bain, Melanie Summit,
and Steve Squyres. At sea, the captain, crew, and technicians of the R/V Atlantis II heroically
maintained position and instrumentation. On land, Perl, Generic Mapping Tools, and Matlab
were crucial in data processing and figure generation. S. Veirs was supported throughout this
research project by a NASA Earth Systems Science Fellowship.
28
References
Archer, S., W.S.D. Wilcock, and G.M. Purdy, Microearthquake activity associated with
hydrothermal circulation on the Endeavour Segment of the Juan de Fuca Ridge, EOS
Transactions AGU, 77(46,supplement), 727, 1996.
Baker, E.T., and G.A. Cannon, Long-term monitoring of hydrothermal heat flux using moored
temperature sensors, Cleft Segment, Juan de Fuca Ridge, Geophysical Research Letters,
20 (17), 1855-1858, 1993.
Baker, E.T., C.R. German, and H. Elderfield, Hydrothermal plumes over spreading-center axes:
Global distributions and geological inferences, in Seafloor Hydrothermal Systems:
Physical, Chemical, Biological, and Geological Interactions, Geophys. Monogr. Ser.,
edited by S.E. Humphris, R.A. Zierenberg, L.S. Mullineaux, and R.E. Thomson, pp. 4771, American Geophysical Union, Washington, D.C., 1995.
Baker, E.T., J.W. Lavelle, and G.J. Massoth, Hydrothermal particle plumes over the southern
Juan de Fuca Ridge, Nature, 316 (25 July), 342-344, 1985.
Baker, E.T., and G.J. Massoth, Characteristics of hydrothermal plumes from two vent fields on
the Juan de Fuca Ridge, northeast Pacific Ocean, Earth and Planetary Science Letters,
85, 59-73, 1987.
Bhat, V., P. Shaw, and J.R. Delaney, The Endeavour Ridge as observed by DSL120 sidescan
sonar: Interplay of volcanic, tectonic, and hydrothermal processes, EOS Transactions
AGU, 1997.
Brikowski, T., and D. Norton, Influence of magma chamber geometry on hydrothermal activity
at mid-ocean ridges, Earth and Planetary Science Letters, 93, 241-255, 1989.
29
Butterfield, D.A., R.E. McDuff, M. Mottl, M.D. Lilley, G.J. Massoth, and J.E. Lupton, Gradients
in the compostition of hydrothermal fluids from the Endeavour Ridge vent field:
Supercritical phase separation and brine loss, Journal of Geophysical Research, 99, 95619583, 1994.
Cannon, G.A., D.J. Pashiniski, and M.R. Lemon, Hydothermal effects west of the Juan de Fuca
Ridge, Deep Sea Research, 40 (1447-1457), 1993.
Cannon, G.A., and R.E. Thomson, Characteristics of 4-day oscillations trapped by the Juan de
Fuca Ridge, Geophysical Research Letters, 23 (13), 1613-1616, 1996.
D'Asaro, E., S. Walker, and E. Baker, Structure of two hydrothermal megaplumes, Journal of
Geophysical Research, 99 (C10), 20,361-20,373, 1994.
Davis, E.E., D.S. Chapman, and C.B. Forster, Observations concerning the vigor of
hydrothermal circulation in young oceanic volcanic crust, Journal of Geophysical
Research, 101, 2927-2942, 1996.
Delaney, J.R., and Crest 91-Flange Research Teams, Jason/Alvin operations on the Endeavour
Segment, Juan de Fuca Ridge, summer 1991, EOS Transactions AGU, 77, 231, 1991.
Delaney, J.R., V. Robigou, R.E. McDuff, and M.K. Tivey, Geology of a Vigorous Hydrothermal
System on the Endeavour Segment, Juan de Fuca Ridge, Journal of Geophysical
Research, 97 (B13), 19663-19682, 1992.
Elderfield, H., and A. Schultz, Mid-ocean ridge hydrothermal fluxes and the chemical
composition of the ocean, Annual Review of Earth and Planetary Science, 24, 191-224,
1996.
30
Fehn, U., K.E. Green, R.P.V. Herzen, and L.M. Cathles, Numerical models for the hydrothermal
field at the Galapagos spreading center, Journal of Geophysical Research, 88 (B2), 10331048, 1983.
Fisher, A.T., Permeability within basaltic oceanic crust, Reviews of Geophysics, 36 (2), 143-182,
1998.
Fornari, D.J., and R.W. Embley, Tectonic and volcanic controls on hydrothermal processes at the
mid-ocean ridge: An overview based on near-bottom and submersible studies, in Seafloor
Hydrothermal Systems: Physical, Chemical, Biological, and Geological Interactions,
Geophys. Monogr. Ser., edited by S.E. Humphris, R.A. Zierenberg, L.S. Mullineaux, and
R.E. Thomson, pp. 1-46, American Geophysical Union, Washington, D.C., 1995.
Haymon, R.M., D.J. Fornari, M.H. Edwards, S. Carbotte, D. Wright, and K.C. MacDonald,
Hydrothermal vent distribution along the East Pacific Rise crest (9o09'-54'N) and its
relationship to magmatic and tectonic processes on fast-spreading mid-ocean ridges,
Earth and Planetary Science Letters, 104, 513-534, 1991.
Holmes, M.L., and H.P. Johnson, Upper crustal densities derived from sea floor gravity
measurements: Northern Juan de Fuca Ridge, Geophysical Research Letters, 20 (17),
1871-1874, 1993.
Johnson, H.P., Processes associated with ocean crustal formation: The Juan de Fuca Ridge,
Geophysical Research Letters, 20 (17), 1847-1850, 1993.
Johnson, H.P., K. Becker, and R.V. Herzen, Near-axis heat flow measurements on the northern
Juan de Fuca ridge: Implications for fluid circulation in oceanic crust, Geophysical
Research Letters, 20 (17), 1875-1878, 1993.
31
Kadko, D.C., N.D. Rosenberg, J.E. Lupton, R.W. Collier, and M.D. Lilley, Chemical reaction
rates and entrainment within the Endeavour Ridge hydrothermal plume, Earth and
Planetary Science Letters, 99, 315-335, 1990.
Kappel, E.S., and W.B.F. Ryan, Volcanic episodicity and a non-steady state rift valley along
northeast Pacific spreading centers: Evidence from Sea MARC I, Journal of Geophysical
Research, 91(B14), 13925-13940, 1986.
Lilley, M.D., R.A. Feely, and J.H. Trefry, Chemical and biochemical transformations in
hydrothermal plumes, in Seafloor Hydrothermal Systems: Physical, Chemical,
Biological, and Geological Interactions, Geophys. Monogr. Ser., edited by S.E.
Humphris, R.A. Zierenberg, L.S. Mullineaux, and R.E. Thomson, pp. 369-391,
American Geophysical Union, Washington, D.C., 1995a.
Lilley, M.D., M.C. Landsteiner, E.A. McLaughlin, C.B. Parker, A.S.M. Cherkaoui, G. Lebon,
S.R. Veirs, and J.R. Delaney, Real-time mapping of hydrothermal plumes on the
Endeavour Segment of the Juan de Fuca, Eos Transactions, AGU, 76, 420, 1995b.
Lister, C.R.B., "Active" and "passive" hydrothermal systems in the oceanic crust: Predicted
physical conditions, in The Dynamic Environment of the Ocean Floor, edited by K.A.
Fanning, and F.T. Manheim, University of Miami, 1982.
Little, S.A., K.D. Stolzenbach, and F.J. Grassle, Tidal current effects on temperature in diffuse
hydrothermal flow: Guaymas Basin, Geophysical Research Letters, 15 (13), 1491-1494,
1988.
Lupton, J.E., Hydrothermal plumes: Near and far field, in Seafloor Hydrothermal Systems:
Physical, Chemical, Biological, and Geological Interactions, Geophys. Monogr. Ser.,
32
edited by S.E. Humphris, R.A. Zierenberg, L.S. Mullineaux, and R.E. Thomson, pp. 317346, American Geophysical Union, Washington, D.C., 1995.
Lupton, J.E., E.T. Baker, N. Garfield, G.J. Massoth, R.A. Feely, J.P. Cowen, R.R. Greene, and
T.A. Rago, Tracking the evolution of a hydrothermal event plume with a RAFOS
neutrally buoyant drifter, Science, 280 (15 May), 1052-1055, 1998.
Lupton, J.E., J.R. Delaney, H.P. Johnson, and M.K. Tivey, Entrainment and vertical transport of
deep-ocean water by buoyant hydrothermal plumes, Nature, 316, 621-623, 1985.
McDougall, T.J., Bulk properties of "hot smoker" plumes, Earth and Planetary Science Letters,
99, 185-194, 1990.
McDuff, R.E., Physical dynamics of deep-sea hydrothermal pumes, in Seafloor Hydrothermal
Systems: Physical, Chemical, Biological, and Geological Interactions, Geophys. Monogr.
Ser., edited by S.E. Humphris, R.A. Zierenberg, L.S. Mullineaux, and R.E. Thomson, pp.
357-368, American Geophysical Union, Washington, D.C., 1995.
McLaughlin, E.A., Microbial hydrogen oxidation associated with deep-sea hydrothermal vent
environments, University of Washington, Seattle, 1998.
Middleton, J.H., Times of rise for turbulent forced plumes, Tellus, 31, 82-88, 1979.
Middleton, J.H., The rise of forced plumes in a stably stratified crossflow, Boundary-Layer
Meteorology, 36, 187-199, 1986.
Middleton, J.H., and R.E. Thomson, Modelling the rise of hydrothermal plumes, in Canadian
Technical Report of Hydrography and Ocean Sciences, Institute of Ocean Sciences,
Department of Fisheries and Oceans, Sidney, B.C. V8L 4B2, 1986.
Morton, B.R., G. Taylor, and J.S. Turner, Turbulent gravitational convection from maintained
and instantaneous sources, Proc. R. Soc. Lond. A, 366, 63-79, 1956.
33
Pond, S., and G.L. Pickard, Introductory Dynamical Oceanography, 329 pp., Pergamon Press,
Tarrytown, New York, 1983.
Robigou, V., J.R. Delaney, and D.S. Stakes, Large massive sulfide deposits in a newly
discovered active hydrothermal system, the High-Rise Field, Endeavour Segment, Juan
de Fuca Ridge, Geophysical Research Letters, 20 (17), 1887-1890, 1993.
Rohr, K.M.M., and U. Schmidt, Regional patterns of hydrothermal alteration of sediments as
interpreted from seafloor reflection coefficients, Middle Valley, Juan de Fuca Ridge,
Geophysical Research Letters, 20 (17), 1867-1870, 1993.
Rona, P.A., D.R. Jackson, T. Wen, C. Jones, K. Mitsuzawa, K.G. Bemis, and J.G. Dworski,
Acoustic maping of diffuse flow at a seafloor hydrothermal site: Monolith Vent, Juan de
Fuca Ridge, Geophysical Research Letters, 24 (19), 2351-2354, 1997.
Rona, P.A., G. Kinkhammer, T.A. Nelson, J.H. Trefry, and H. Elderfield, Black smokers,
massive sulphides and vent biota at the Mid-Atlantic Ridge crest at 26oN, Nature, 321,
33-37, 1986.
Rona, P.A., and D.A. Trivett, Discrete and diffuse heat transfer at ASHES vent field, Axial
Volcano, Juan de Fuca Ridge, Earth and Planetary Science Letters, 109, 57-71, 1992.
Rosenberg, N.D., F.J. Spera, and R.M. Haymon, The relationship between flow and permeability
field in seafloor hydrothermal systems, Earth and Planetary Science Letters, 116, 135153, 1993.
Speer, K.G., and P.A. Rona, A model of an Atlantic and Pacific hydrothermal plume, Journal of
Geophysical Research, 94 (C5), 6213-6220, 1989.
Stein, C.A., and S. Stein, Constraints on hydrothermal heat flux through the oceanic lithosphere
from global heat flow, Journal of Geophysical Research, 99 (B2), 3081-3095, 1994.
34
Thomson, R.E., E.E. Davis, and B.J. Burd, Hydrothermal venting and geothermal heating in
Cascadia Basin, Journal of Geophysical Research, 100 (B4), 6121-6141, 1995.
Thomson, R.E., J.R. Delaney, R.E. McDuff, D.R. Janecky, and J.S. McClain, Physical
characteristics of the Endeavour Ridge hydrothermal plume during July 1988, Earth and
Planetary Science Letters, 111, 141-154, 1992.
Tivey, M.A., and H.P. Johnson, The central anomaly magnetic high: Implications for ocean crust
construction and evolution, Journal of Geophysical Research, 92 (B2), 12,685-12,694,
1987.
Travis, B.J., D.R. Janecky, and N.D. Rosenberg, Three-dimensional simulation of hydrothermal
circulation at mid-ocean ridges, Geophysical Research Letters, 18 (8), 1441-1444, 1991.
Trivett, D.A., Effluent from diffuse hydrothermal venting: 1. A simple model of plumes from
diffuse hydrothermal sources, Journal of Geophysical Research, 99 (C9), 18403-18415,
1994.
Trivett, D.A., and A.J.Williams III, Effluent from diffuse hydrothermal venting 2. Measurement
of plumes from diffuse hydrothermal vents at the southern Juan de Fuca Ridge, Journal
of Geophysical Research, 99 (C9), 18417-18432, 1994.
Von Damm, K.L., Controls on the chemistry and temporal variability of seafloor hydrothermal
fluids, in Seafloor Hydrothermal Systems: Physical, Chemical, Biological, and
Geological Interactions, Geophys. Monogr. Ser., edited by S.E. Humphris, R.A.
Zierenberg, L.S. Mullineaux, and R.E. Thomson, pp. 222-247, American Geophysical
Union, Washington, D.C., 1995.
35
Wheat, C.G., and M.J. Mottl, Hydrothermal circulation, Juan de Fuca ridge eastern flank: Factors
controlling basement water composition, Journal of Geophysical Research, 99 (B2),
3067-3080, 1994.
Wilcock, W.S.D., Cellular convection models of mid-ocean ridge hydrothermal circulation and
the temperatures of black smoker fluids, Journal of Geophysical Research, 103 (B2),
2585-2596, 1998.
Wilcock, W.S.D., and J.R. Delaney, Mid-ocean ridge sulfide deposits: Evidence for heat
extraction from magma chambers or cracking fronts?, Earth and Planetary Science
Letters, 145, 49-64, 1996.
Wilcock, W.S.D., and A. McNabb, Estimates of crustal permeability on the Endeavour segment
of the Juan de Fuca mid-ocean ridge, Earth and Planetary Science Letters, 138, 83-91,
1996.
Wright, D.J., R.M. Haymon, and D.J. Fornari, Crustal fissuring and its relationship to magmatic
and hydrothermal processes on the East Pacific Rise crest (9o12'-54'N), Journal of
Geophysical Research, 100 (B4), 6097-6120, 1995.
36
Figure Captions
Figure 1: SEABEAM bathymetric map (50 m contours) of the study site on the Endeavour
Segment of the Juan de Fuca Ridge showing the axial ridge crest and a half-ridge to the west
(outlined by the bold 2200 m contour). The 1-2 km wide half-ridges have steeper slopes facing
the spreading axis and are separated by 5-6 km wide, ~300 m deep basins. Bold rectangles
bound known fields of focussed hydrothermal venting: Mothra (Mo), Main Endeavour (Ma),
High-Rise (HR) and Salty Dawg (SD). Smaller squares mark diffuse flow sites: Quebec (Q),
Clam Bed (CB), Cirque (C), and Dune (D). The areal coverage of the hydrographic survey is
indicated by the underlying white tow traces.
Figure 2: Bold contours of potential density (2) over the potential temperature anomaly ()
field (grey shades and contours) measured during a vertically oscillating tow through the axial
valley over the Main Endeavour and Mothra areas demonstrate that thermal anomalies and
instabilities do not necessarily coincide. Equilibrated (stable) effluent above the Main
Endeavour field (~2000 m depth) has a  signature indistinguishable from that of fluid high
above the Mothra field (~2100 m depth) that contains an instability (upper circle) which proves it
is still rising from its source. In contrast, the serendipitous transect of a buoyant plume directly
above an intense thermal source in the Mothra vicinity exhibits both exceptional thermal
anomalies and instabilities (lower circles). The instrument path (grey saw-tooth line) follows the
SEABEAM bathymetry and traverses isopycnals that are generally near horizontal, but can
reveal strong inversions. The indicated location and extent of the Main and Mothra fields are
approximate.
37
Figure 3: Depth profiles of raw (1 Hz) potential temperature (,oC), potential density (2,
kg/m3), and % light transmissivity (c) data, through plumes that contain instabilities above (A)
the sulfide structure Dudley in the Main Endeavour vent field, (B) a source of high temperature,
clear effluent in the Mothra area, (C) the up slope vicinity of diffuse sites Cirque and Dune, and
(D) a potential vent site on the outer slope of the southeast ridge crest. Reference cast names
(above each plot) to Figure 5 for bathymetric orientation.
Figure 4: Plume trajectories determined using a model of plume rise in an ambient cross flow
[Middleton, 1986]. While the modeled rise height (Z) is sensitive to (A) ambient cross flow
velocity, U, (B) entrainment ratio, , (C) stratification, characterized by buoyancy frequency, N,
and (D) source radius, bo, the downstream distance (X) after which instabilities are unlikely is
most sensitive to U and N. The dashed trajectory common to all four plots has U=0.14 m/s,
=0.33, N=0.002 s-1, and bo=0.02 m. As illustrated in (A), an instability observed at an altitude
of Z~75 m when U=0.14 m/s indicates that its source is located ~120 m upstream.
Figure 5: Plan view bathymetric map indicating the location of each instability (+) detected as
well as the bearing (solid black line) and maximum range (open circle) to its source. The
regional instability distribution (A) is shown relative to the known venting areas (grey boxes,
labeled as in Figure 1) and the survey coverage (underlying light grey traces). The detailed
relationship between observations of sea floor geology, the instabilities, and the inferred
hydrothermal source locations is depicted for the vicinity of (B) Main Endeavour and (C) Salty
38
Dawg, Cirque, and Dune. The position of the current meter mooring is indicated by a triangle
ENE of Main Endeavour. Select instabilities are scrutinized in the depth profiles of Figure 3.