T) of the
plume relative to ambient water of the same potential density is calculated
using the formula [Lupton
et al., 1985]U.S. Dept. of Commerce / NOAA / OAR / PMEL / Publications
When the megaplume at the north end of the Cleft segment of the Juan de Fuca
Ridge was discovered in August 1986, an extensive steady-state hydrothermal
plume was also found at the same location [Baker
et al., 1987; Lupton
et al., 1989]. The steady-state plume has been annually mapped by
near-bottom conductivity-temperature-depth transmissometer (CTDT) tows and casts,
and sampled for He and other constituents using rosette-mounted Niskin bottles.
The temperature anomaly ( T) of the
plume relative to ambient water of the same potential density is calculated
using the formula [Lupton
et al., 1985]
where 
and 
are the potential temperature and potential density, respectively, in the plume,
and m and b are the slope and intercept, respectively, of the
linear regression of the -
curve of water immediately above the plume where the hydrographic effect of
hydrothermal emissions is negligible (Fig. 1).
The slope m is constant at -4.865°C m
Kg
each year, whereas the intercept b
is adjusted to compensate for changes in
arising from annual variations of ±0.002
in the
salinity calibration. The calculated T
values may be an underestimate of the true hydrothermal heating by as much as
a factor of two owing to cooling of the plume by entrainment of cold bottom
water [Baker
et al., 1989; Speer
and Rona, 1989], and hydrographic aliasing if vent waters have a salinity
substantially different from that of ambient sea water [McDuff,
1988]. Despite some uncertainty in the absolute value of
T, we assume that the year-to-year values are correct relative to each
other because the consistency of the ambient hydrographic structure and the
rise height of the plume (Fig. 1) implies a consistency
of plume entrainment processes and vent-fluid salinity.
Fig. 1. Potential density plotted against potential temperature for selected
axial valley CTDT casts in 1986, 1987 and 1988. Each curve has a slope of -4.865°C
m Kg (shown
by the straight line) except where hydrothermally affected. 1986 casts in the
axial valley sampled only a thin horizon of ambient water (~27.635-27.640 )
between the steady-state plume and the overlying megaplume. Density of the water
in the axial valley was significantly lower in 1986 because of enhanced vertical
mixing of the lowermost 1,000 m by the megaplume discharge. Such vertical mixing
does not affect the ambient -
relationship. The maximum temperature anomaly, defined as the temperature deviation
from the background trend along an isopycnal, was about the same on each cast.
Insets show vertical profiles of the temperature and light-attenuation (proportional
to the concentration of fine-grained hydrothermal precipitates) anomalies for
each cast. Increased anomalies above 1,950 m in 1986 indicate the megaplume
presence. Depth of the steady-state plume maximum was ~2,100 m each year. For
each cast the light-attenuation and temperature anomalies go to zero at the
same depth, supporting our interpretation of the vertical distribution of temperature
anomaly.
Figure 2 shows areal maps of T
in the steady-state plume created by plotting and contouring maximum T
values along the tow track-lines and at the vertical cast locations. The size
and intensity of the T plume indicate
that venting extends more than 20 km along the ridge axis with a vigour comparable
to the vent field at 48°N on the Endeavour segment of the Juan de Fuca Ridge,
where the heat flux is calculated to be 1-5 × 10
W [Rosenberg
et al., 1988; Baker
and Massoth, 1987]. When the megaplume vent field was discovered in 1986,
T values 
0.04°C
extended from a few kilometres west of the axial valley to the limits of the
survey area to the east and north (Fig. 2a).
Plume mapping in 1986 was not detailed enough to locate precisely the vent field
creating the plume, but the fact that the highest light-attenuation anomalies
were found in the axial valley near station SC22 suggests that the source was
in the axial valley and the plume distribution was the result of advection to
the east.
Fig. 2. Areal maps of the steady-state plume as described by the maximum
temperature anomaly between 2,000 m and 2,200 m along CTDT tow tracks (dashed
lines) and vertical casts (crosses or solid symbols) in 1986, 1987 and 1988.
He samples were collected during tows (open
symbols) or on vertical casts (solid symbols); the locations of the Fig. 1 profiles
are indicated. Hatched areas are the walls of the axial valley as defined by
the 2,200-m bathymetric contour. In each year the plume was ~200 m thick and
centred ~150 m above the floor of the axial valley (see Fig. 1). The plume was
more extensive in 1986 than in either 1987 or 1988. The size and location of
the 1986 megaplume is shown by shading.
In 1987 and 1988 the T plumes were
less extensive and more symmetrical about the axial valley (Fig.
2b, c). Temperature anomalies in the axial valley averaged 0.03-0.04°C with
localized highs reaching 0.06°C. Although quantitative flux estimates cannot
be derived from plume maps alone, it seems likely that the release of the August
1986 megaplume was accompanied by a higher heat and mass flux in the underlying
steady-state plume than has since been characteristic of this vent field.
Unequivocal evidence for temporal variability in hydrothermal processes at
this site comes from the progressive decrease in the He/T
ratio of the steady-state plume (Fig. 3). In
August 1986, the He/T
ratio in the plume was (4.4 ± 0.64) × 10
cm STP g
°C (1
uncertainty), the highest ratio yet reported for hydrothermal emissions on a
mid-ocean ridge and a factor of 14 higher than the He/
T ratio in the overlying megaplume [Lupton
et al., 1989]. Concentrations of He
in the plume were as high as 0.30 × 10 cm
STP g (
He
= 238 ; He
= [(He/
He)
/(He/He)
]-1).
In September 1987 the ratio in the steady-state plume decreased to (2.4 ± 0.14)
× 10 cm
STP g°C
(1), a factor of 1.8 below the 1986 value.
By September 1988 the ratio had decreased further, to (1.3 ± 0.1) × 10
cm STP g°C
(1), a factor of 3.4 below the 1986 value.
The yearly decreases apparently resulted from a decreasing concentration of
He in the plume, as the range of
T values is about the same each year. In the following discussion we use
the conservative assumption that hydrographic cooling and vent-fluid salinity
changes had a negligible influence on the observed T
values, so that the He/T
ratios in the plumes are equivalent to the He/heat
ratios in the original fluids. We cannot verify this because no high-temperature
vents have yet been located at this site.
Fig. 3. He concentration plotted against
temperature anomaly for samples in steady-state plume from 1986, 1987 and 1988.
Sample locations shown in Fig. 2. Least-squares trend of 1986 megaplume samples
(which had a maximum temperature anomaly of ~0.26°C) shown for comparison. Least-squares
regression shows that the He/T
ratio decreased from 4.4 to 2.4 to 1.3 × 10
cm STP g
°C from 1986 to 1988. A t-test analysis
indicates that these ratios are all significantly different from each other
at the 95% confidence level.
We propose that the observed He/heat trend
represents accelerated degassing from a magma body whose solidification rate
was abruptly increased either by vertical intrusion into cooler crust or by
a local deepening of the boundary between brittle and ductile deformation. Sub-surface
releases of He have been reported at magmatically
active terrestrial sites in Japan [Wakita
et al., 1978; Sano
et al., 1986] and Long Valley Caldera in California [Welhan
et al., 1988] in association with earthquakes and local faulting.
Two degassing processes can enhance the expulsion of He from a magma body and
thereby temporarily increase the He/heat ratio
of the surrounding hydrothermal fluids: pressure release or crystal-melt fractionation.
Because of the low and pressure-dependent solubility of C0
in silicate liquids [Stolper
and Holloway, 1988], CO exsolution
forms vesicles in a rising magma. Basalts from the Juan de Fuca Ridge have a
mean vesicularity, of ~1% [Dixon
et al., 1988]. Determinations of the solubility constant of He in
a tholeiitic basalt melt [Jambon
et al., 1986] show that for 1% vesicularity, 60%
of the total He is contained in vesicles and is thus more readily available
for loss from the melt. The discovery of basalts with [He] > 50 × 10
cm STP g,
ten times larger than that of typical mid-ocean ridge basalts [Sarda
and Graham, 1990], suggests that a large proportion of magmatic He is commonly
lost by pre-eruptive degassing.
As the intrusion solidifies, He is progressively concentrated in the remaining
melt by crystal-melt fractionation because the distribution coefficient for
He between crystals and a basaltic melt is 
0.1-0.01
[Kirsten,
1968; Kurz
et al, 1982; Hiyagon
and Ozima, 1986]. Freezing of a basaltic melt at a liquidus temperature
of 1,150°C excludes 90-99% of the He from individual crystals but liberates
only ~100 cal g, ~25% of the total heat
lost on cooling to seawater temperature. A thick and slowly cooling intrusion
might thus supply an elevated, but decreasing, He/heat
flux for several years.
An alternative to magma intrusion is the incremental deepening of the cracking front [Lister, 1983] above a magma chamber (J. R. Delaney, personal communication). A sudden increase in the permeability of the crust, perhaps caused by tectonic stretching, might allow deeper penetration of cold sea water and an accelerating solidification of the magma.
Lupton et al. have suggested that He/heat
variations in mid-ocean-ridge vent systems do not arise from geographic inhomogeneities
in parent magmas but are symptomatic of dynamic conditions such as the age and
permeability of the circulation system feeding the vent field. The changing
He/heat ratios reported here support that prediction
and may thus describe a changing hydrothermal circulation system. A sudden permeability
increase caused by crustal fracturing associated with magma solidification could
accelerate the release of preequilibrated fluids having a "normal"
He/heat ratio, causing a megaplume event. Fluids
remaining in the now rejuvenated, highly permeable, circulation system are enriched
with degassing He faster than they are heated, causing He/heat
ratios to exceed the theoretical ratio of ~2 × 10
cm STP cal
expected for upper mantle magmas. Indeed, the He/heat
ratio in the steady-state plume immediately after the megaplume disruption was
more than twice this theoretical ratio.
As the system ages, circulation approaches equilibrium and the fluid pathways
become more tortuous as a result of mineral precipitation. He/heat
ratios decline as the residence time of fluids in the circulation system increases
and the extraction rates for heat and He converge. The compositionally unchanging
fluids at the 21°N vent field on the East Pacific Rise [Campbell
et al., 1988] have a low He/heat
ratio (0.4 × 10 cm
STP cal) [Welhan
and Craig, 1983; Lupton
et al., 1980] that may reflect equilibrium conditions. Recent studies
of Icelandic geothermal systems [Poreda
et al., in press] and the degassing behaviour of mid-ocean-ridge
magma [Sarda
and Graham, 1990] also suggest that He/heat
ratios can decline as a magmatic system ages. Based on our observations, the
timescale for a return to equilibrium conditions can be of the order of a few
years, in agreement with radiochemical estimates of the residence time of hydrothermal
fluids in the crust [Kadko
and Moore, 1988].
Acknowledgements. We thank G. Massoth for collecting the He samples, S. Walker for processing the CTDT data, K. Thornberry, D. Dion and A. Faizullabhoy for the He isotope measurements, and D. Graham for comments on the manuscript. This work was supported by the NOAA and the NSF.
Go back to abstract or forward to References