methane dynamics in hydrothermal plumes over a superfast ... et al 2005 jgr.pdfb10101 gharib et al.:...

Methane dynamics in hydrothermal plumes over a superfast

spreading center: East Pacific Rise, 27.5�––32.3�SJ. J. Gharib,1 F. J. Sansone,2 J. A. Resing,3 E. T. Baker,4 J. E. Lupton,5

and G. J. Massoth4,6

Received 4 November 2004; revised 19 April 2005; accepted 17 June 2005; published 1 October 2005.

[1] Samples were collected from hydrothermal plumes along the East Pacific Rise (EPR)from 28� to 32�S during the Ridge Axis Plume and Neotectonic Unified Investigation(RAPANUI) cruise (5 March to 12 April 1998). Forty-five vertical casts and tow-yos wereconducted: 3 off axis and 42 over the axes of two overlapping propagating ridges, theWest and East ridges. These ridges are composed of several nontransform offset ridgesegments. Spreading rates range from 149 mm/yr in the segment interiors to 0 mm/yr atthe propagating rifts. These maximum spreading rates are considered the fastest on themid-ocean ridge system and affect the structure of the ridge. Segment-averaged methaneplume maxima ranged from 1.7 to 7.2 nM. Mean methane concentrations on the WestRidge were nearly double those of the East Ridge. Westwardly advecting hydrothermalmethane persisted to our most distal station, nearly 480 km west of the East Pacific Rise.Background concentrations were less than 1 nM. The highest methane concentrationmeasured was 50 nM in a buoyant plume. Methane did not covary with manganese or anyother hydrothermal tracer in plumes over portions of a segment that exhibited recentmagmatic activity, possibly as a result of the hydrothermal system’s recovery from a phaseseparation event. In contrast, methane/manganese ratios on the other segments rangedfrom 0.077 to 0.091. Methane d13C values in plume maxima ranged from �27 to �33%versus Peedee belemnite; background values were around �40%. These data arecompared with hydrothermal plume methane data from slower spreading ridges andillustrate similarities in hydrothermal processes and sources between these systems.

Citation: Gharib, J. J., F. J. Sansone, J. A. Resing, E. T. Baker, J. E. Lupton, and G. J. Massoth (2005), Methane dynamics in

hydrothermal plumes over a superfast spreading center: East Pacific Rise, 27.5�–32.3�S, J. Geophys. Res., 110, B10101,

doi:10.1029/2004JB003531.

1. Introduction

[2] A variety of chemical and physical tracers arerequired for the use of water column surveys to understandridge crest processes. Methane is one of the most dynamicand informative chemical tracers because it has severalpossible sources and sinks. Magmatic methane can bedirectly outgassed from a magmatic melt into hydrothermalfluids or it can become trapped in fluid inclusions in

solidifying rock and then liberated by later contact withhydrothermal fluid; another source of hydrothermalmethane is from the thermogenic reduction of either organichydrocarbons in regions of high sedimentary material, orinorganic carbon (typically CO2 in seawater) [Welhan,1988; Kelley and Fruh-Green, 1996]. Along poorly sedi-mented mid-ocean ridges, the conventional model ofthermogenic reduction is of CO2 reduction by hydrogenprimarily generated from the oxidation of ferrous to ferriciron and in the presence of certain catalyzing metal alloys.The CO2 can be reduced to methane either while trapped influid inclusions or directly within the hydrothermal fluid.The main source of hydrogen to reduce the CO2 at fastspreading ridges is from the oxidation of ferrous iron inbasalt; another hydrogen source is from the serpentinizationof iron-bearing ultramafic minerals, such as that occurringalong slow spreading ridges [Lilley et al., 1982; Welhan andCraig, 1983; Charlou et al., 1991; Charlou and Donval,1993; Yoshida et al., 1993; Berndt et al., 1996; Horita andBerndt, 1999].[3] At present, there is little evidence elucidating whether

microbially produced methane, common in regions of highsediment and organic carbon content, is important in unse-

JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 110, B10101, doi:10.1029/2004JB003531, 2005

1Department of Geology and Geophysics, University of Hawaii,Honolulu, Hawaii, USA.

2Department of Oceanography, University of Hawaii, Honolulu,Hawaii, USA.

3Joint Institute for the Study of Atmosphere and Ocean, University ofWashington, Seattle, Washington, USA.

4PacificMarine Environmental Laboratory, NOAA, Seattle,Washington,USA.

5Pacific Marine Environmental Laboratory, NOAA, Newport, Oregon,USA.

6Institute of Geological and Nuclear Sciences, Ltd., Lower Hutt, NewZealand.

Copyright 2005 by the American Geophysical Union.0148-0227/05/2004JB003531$09.00

B10101 1 of 16

dimented hydrothermal systems [Baross et al., 1982; Lilleyet al., 1982; Welhan and Craig, 1983; Welhan, 1988; Lilleyet al., 1993].[4] Methane is also a useful indicator of hydrothermal

activity because methane analyses can be conducted rapidly,economically, and reliably while at sea. Furthermore, meth-ane concentrations in a vent fluid are typically 105–107

times enriched over background concentrations [Welhanand Craig, 1983; Von Damm, 1990; DeAngelis et al.,1993], and when diluted 104–105 times with ambientseawater [Lupton et al., 1985] by entrainment into a plume,are still enriched 10 to 1000 times over background. Whiledilution of methane-rich hydrothermal fluid continues withtime in a hydrothermal plume, methane concentrations canalso be diminished by microbial oxidation [e.g., Sansoneand Martens, 1978; Coleman et al., 1981; DeAngelis et al.,1993]. It has been demonstrated, however, that there is nodirect correlation between the methane concentrations andthe specific oxidation rates within hydrothermal plumes;apparently, other factors control the rate of methane oxida-tion, and these are still poorly understood [Sansone andMartens, 1978; DeAngelis et al., 1993]. Methane in hydro-thermal plumes has been estimated to have a mean resi-dence time of 11 days [Kadko et al., 1990] to 17 days[DeAngelis et al., 1993]. Thus the ratio of methane with amore conservative species, such as dissolved manganese,with a residence time of several months to a year [Cowen etal., 1990], or a fully conservative species such as 3He[Lupton, 1998], can also be used to provide informationabout the age of a plume.[5] Comparing the ratios of methane to other hydrother-

mal tracers is a powerful method for determining source-dependent and temporal differences among hydrothermalsources. For example, Charlou et al. [1991] proposed thatthe ratio of methane to manganese can be indicative of threedifferent hydrothermal regimes on the seafloor. The differ-ences in venting characteristics that can affect the CH4/Mnratio are (1) variations in the temperature and pressureconditions in which the hydrothermal reactions occur,including possible phase separation; (2) variations in thegeology of the region in which the seawater-rock interac-tions take place; and (3) variations in the fashion ofdischarge from the vent (i.e., diffuse venting, higher tem-perature ‘‘black smoker’’ venting, or cataclysmic ‘‘mega-plume’’ venting) [Charlou et al., 1991].[6] In addition to the relationship between methane and

other hydrothermal tracers, the isotopic composition (d13C)of methane can be used to identify hydrothermal processesaffecting methane distributions in hydrothermal areas. Therange of d13C values in methane from hydrothermal vents istypically from as heavy as �15% to as light as �50% orlighter, relative to the Peedee belemnite (PDB) standard[Welhan, 1988]. Some studies have shown that unsedi-mented mid-ocean ridges produce heavy d13C signatures;for example, vents at 21�N on the East Pacific Rise had asignature ranging from �18 to �15% [Welhan, 1988] and�24 to �22% in fluid end-members between 17�S and19�S [Charlou et al., 1996]. This is comparable to typicald13C values believed to be from unsedimented abiogenicmethane such as �17 to �15% from the Izu-Bonin arc[Tsunogai et al., 2000]. Some heavier values are alsoreported, such as �8.6% from the Rodriguez Triple Junc-

tion in the Indian Ocean [Gamo et al., 2001], from �10 to�16% from the Grimsey hydrothermal field near Iceland[Riedel et al., 2001], and from �6.9 to �13% from theTAG site on the Mid-Atlantic Ridge (MAR) [Charlou et al.,1996]. Methane that has organic-rich sediment as its sourceis isotopically lighter than inorganic methane because of thevery high fractionation that occurs during microbiallymediated methanogenesis. A recent study has experimen-tally produced abiogenic hydrothermal methane with d13Cvalues as light as �50%, suggesting that some abiogenicmethane may overlap the light values typically expected formicrobial production, but the microbial values do not tendto be heavier than �50% [Horita and Berndt, 1999].[7] Along the mid-ocean ridge system, it appears that

there are direct relationships among the spreading rate, themagmatic budget, the axial heat flux, and the coverage of aridge axis by hydrothermal plumes [Baker et al., 1996]. Asspreading rate increases, so does the rate of heat input perunit length along the ridge axis and the frequency ofhydrothermal activity. In theory, the faster the spreading,the higher the rate of heat input per unit length along theridge axis. Furthermore, differences in plate boundary con-ditions, specifically the replacement of transform faultswith propagating rifts and microplates, occur at somequalitative transition between ‘‘fast’’ (less than 136 mm/yr)and ‘‘superfast’’ (greater than 142 mm/yr) spreading ridges[Lonsdale, 1977; Hey et al., 2004]. Hydrothermal systemsassociated with transform faults are often influenced bymantle rocks. This is most apparent at ‘‘slow’’ spreadingridges (spreading rates of less than 55 mm/yr but greater than20 mm/yr, the latter being considered ‘‘ultraslow’’ [Dick etal., 2003; Baker et al., 2004]) where several hydrothermalsites vent fluids that have had considerable interaction withultramafic mantle rocks. The CH4/Mn ratios of these fluidstend to be quite sensitive of this interaction [Charlou et al.,1991], consequently there is a relationship between themethane chemistry and the spreading rate between slowspreading and fast spreading mid-ocean ridges, as a result ofincreased ultramafic rock interaction.

2. Field Site

[8] The samples reported here were collected on theRidge Axis Plume and Neotectonic Unified Investigation(RAPANUI) cruise, which took place on the R/V Melvillefrom 5 March to 12 April 1998. The research conducted wasmultidisciplinary, with both geophysical and water columngeochemical surveys. The field site for this study, a sectionof the East Pacific Rise (EPR) between 28 and 32�S, islocated between the Easter and Juan Fernandez microplatesand contains the largest known pair of overlapping propa-gating ridges [Hey et al., 2004] (Figure 1). These propa-gating ridges strike roughly north-south, with a separationof approximately 120 km. They are the shallowest of anysection along the East Pacific Rise [Hey et al., 1995; andreferences therein], implying a vigorous magma supply[Macdonald and Fox, 1988]. These propagating ridges aretermed the West and East ridges and are further subdividedinto nontransform offset segments, six of which werestudied during this cruise: W3, W4, E1, E2, E3, and E4.[9] The full spreading rate is up to �149 mm/yr near

31�S on the East Ridge, as measured by magnetic anoma-

B10101 GHARIB ET AL.: METHANE DYNAMICS IN HYDROTHERMAL PLUMES

2 of 16

B10101

lies, which is the fastest spreading rate yet reported for anyridge system [Martinez et al., 1997]. Segment W3 has boththe most inflation and the greatest cross-sectional area of thesix segments [Martinez et al., 1997], followed by segmentE4 (F. Martinez, personal communication, 1999); the seg-ments closest to the propagating tips are the least inflated.The rapid spreading and the inflated axial magma chamberimply a high rate of magma activity, which in turn implies avigorous hydrothermal system [Macdonald and Fox, 1988].In addition, differences in plate boundary conditions, suchas the substitution of transform faults by propagating riftsand microplates, have been recognized between fast andsuperfast spreading ridges.

3. Methods

[10] Seawater samples were taken by either of twomethods: vertical casts, in which the water sampler rosettewas lowered into the water column and then raised at onelocation, and tow-yos, in which the rosette was continuallyraised and lowered vertically in the water column as the ship

moved [Baker et al., 1995]. A total of 45 casts were taken;35 were vertical casts and 10 were tow-yos (Figure 1).[11] Once the rosette was brought onboard, samples were

collected in 250-mL glass serum bottles, preserved with0.5 mL of saturated mercuric chloride solution, and sealedwith grey butyl rubber septa secured with aluminum crimpseals. Shipboard methane analyses were conducted less than24 hr after sample collection.[12] Methane analyses used during the RAPANUI cruise

used a modification of the method of Brooks et al. [1981].The cold trap, which consisted of a 145 cm long column ofPorapak QS, was submerged in ethanol that was cooled toaround �60�C by a portable cooler. Optimized parametersfor stripping the gases from a seawater sample involve 8 minof stripping at a temperature of �195�C and at a flow rate ofover 100 mL/min. However, with a less efficient cold trap at�60�C, stripping lasted for 5 min at a flow rate of 35 mL/min, otherwise the sample would pass through the cold trap.Shore-based laboratory experiments comparing the fieldmethod to the Brooks et al. [1981] method determined thefield method to have an extraction efficiency of 67%, and

Figure 1. Location of study area with an inset larger-scale map showing spreading rate (courtesy ofF. Martinez). The six nontransform offset segments studied during this project are shown, along with thelocations of vertical casts and tow-yos.


3 of 16

B10101

shipboard results were appropriately adjusted. Error attrib-utable to the shipboard methane analysis was calculated at2% from the variations in the values given for standards,but a combined analytical and sampling error of about 8%was calculated from replicate samples [Gharib, 2000].

Shore-based isotopic analyses were performed at theUniversity of Hawaii using an isotope-ratio-monitoringgas chromatography/mass spectrometry technique with aprecision of less than 0.8% for replicate samples [Sansoneet al., 1997].

Figure 2. Methane concentration relative to depth and location over the West and East ridges, showingthe location of the methane plume over the axis. White circles are the sample locations. These figureswere generated by averaging values in a 20 m grid; thus the scale is indicative of the average methaneconcentration within each grid and not that of the individual sample.


4 of 16

B10101

[13] Manganese concentrations were measured usingshipboard flow injection analysis [Resing and Mottl,1992]. DQ and light backscatter data were acquired duringshipboard CTD casts and tows [Baker et al., 2002]. 3Hesamples were collected in copper tubing, and later extractedand analyzed by mass spectrometry [Young and Lupton,1983].

4. Results

[14] A total of over 500 samples were analyzed formethane from the tow-yos and vertical casts and wereplotted as a function of latitude (Figures 2 and 3). Acontinuous signal for a hydrothermal methane plumeextended over the entire length of both ridges, except forthe southernmost samples over the West Ridge, and thenorthernmost samples over the East Ridge, where themethane concentrations were the lowest. These arethe regions that are closest to the propagating tips of eachridge, and therefore the least inflated. The vertical andhorizontal extents of the plumes, including the depth ofthe plume maximum or presence of multiple maxima, weredefined well by the methane concentration profiles.[15] There was evidence for multiple methane maxima in

casts V3, V9, V27, V31 (Figure 3). Two mechanisms have

been proposed for the formation of a plume with more thanone maximum. The first is the overlap of laterally spreadinghydrothermal plumes from different sources; these plumeshave either slightly different densities or slightly differentchemical signatures, or some combination thereof [Baker etal., 1995; McDuff, 1995]. This mechanism is exacerbated inregions of high topography; for example plume dispersionis strongly controlled by some of the high-relief axialvalleys of the Rainbow vent site on the Mid-Atlantic Ridge[German et al., 1998]. The second mechanism is whatSpeer and Rona [1989] refer to as ‘‘turbulent fields withinplumes’’, which contain background seawater that has eitherbeen trapped in the rising plume or influenced by the actionof currents and has not yet fully mixed with the hydrother-mal fluid [Speer and Rona, 1989; McDougall, 1990;McDuff, 1995].[16] The presence of bottom plumes of methane were

observed in several casts, for example, V2, V3, V12, V13(Figure 3). A bottom plume forms as the methane concen-tration increases toward the seafloor, rather than forming amaximum in the overlying water column. Because thevertical casts typically extended to within 10 to 20 m ofthe seafloor, this type of methane maximum was observeddirectly above the seafloor and probably resulted fromdiffuse venting of low-temperature methane-rich fluids

Figure 3. Depth profiles from vertical cast data. See Figure 1 for cast locations. All axes are the sameexcept for cast V33 (�80 km west of the ridge axis) and cast V34 (�480 km west of the ridge axis). Notethe consistency of the plume depth both along axis and off axis, as well as several examples of multipleplume maxima and of bottom plumes. Horizontal line at base of cast indicates seafloor; asterisk indicatesa cast taken off the ridge axis.


5 of 16

B10101

[e.g., McDougall, 1990; Lilley et al., 1991; Rona andTrivett, 1992]. The presumed lower temperatures and there-fore lower buoyancies of this style of venting mightpreclude upward advection through the water column tothe primary plume.[17] An off-axis cast, V33, was taken approximately

80 km to the west of the East Ridge and 100 km south ofthe West Ridge. A hydrothermal plume methane signal at asimilar depth to the signals over the ridges was stillmeasurable even at this distance, indicating westwardand/or southward current flow. There also was a detectablehydrothermal methane signal in the background cast, V34,which was approximately 480 km to the west of the ridge(Figure 3).[18] Segment-averaged methane concentrations were cal-

culated using values from both vertical cast maxima andfrom tow-yo samples, on the assumption that the latter werepredominantly collected within the neutrally buoyantplume. Plume maxima averaged 5.4 nM on segment W3and 7.2 nM on segment W4. The four eastern segments, E1,E2, E3, and E4, averaged 3.2, 1.7, 3.9, and 3.7 nM,respectively (Figure 4). These values represent averages ofthe maximum (or near maximum) measured concentrationsin the plume for each particular segment. Not all of the tow-yos encountered a vigorous plume, though, so theseaverages are approximate. Methane concentrations wereconsiderably different between the two propagating spread-ing centers. The average maximum concentration in the

plumes along the west propagator was 6.0 nM, almost twicethe average in the east, 3.4 nM.[19] The measured methane stable carbon isotopic ratios

(d13C-CH4) generally ranged from �27% to �35%, withthe heaviest value being �24.5%. These data are shown inTable 1, in addition to corresponding depth, methaneconcentration, temperature anomaly, 3He concentration,and backscatter. The isotopic signature on the West Ridgewas slightly lighter than that of the East Ridge.

5. Discussion

5.1. Methane Concentrations

[20] The average maximum methane concentrations fromthe two segments on the West Ridge of this study area wereboth much higher than the average maximum methaneconcentrations from the segments on the East Ridge (seg-ments W3 and W4 were 5.4 nM and 7.2 nM; segments E1–E4 were 3.2, 1.7, 3.9, and 3.7 nM, respectively) (Figure 4).However, because off-axis sampling indicated that themiddepth currents are flowing westward, there may besome contamination of the plume over the northern portionof segment W4, the location of casts V8 and V9, bywestwardly advecting plume water from segment W3.[21] A measurable hydrothermal methane signal in

the water column as far as 80 km from the ridge axis (castV33), and a detectable signal in the background castapproximately 480 km from the ridge axis (cast V34) was

Figure 4. Comparisons of mean CH4 maximum concentrations along the EPR, with data from thisstudy shaded. The mean concentrations are 6.0 nM from the West Ridge (W3–W4) and 3.4 nM from theEast Ridge (E1–E4). Data not from this study are from Mottl et al. [1995, and references therein] andIshibashi et al. [1997, and references therein]. In each case, where the CH4 concentration is greater than20 nM, phase separation has been invoked as its cause.


6 of 16

B10101

observed (Figures 1 and 3). Methane, with a residence timeof approximately two weeks [DeAngelis et al., 1993;Kadko et al., 1990], is considered to be one of theshorter-lived hydrothermal tracers. Advection of hydrother-mal methane to such great distances from a source has notbeen previously reported [Winn et al., 1986; Rosenburg etal., 1988; Winn et al., 1995]. The 28–32�S data suggestthat hydrothermal methane might persist to greater dis-tances than previously thought, especially when the con-centration drops below some threshold level making it aninefficient energy source for a microbial community, andthe role hydrothermal methane plays in the oceanic meth-ane reservoir might be more influential than previouslyconsidered.[22] Hey et al. [2004] and Baker et al. [2002] noted that

the intensity of hydrothermal venting correlated positivelywith high axial inflation and, in contrast, Hey et al. [2004]determined that major plume concentrations on these seg-ments were usually found where fracture and fissure densitywere relatively low. This implies that the hydrothermalactivity is more strongly controlled by the heat supply thanby crustal permeability. The methane data presented hereagree with this trend (Figure 2) in that the highest methaneconcentrations were found over the more inflated portionsof the ridge.[23] The segment-averaged maximum plume methane

concentrations from 28�S to 32�S are compared inFigure 4 with each other and with published concentrationsof methane from other chronic hydrothermal plumes located

along the East Pacific Rise. The values are generally withina factor of two of each other, except for the 9–10�N and the16–19�S systems, which are discussed below. This lack ofvariation between plume methane concentrations andspreading rates suggests that the regional spreading rate(i.e., fast versus superfast spreading) has no significantsystematic effect on the methane concentrations in plumesover the EPR.[24] Our measured concentrations are considerably lower

than those observed by Mottl et al. [1995] at 9–10�N on theEPR, where methane levels reached 90 nM. They attributethe elevated methane to recent volcanic activity and phaseseparation of the hydrothermal fluid [Mottl et al., 1995].Elevated methane concentrations attributed to recent volca-nism were also observed at 16–19�S EPR, although theseconcentrations were lower than at 9–10�N [Charlou et al.,1996; Ishibashi et al., 1997]. This implies that the 28–32�Ssystem was predominantly undergoing steady state hydro-thermal activity during this study, without the elevatedmethane concentrations engendered by the recent volcanicperturbations that affected these other sites.[25] Although higher methane concentrations than those

found on the EPR have been observed at several sites(Table 2), these locations typically have different meansof methane production. At the Endeavour Segment alongthe Juan de Fuca Ridge, and in the Manus and Fiji Basins,microbial methanogenesis of buried sedimentary materialproduces methane-rich plumes; at sites along the Mid-Atlantic Ridge, serpentinization reactions produce elevated

Table 1. Methane Stable Isotope Measurements With Corresponding Depth, Methane Concentration,

Temperature Anomaly, 3He Concentration, and Backscattera

Vertical (V)or Tow-Yo(T) Cast

RidgeSegment

Depth,m

d13C, %Relativeto PDB

CH4,nM

DQ,�C

3He,nM

Backscatter,V

V2 W3 2098 �28.98 4.99 0.009 56.64 0.026V2 W3 2001 �31.47 5.00 0.006 52.69 0.023V2 W3 2030 �29.35 4.54 0.041 141.05 0.167V5 W3 2039 �34.75 2.20 0.027 114.99 0.138V5 W3 2040 �32.66 2.23 0.029 116.20 0.140V8 W4 2220 �30.62 16.81 0.024 86.96 0.037V8 W4 2160 �32.29 10.73 0.013 69.84 0.036V9 W4 2140 �28.00 10.08 0.017 70.87 0.032V9 W4 2090 �32.81 8.46 0.013 67.08 0.042T8 E2 2271 �33.25 0.69 0.008 47.87 0.022T8 E2 2239 �30.00 1.08 0.000 68.76 0.024T9 E3 2245 �24.53 23.77 0.013 74.14 0.057T9 E3 2212 �31.84 3.00 0.004 55.16 0.045T9 E3 2308 �28.50 1.61 0.008 63.75 0.022T9 E3 2308 �28.39 2.72 0.002 39.22 0.023T9 E3 2258 �30.93 3.12 0.014 63.06 0.025T9 E3 2254 �34.65 4.25 0.016 69.25 0.043V23 E3 2220 �28.42 4.01 0.030 93.80 0.037V25 E3 2289 �34.09 4.80 0.032 79.62 0.050V26 E4 2229 �27.43 7.66 0.033 99.05 0.045V26 E4 2192 �27.64 7.33 0.033 108.58 0.037V31 E4 2297 �29.74 5.90 0.022 84.12 0.064V31 E4 2273 �30.44 4.90 0.016 72.83 0.056V31 E4 2221 �31.95 4.81 0.014 68.88 0.067V33 off-axis 2350 �24.80 1.02 0 48.02 0.024V33 off-axis 2300 �24.62 0.98 0 50.39 0.024V33 off-axis 2249 �24.73 0.92 0 53.06 0.025V34 off-axis 2500 �41.59 0.10 0 27.40 0.017V34 off-axis 2000 �38.28 0.19 0 30.07 0.017

aRefer to Figure 1 for locations.


7 of 16

B10101

methane [Mottl et al., 1995; Ishibashi et al., 1997, andreferences therein].

5.2. Covariation of Methane With OtherHydrothermal Tracers

[26] Source-dependent differences among the plumes aremore easily identified by comparing methane to otherhydrothermal tracers. In the following section we use aslope of a model I (least squares) linear regression betweenmethane and the other tracers to calculate the ratio betweenthe two tracers. This facilitates comparisons with publisheddata using the same procedure, rather than the model IIlinear regression, which is more appropriate for two inde-pendent variables [Laws, 1997].5.2.1. CH4/Mn[27] The data plotted in Figure 5 are all of the RAPANUI

plume methane concentrations from on-axis casts versus thecorresponding manganese concentrations. All samples wereused, instead of only the cast maxima, on the assumptionthat the methane and manganese encountered equal dilutionwith ambient seawater and that this should not affect their

concentration ratio. There are not many axial grabens oraxial summit collapse structures [Hey et al., 2004] thatmight impair off-axis dispersion of the hydrothermal plume,and the extent to which the hydrothermal signal wasdetectable, to over 480 km, implies strong off-axis advec-tion. These observations support the assumption that thesesamples from on-axis casts are close to near-zero age. Therelative consistency of the d13C values further supports thisassumption. Also, the depth of the methane maximum in anindividual cast tended to be coincident with the depth of themanganese maximum, except for some of the cases wherethe methane maximum was determined to be a bottomplume (data not shown). The removal or inclusion of thecasts with sizable bottom plumes was found to not signif-icantly affect the CH4/Mn ratios for each segment.[28] The slopes for segments W4 and E1–E4 were all

statistically similar, ranging from 0.077 to 0.091. These arecomparable to ratios from elsewhere on the EPR, whichrange from 0.05 to 0.2, and are close to the 0.1 to 1 rangethat Mottl et al. [1995] call ‘‘typical’’ for CH4/Mn ratios.One segment, W3, was different from the other five seg-ments and did not have increasing methane with increasingmanganese. Possible reasons for the lack of correlationbetween the methane and manganese on segment W3 willbe discussed later. Figure 6 is a comparison between the28–32�S data and other published EPR CH4/Mn ratios. Asstated previously, 9�N and 16–18�S experienced recentvolcanic activity before sampling and the CH4/Mn ratiosreached as high 10.5 and 3.9, respectively. The highest CH4/Mn ratios are predominantly functions of microbial meth-anogenesis, the presence of ultramafic host rock, and phaseseparations.[29] Colonies of bacteria inhabiting hydrothermal vents

and the shallow subsurface of hydrothermal systems havebeen inferred from high concentrations of ATP entrained inventing water [Winn et al., 1986]. Hydrothermal methaneresulting from bacterial activity is especially common inregions of high organic carbon, such as heavily sedimentedhydrothermal systems. Methanogenesis has been proposedto occur on the Endeavour Segment of the Juan de FucaRidge [Lilley et al., 1993] but may be fueled by possiblenearby organic material rather than inorganic carbon.[30] Hydrothermal systems that are influenced by the

presence of ultramafic mantle rocks produce fluids withlower Mn concentrations than fluids produced by basalt-seawater interactions. CH4/Mn ratios in regions such as theMid-Atlantic Ridge can have a large range of values. Thisrange reflects the relative inputs of ‘‘typical’’ CH4/Mn ratiosgenerated by a basalt-seawater hydrothermal component,such as from the TAG and MAR at Kane (MARK)areas, versus high CH4/Mn ratios generated by anultramafic seawater hydrothermal component, such as at15�N [Charlou et al., 1991; Charlou and Donval, 1993].Hydrothermal systems with ultramafic influences are typi-cally found in slow spreading ridges where a significantcomponent of mantle rock is exposed to hydrothermalseawater along the spreading ridge.[31] Faster spreading ridge systems, such as those found

in the EPR, tend to be both relatively unsedimented andalso contain fewer exposures of ultramafic material. Con-sequently, high CH4/Mn ratios are less likely to be influ-enced by these two factors. Instead, the greatest methane

Table 2. Mean CH4 Maximum Concentrations in Hydrothermal

Plumesa

Location CH4, nM CH4/Mn

21�N 2.7 0.04512�090–12�550N 8.1 0.1910�300–11�520N 5.9 0.0759�530–10�070N 35 0.859�390–9�530N 89 10.59�080–9�390N 33 0.518�420–9�080N 5.5 0.06913�320–14�290S 2.8 0.114�360–15�160S 2.7 0.115�430–16�260S 8.8 0.7316�350–17�20s 11.9 1.617�280–17�340S 22.8 1.317�450–18�140S 13.9 0.118�260–18�390S 39.3 3.919�250S 11 0.13Segment W3 (this study) 5.4 0.012Segment W4 (this study) 7.2 0.077Segment E1 (this study) 3.2 0.077Segment E2 (this study) 1.7 0.085Segment E3 (this study) 3.9 0.078Segment E4 (this study) 3.7 0.091Endeavour 48�N 390 6–16MAR around FAMOUS site 5 1.2MAR around AMAR site 4.7 0.8MAR around RAINBOW site 5.1 0.7MAR 15.05�N 18 8.7Teahitia Seamount 14 0.013–10.6Macdonald Seamount 351 1.4–2.8Juan de Fuca South Cleft 45�N 3 0.11Gorda Ridge 42.45�N 5.1 0.53Reykjanes Ridge 63.06�N 18 0.30TAG 26.08� buoyant plume 67.3 .079–0.12TAG 26.08� nonbuoyant plume 9.4 0.26–0.28MARK 23.22� buoyant plume 327 0.074–0.13East Manus Basin shallow plume 6.5 0.29North Fiji Basin shallow plume 4.2 0.24–0.39Loihi Seamount 292 0.26–0.33East Manus Basin deep plume 4.3 0.02–0.05North Fiji Basin deep plume 1.6 0.056

aFrom Mottl et al. [1995, and references therein], Ishibashi et al. [1997,and references therein], and Bougault et al. [1998, and references therein].First part of table presents data from the EPR including this study, followedby data from other locations.


8 of 16

B10101

concentrations and CH4/Mn ratios measured along the EPRwere generated through phase separations (e.g., 9�N and16–18�S EPR) [Mottl et al., 1995; Ishibashi et al., 1997].5.2.2. CH4//#0

[32] DQ is the hydrothermal temperature anomaly in thewater column [Baker et al., 1995]. Both DQ (data notshown) and methane had coincident maxima relative todepth [Baker et al., 2002]. Segments W3 and E1 were the

only segments where methane and DQ did not correlate.Segment E1 had a cluster of values with no linear trend;segment W3 had increasing DQ values with no increase inmethane (Figure 7).[33] The volatile-to-temperature-anomaly ratio in a

hydrothermal plume is another indication of magmaticperturbation in the ridge segment producing the plume,along with the volatile-to-metal ratio [Mottl et al., 1995].

Figure 5. Methane concentration versus manganese concentration. The slope of the least squares linearregression is used for the CH4/Mn ratio.


9 of 16

B10101

On the East Ridge, the CH4/DQ ratio increased withincreasing distance from the propagating tip, suggestingthat the methane source was more robust toward the moreinflated and magmatically active interiors of the segments.The lack of covariation between CH4 and DQ at segmentW3 resembles the CH4 and Mn behavior at segment W3,again indicating that the methane plume chemistry at thissegment is behaving differently than at the other segments.5.2.3. CH4//Backscatter[34] A nephelometer measures the backscatter of light in

the water column, and is therefore a measure of thesuspended particle mass concentration in the water. Whenbuoyant plume samples are disregarded, methane covariedwell with suspended particle density at our study site on allsegments except segment W3 (Figure 8). W3 had a light-scattering intensity that was more than double most of theother segments, suggesting vigorous venting of particulatematerial on that segment, but without any systematicincrease in methane. This behavior is similar to the along-axis trend observed at 17.5–18.6�S EPR where recentmagmatic activity was assumed to have occurred [Ishibashiet al., 1997; Baker et al., 2002].5.2.4. CH4//He[35] Helium was measured on 28 of the collected plume

samples. Although the data are sparse, the values formethane and helium show three populations (Figure 9).The values from segment W4 and the East segments bothcovary well, although segment W4 is much steeper becauseof higher methane concentrations. The third population isfrom segment W3, and those data are scattered, showing norelation. Unfortunately the segment W3 samples with higher

methane concentration do not have corresponding heliummeasurements for comparison.[36] Kelley et al. [1998] observed a considerable enrich-

ment of He over an eruption site on the Gorda Ridge thatwas not reflected in the methane at that depth. This wasattributed to continued degassing of helium from an aginglava flow, which as it cooled was decreasing its methaneproduction from high-temperature water-rock interactions.These Gorda Ridge data demonstrate the sensitivity of theCH4/

3He ratio to recent magmatic perturbations. Despite thelimited number of CH4/

3He ratios from segment W3,Figure 9 clearly shows segment W3 to have the greatestincrease in 3He relative to CH4 of any segment in this studysite, suggesting recent magmatic activity at W3.

5.3. Methane Stable Isotopes

[37] Conservative methane mixing curves between ahydrothermal end-member and the background seawaterare shown in Figure 10 using the methods described bySpiker [1980] and Sansone et al. [1999]. Except for a singleoutlier, the range in the East Ridge data is from �34.6 to�27.4%, and the range in the West Ridge data is from�34.7 to �29.0%. In theory, field data deviating from theselines represent deviations from conservative behaviorcaused by either production or removal of methane. Unfor-tunately, no true end-member was available because thehydrothermal vents were not sampled, so plume sampleswith the highest concentrations of the mixing parameter(3He concentration or DQ) were used as surrogate end-members (Table 1). The interpretations of these mixingcurves are that both ridges have observed isotopic values

Figure 6. A comparison of the log of CH4/Mn ratios along the EPR (data not from this study are fromMottl et al. [1995, and references therein] and Ishibashi et al. [1997, and references therein]).


10 of 16

B10101

that are scattered and below the calculated mixing curves,indicating lighter isotopic signatures than predicted.[38] Observed data deviating from the calculated conser-

vative behavior indicate either production or removal ofmethane. Nonconservative behavior in the plumes is likelyremoval of methane by methanotrophic oxidizers in theplume with time, although some methanogens might bepresent as well, as it has been inferred that microorganismsinhabiting the shallow subsurface of a hydrothermal systemmay get entrained in venting water [Winn et al., 1986;

DeAngelis et al., 1993] and microbial methanogenesistypically produces lighter methane than other sources[Welhan, 1988]. However, it seems unlikely that a microbialcommunity of methanogens would be able to proliferate inthe predominantly oxidizing conditions of a plume.[39] In addition to deviations from the calculated conser-

vative behavior indicating production or removal of meth-ane, there is also the possibility that this scatter results froma range in the source material. Differences in the isotopicsignatures of these sources become integrated into the

Figure 7. Methane concentration versus DQ in Celsius.


11 of 16

B10101

hydrothermal plume. These data are from different casts,some with hundreds of kilometers of separation betweenthem. These isotopic differences can be caused by varyingdegrees of methanogenesis and methanotrophy in thedeep biosphere of the ridge [e.g., Shanks et al., 1995],and/or different lithologies influencing and catalyzing theabiogenic production of methane [e.g., Horita and Berndt,1999], and/or other processes, such as phase separation, thatsegregate hydrothermal fluids into different point sourceswith distinct characters [e.g., Bowers et al., 1988]. Theseprocesses lead to inhomogeneities within the plume; the

conservative mixing curve calculation is based upon asimple two end-member model and therefore does not takethese inhomogeneities into account.[40] As mentioned above, microbial methane oxidation

preferentially removes isotopically light methane, leavingheavier methane with more positive d13C-CH4 ratios. Thisfractionation was noticeable for cast V33, labeled as ‘‘offaxis’’ in Figure 10. These samples were collected approx-imately 80 km to the west of the ridge axis, but still withinthe westwardly advecting hydrothermal plume. Apparentbiological kinetic isotopic fractionation factors [Coleman et

Figure 8. Methane concentration versus light backscatter in volts.


12 of 16

B10101

al., 1981] of 1.0064, 1.0052, and 1.0040 were calculated.These values are similar to the 1.0042–1.025 range formicrobial oxidation obtained empirically from previouslystudies [Coleman et al., 1981; Sansone et al., 1999; Cowenet al., 2002, and references therein], suggesting that theobserved loss of methane in the distal plume is a result ofmicrobial oxidation. Winn et al. [1995] and DeAngelis et al.[1993] also identified microbial activity in distal plumes,with maximum biomass and oxidation rates at some dis-tance from the ridge axis. The other plume samples in ourstudy, from over the ridge axes, were near zero age andwere mainly affected by dilution instead of oxidation, andtherefore plotted close to the conservative mixing lines(Figure 10).

5.4. Effects of Recent Magmatic Activity onSegment W3

[41] Segment W3 is the most inflated segment of thestudy area [Martinez et al., 1997] and is bathymetricallythe shallowest. This inflation of segment W3 relative to theother segments is a likely indicator of recent magmamovement or dike injection [Macdonald and Fox, 1988].More evidence from this region for recent magmatic activityare the interpretations of the backscatter images collected bythe Woods Hole Oceanographic Institution DSL-120 kHzhigh-resolution side-scan sonar during the RAPANUI cruiseand the observed nephelometry measurements. Regionsinterpreted as very low viscosity sheet lava flows arepresent at the tips of fissures along the ridge on segmentW3 and in the northern portion of segment W4 but not thesouthern portion of segment W4 (the northern portion ofsegment W4 is similar in latitude to the southern portion of

segment W3) [Kleinrock et al., 1998]. These backscatterimages are interpreted as very smooth, unsedimented sea-floor surfaces, implying a very young age for these sheetflows. Furthermore, the average nephelometry signal mea-sured along segment W3 was more than double that of mostof the other segments. Baker et al. [2001] observed thatabnormally high light scattering can occur in regions ofrecent volcanic activity as a result of high levels ofparticulate sulfur and microbial material discharged intothe water column.[42] The increased heat flux from a volcanic event or dike

injection segregates hydrothermal fluid into more buoyantchloride-poor, volatile-rich fluid and less buoyant chloride-rich, metal-rich brine [Butterfield et al., 1997]. The morebuoyant volatile-rich fluid is vented first, while the brine isinitially retained near the heat source. The volatile-rich fluidvents quickly, and the transition to the brine phase isbelieved to be on the order of months to years afterthe initial event [Mottl et al., 1995; Urabe et al., 1995;Butterfield et al., 1997].[43] For example, the high methane concentrations and

methane-to-manganese ratios observed at 9�N EPR werelikely the result of the venting of a volatile-rich phasefollowing the volcanic event of seven months earlier [Lilleyet al., 1991, 1992; Mottl et al., 1995; M. Berndt, personalcommunication, 1998]. The same evolution in the hydro-thermal chemistry was also observed at Loihi Seamountfollowing the 1996 seismic event, where initial venting ofchloride-poor and volatile-rich fluid was followed by anincrease in chlorinity and metals, and a decrease in volatiles[Sansone et al., 1998]. The variability in the CH4/Mn molarratios along 16–18�S EPR was also attributed to recent

Figure 9. Methane concentration versus helium concentration.


13 of 16

B10101

magmatic heat injections [Charlou et al., 1996; Ishibashi etal., 1997].[44] Baker et al. [2002] identified three sites of likely

recent magmatic activity in the 28–32�S EPR system. Twoare sites of small but intense plumes on the East Ridge; thethird is segment W3. In addition to the anomalous behaviorof methane, Baker et al. [2002] identify unusually highparticulate sulfur to particulate metal ratios and DpH valuesat segment W3 as evidence for recent magmatism.[45] The plume signature of segment W3 was interesting

in that it had some of the highest average maximumconcentrations of methane, manganese, and light backscat-tering. The methane did not covary with any of these,whether bottom plumes were present or not, and the isotopicdata showed that this lack of covariance was not a result ofmethane oxidation. The poor relation between the methaneand helium was also notable; on the Gorda Ridge this wasinterpreted to result from the cooling of a recent lava flow[Kelley et al., 1998]. This, coupled with the other aforemen-tioned evidence for recent magmatic activity at this segment,

suggests that the anomalous behavior of methane can beexplained as the recovery from recent magmatic activity.[46] Hydrothermal plumes integrate the fluids from many

vents, and these vents may produce fluids with a widerange of compositions. These compositional variations canresult from different degrees of perturbation as well as fromdifferences in rates of recovery from events (e.g., phaseseparation events), generating a range of metal-rich fluids,volatile-rich fluids, and fluids with varying amounts ofsuspended particles. The mixing, in the hydrothermalplumes, of these different fluids would preclude anycovariation between the methane and the various tracers.

6. Conclusions

[47] The East Pacific Rise at 28–32�S is among theEarth’s fastest spreading mid-ocean ridge segments, andthe extensive hydrothermal plume coverage of this ridgesuggests robust hydrothermal venting. However, the seg-ment-averaged maximum methane concentrations and

Figure 10. Calculated mixing diagrams for d13C-CH4 versus He concentration and for d13C-CH4 versusDQ. The solid line indicates the calculated conservative mixing lines for dilution between a hydrothermalplume end-member and a background seawater end-member, that is, the behavior if conservative mixingwere the only removal mechanism of methane from the plume system. Most points are distributed closeto this line. The scatter likely results from errors in the use of surrogate end-member values and/ormethanotrophy/methanogenesis. The triangles are samples of background seawater, the diamonds are theplume samples, and the crosses are the samples from the older distal hydrothermal plume, taken off axis.There is obvious oxidation that has created a heavy signal of these older hydrothermal samples, incomparison with the light signal of the background samples.


14 of 16

B10101

methane/manganese ratios were not indicative of any differ-ences as a result of the superfast spreading conditions andthe lack of transform faults between the segments. Theaverage maximum methane concentrations in the variousplumes were comparable to published values from otherlocations along the East Pacific Rise. The methane concen-trations did vary on a local scale at 28–32�S, and the twosegments on the West Ridge had nearly double the averageplume maximum concentration of the East Ridge. Themethane/manganese ratios showed that segment W4 onthe West Ridge and the segments on the East Ridge wereroughly comparable to those observed at other slowerspreading locations along the EPR. These methane/manga-nese ratios would be classified as ‘‘typical’’ according toMottl et al. [1995]. In contrast, the methane did not covarywith manganese or any other plume tracer on segment W3on the West Ridge.[48] The persistence of the westwardly advecting hydro-

thermal methane signal in the off-axis and background castsillustrates the influence that the advection of hydrothermalmethane can have on methane levels far off axis ordownfield.[49] Microbial oxidation was occurring in the hydrother-

mal plumes, as shown by the isotopic data from the older,off-axis samples (cast V33), but the isotopic mixing curvesshowed oxidation to be an insignificant removal mechanismcompared to the dilution by ambient seawater for thesamples taken from younger plumes over the ridge, suggest-ing that there is little microbial oxidation occurring in theplumbing of these hydrothermal systems prior to venting.[50] Microbial oxidation does not seem to be a good

explanation for the lack of covariation of methane with theother tracers in the plume over segment W3; instead, mixingof different methane-bearing fluids, possibly the result ofphase separation, is proposed. The inflation of segment W3,the images interpreted as recent lava flows, the high lightbackscattering, and an overall high methane concentrationbut poor covariance of the methane with other tracers allsuggest recent magma injections along this segment. Thusthe mixture of fluids from individual vents that are recov-ering from such a perturbation would preclude any ‘‘typi-cal’’ or ‘‘steady state’’ methane behavior. Apart fromconditions on W3, the concentration ratio of methane toother plume tracers along these superfast spreading seg-ments is comparable to that found at other steady statehydrothermal systems along the EPR.

[51] Acknowledgments. This work was funded by NSF grants OCE-990689E and OCE-9911649, and by the NOAA VENTS Program. Wethank Jean-Luc Charlou, Jim Cowen, and an anonymous Journal ofGeophysical Research Associate Editor for their thoughtful reviews of thismanuscript. We also thank Fernando Martinez for contributing unpublishedmagmatic inflation data, the PMEL Vents Group, and Captain Curl and theofficers and crew of the R/V Melville. This is SOEST contribution 6612,and PMEL contribution 2601, and JISAO publication 1175.

ReferencesBaker, E. T., C. R. German, and H. Elderfield (1995), Hydrothermal plumesover spreading-center axes: Global distributions and geological infer-ences, in Seafloor Hydrothermal Systems: Physical, Chemical, Biologi-cal, and Geological Interactions, Geophys. Monogr. Ser., vol. 91, pp.47–71, edited by S. Humphris et al., AGU, Washington, D. C.

Baker, E. T., Y. J. Chen, and J. Phipps Morgan (1996), The relationshipbetween near-axis hydrothermal cooling and the spreading rate of mid-ocean ridges, Earth Planet. Sci. Lett., 142, 137–145.

Baker, E. T., D. A. Tennant, R. A. Feely, G. T. Lebon, and S. L. Walker(2001), Field and laboratory studies on the effect of particle size andcomposition on optical backscattering measurements in hydrothermalplumes, Deep Sea Res., 48, 593–604.

Baker, E. T., et al. (2002), Hydrothermal venting along Earth’s fastestspreading center: East Pacific Rise, 27.5�–32.3�, J. Geophys. Res.,107(B7), 2130, doi:10.1029/2001JB000651.

Baker, E. T., H. N. Edmonds, P. J. Michael, W. Bach, H. J. B. Dick, J. E.Snow, S. L. Walker, N. R. Banerjee, and C. H. Langmuir (2004), Hydro-thermal venting in magma deserts: The ultraslow-spreading Gakkel andSouthwest Indian Ridges, Geochem. Geophys. Geosyst., 5, Q08002,doi:10.1029/2004GC000712.

Baross, J. A., C. N. Dahm, A. K. Ward, M. D. Lilley, and J. R. Sedell(1982), Initial microbial response in lakes to the Mt. St. Helens eruption,Nature, 296, 49–52.

Berndt, M. E., D. E. Allen, and W. E. Seyfried (1996), Reduction of CO2

during serpentinization of olivine at 300�C and 500 bar, Geology, 24,351–354.

Bougault, H., M. Aballea, J. Radford-Knoery, J. L. Charlou, P. JeanBaptiste, P. Appriou, H. D. Needham, C. German, and M. Miranda(1998), FAMOUS and AMAR segments on the Mid-Atlantic Ridge:Ubiquitous hydrothermal Mn, CH4,

3He signals along the rift valleywalls and rift offsets, Earth Planet. Sci. Lett., 161, 1–17.

Bowers, T. S., A. C. Campbell, C. I. Measures, A. J. Spivack, M. Khadem,and J. M. Edmond (1988), Chemical controls on the composition of ventfluids at 13–11�N and 21�N, East Pacific Rise, J. Geophys. Res., 93,4522–4536.

Brooks, J., D. Reid, and B. Bernard (1981), Methane in the upper watercolumn of the northwestern Gulf of Mexico, J. Geophys. Res., 86,11,029–11,040.

Butterfield, D. A., et al. (1997), Seafloor eruptions and evolution of hydro-thermal fluid chemistry, Philos. Trans. R. Soc. London, Ser. A, 355, 369–386.

Charlou, J.-L., and J.-P. Donval (1993), Hydrothermal methane ventingbetween 12�N and 26�N along the Mid-Atlantic Ridge, J. Geophys.Res., 98, 9625–9642.

Charlou, J. L., H. Bougault, P. Appriou, P. Jean-Baptiste, J. Etoubleau, andA. Birolleau (1991), Water column anomalies associated with hydrother-mal activity between 11�400 and 13�N on the East Pacific Rise: Discre-pancies between tracers, Deep Sea Res., 38, 569–596.

Charlou, J. L., Y. Fouquet, J. P. Donval, J. M. Auzende, P. Jean-Baptiste,M. Stievenard, and M. Stephane (1996), Mineral and gas chemistry ofhydrothermal fluids on an ultrafast spreading ridge: East Pacific Rise,17� to 19�S (Naudur cruise, 1993) phase separation processes controlledby volcanic and tectonic activity, J. Geophys. Res., 101, 15,899–15,919.

Coleman, D. D., J. B. Risatti, and M. Schoell (1981), Fractionation ofcarbon and hydrogen isotopes by methane-oxidizing bacteria, Geochim.Cosmochim. Acta, 45, 1033–1037.

Cowen, J. P., G. J. Massoth, and R. A. Feely (1990), Scavenging rates ofdissolved manganese in a hydrothermal vent plume, Deep Sea Res., 37,1619–1637.

Cowen, J. P., X. Wen, and B. N. Popp (2002), Methane in aging hydro-thermal plumes, Geochim. Cosmochim. Acta, 66, 3563–3571.

DeAngelis, M. A., M. D. Lilley, E. J. Olson, and J. A. Baross (1993),Methane oxidation in deep-sea hydrothermal plumes of the EndeavourSegment of the Juan de Fuca Ridge, Deep Sea Res., 40, 1169–1186.

Dick, H. J. B., J. Lin, and H. Schouten (2003), An ultraslow-spreading classof ocean ridge, Nature, 426, 405–412.

Gamo, T., et al. (2001), Chemical characteristics of newly discovered blacksmoker fluids and associated hydrothermal plumes at the Rodriguez Tri-ple Junction, central Indian Ridge, Earth Planet. Sci. Lett., 193, 371–379.

German, C. R., K. J. Richards, M. D. Rudnicki, M. M. Lam, J. L.Charlou, and FLAME Scientific Party ‘3’ (1998), Topographic controlof a dispersing hydrothermal plume, Earth Planet. Sci. Lett., 156, 267–273.

Gharib, J. J. (2000), Methane in hydrothermal plumes along the East PacificRise, 28–32�S, M.S. thesis, Univ. of Hawaii, Honolulu.

Hey, R. N., P. Johnson, F. Martinez, J. Korenaga, M. Somers, Q. Huggett,T. P. LeBas, R. I. Rusby, and D. F. Naar (1995), Plate boundaryreorganization at a large-offset, rapidly propagating rift, Nature, 378,167–170.

Hey, R., et al. (2004), Tectonic/volcanic segmentation and controls onhydrothermal venting along Earth’s fastest seafloor spreading system,EPR 27�–32�S, Geochem. Geophys. Geosyst., 5, Q12007, doi:10.1029/2004GC000764.

Horita, J., and M. E. Berndt (1999), Abiogenic methane formation andisotopic fractionation under hydrothermal conditions, Science, 285,1055–1057.


15 of 16

B10101

Ishibashi, J., H. Wakita, K. Okamura, E. Nakayama, R. A. Feely, G. T.Lebon, E. T. Baker, and K. Marumo (1997), Hydrothermal methane andmanganese variation in the plume over the superfast-spreading southernEast Pacific Rise, Geochim. Cosmochim. Acta, 61, 485–500.

Kadko, D. C., N. D. Rosenberg, J. E. Lupton, R. W. Collier, and M. D.Lilley (1990), Chemical reaction rates and entrainment within theEndeavour Ridge hydrothermal plume, Earth Planet. Sci. Lett., 99,315–335.

Kelley, D. S., and G. Fruh-Green (1996), What is the origin of methane insubmarine hydrothermal systems? (abstract), Eos Trans. AGU, 77(46),Fall Meet. Suppl., F725.

Kelley, D. S., M. D. Lilley, J. E. Lupton, and E. J. Olson (1998), EnrichedH2, CH4, and

3He concentrations in hydrothermal plumes associated withthe 1996 Gorda Ridge eruptive event, Deep Sea Res., Part II, 45, 2665–2682.

Kleinrock, M., R. N. Hey, D. F. Naar, D. W. Bohnenstiehl, F. Martinez,D. Pardee, E. T. Baker, and J. E. Lupton (1998), Volcanism on theFastest Spreading Center: EPR 27�–32� (abstract), Eos Trans. AGU,79(45), Fall Meet. Suppl., F831.

Laws, E. (1997), Mathematical Methods for Oceanographers: An Introduc-tion, John Wiley, Hoboken, N. J.

Lilley, M. D., M. A. DeAngelis, and L. I. Gordon (1982), CH4, H2, CO, andN2O in submarine hydrothermal vent waters, Nature, 300, 48–50.

Lilley, M. D., E. J. Olson, E. A. McLaughlin, and K. L. Von Damm (1991),Methane, hydrogen, and carbon dioxide in vent fluids from the 9�Nhydrothermal system (abstract), Eos Trans. AGU, 72(44), Fall Meet.Suppl., 481.

Lilley, M. D., E. J. Olson, J. E. Lupton, and K. L. Von Damm (1992),Volatiles in the 9N hydrothermal system: A comparison of the 1991 and1992 data (abstract), Eos Trans. American Geophysical Union, 73(43),Fall Meet. Suppl., 524.

Lilley, M. D., D. A. Butterfield, E. J. Olson, J. E. Lupton, S. A. Macko, andR. E. McDuff (1993), Anomalous CH4 and NH4

+ concentrations atan unsedimented mid-ocean-ridge hydrothermal system, Nature, 364,45–47.

Lonsdale, P. (1977), Structural geomorphology of a fast-spreading risecrest: The East Pacific Rise near 3�250S, Mar. Geophys. Res., 3, 251–293.

Lupton, J. E. (1998), Hydrothermal helium plumes in the Pacific Ocean,J. Geophys. Res., 103, 15,853–15,868.

Lupton, J. E., J. R. Delaney, H. P. Johnson, and M. M. Tivey (1985),Entrainment and vertical transport of deep-ocean water by buoyant hy-drothermal plumes, Nature, 316, 621–623.

Macdonald, K., and P. J. Fox (1988), The axial summit graben and cross-sectional shape of the East Pacific Rise as indicators of axial magmachambers and recent volcanic eruptions, Earth Planet. Sci. Lett., 88,119–131.

Martinez, F., R. N. Hey, and P. Johnson (1997), The east ridge system28.5–32�S East Pacific Rise: Implications for overlapping spreadingcenter development, Earth Planet. Sci. Lett., 151, 13–31.

McDougall, T. (1990), Bulk properties of ‘‘hot smoker’’ plumes, EarthPlanet. Sci. Lett., 99, 185–194.

McDuff, R. E. (1995), Physical dynamics of deep-sea hydrothermal plumes,in Seafloor Hydrothermal Systems: Physical, Chemical, Biological,and Geological Interactions, Geophys. Monogr. Ser., vol. 91, edited byS. Humphris et al., pp. 47–71, AGU, Washington, D. C.

Mottl, M. J., F. J. Sansone, C. G. Wheat, J. A. Resing, E. T. Baker, and J. E.Lupton (1995), Manganese and methane in hydrothermal plumes alongthe East Pacific Rise, 8�400 to 11�500N, Geochim. Cosmochim. Acta, 59,4147–4165.

Resing, J. A., and M. J. Mottl (1992), Determination of manganese inseawater using flow injection analysis with on-line preconcentrationand spectrophotometric detection, Anal. Chem., 64, 2682–2687.

Riedel, C., M. Schmidt, R. Botz, and F. Theilen (2001), The Grimseyhydrothermal field offshore north Iceland: Crustal structure, faultingand related gas venting, Earth Planet. Sci. Lett., 193, 409–421.

Rona, P. A., and D. Trivett (1992), Discrete and diffuse heat transfer atASHES vent field, Axial Volcano, Juan de Fuca Ridge, Earth Planet. Sci.Lett., 109, 57–71.

Rosenburg, N. D., J. E. Lupton, D. S. Kadko, R. W. Collier, M. D. Lilley,and H. Pak (1988), Estimation of heat and chemical fluxes from a sea-floor hydrothermal vent field using radon measurements, Nature, 334,604–607.

Sansone, F. J., and C. S. Martens (1978), Methane oxidation in CapeLookout Bight, North Carolina, Limnol. Oceanogr., 23, 349–355.

Sansone, F. J., B. N. Popp, and T. M. Rust (1997), Stable carbon isotopicanalysis of low-level methane in water and gas, Anal. Chem., 69, 40–44.

Sansone, F. J., B. P. Midson, C. G. Wheat, J. A. Resing, C. L. Moyer, andH. W. Jannasch (1998), Geochemical evolution of Loihi Seamount fol-lowing the 1996 seismic event, Eos Trans. American Geophysical Union,79(45), Fall Meet. Suppl., F1022.

Sansone, F. J., M. E. Holmes, and B. N. Popp (1999), Methane stableisotopic ratios and concentration as indicators of methane dynamics inestuaries, Global Biogeochem. Cycles, 13, 463–474.

Shanks, W. C., J. K. Bohlke, and R. R. Seal (1995), Stable isotopes in mid-ocean ridge hydrothermal systems: Interactions between fluids, minerals,and organisms, in Seafloor Hydrothermal Systems: Physical, Chemical,Biological, and Geological Interactions, Geophys. Monogr. Ser., vol. 91,edited by S. Humphris et al., pp. 47–71, AGU, Washington, D. C.

Speer, K. G., and P. A. Rona (1989), A model of an Atlantic and a Pacifichydrothermal plume, J. Geophys. Res., 94, 6213–6220.

Spiker, E. C. (1980), The behavior of 14C and 13C in estuarine water:Effects of in situ CO2 production and atmospheric exchange, Radiocar-bon, 22, 647–654.

Tsunogai, U., N. Yoshida, J. Ishibashi, and T. Gamo (2000), Carbon iso-topic distribution of methane in deep-sea hydrothermal plume, MyojinKnoll Caldera, Izu-Bonin arc: Implications for microbial methane oxida-tion in the oceans and applications to heat flux estimation, Geochim.Cosmochim. Acta, 64, 2439–2452.

Urabe, T., et al. (1995), The effect of magmatic activity on hydrothermalventing along the superfast-spreading East Pacific Rise, Science, 269,1092–1095.

Von Damm, K. I. (1990), Seafloor hydrothermal activity: Black smokerchemistry and chimneys, Annu. Rev. Earth Planet. Sci., 18, 173–204.

Welhan, J. A. (1988), Origins of methane in hydrothermal systems, Chem.Geol., 71, 183–198.

Welhan, J. A., and H. Craig (1983), Methane, hydrogen, and helium inhydrothermal fluids at 21�N on the East Pacific Rise, in HydrothermalProcesses at Seafloor Spreading Centers, edited by P. Rona et al., pp.391–409, Springer, New York.

Winn, C. D., D. M. Karl, and G. J. Massoth (1986), Microorganisms indeep-sea hydrothermal plumes, Nature, 320, 744–746.

Winn, C. D., J. P. Cowen, and D. M. Karl (1995), Microbes in deep-seahydrothermal plumes, in The Microbiology of Deep-Sea HydrothermalVents, edited by D. Karl, pp. 255–273, CRC Press, Boca Raton, Fla.

Yoshida, T., K. Nishizawa, M. Tabata, H. Abe, T. Kodama, M. Tsuji,and Y. Tamaura (1993), Methanation of CO2 with H2-reduced magne-tite, J. Mater. Sci., 28, 1220–1226.

Young, C., and J. E. Lupton (1983), An ultra-tight fluid sampling systemusing cold-welded copper tubing (abstract), Eos Trans. AGU, 64, FallMeet. Suppl., 735.

��E. T. Baker and J. A. Resing, Pacific Marine Environmental Laboratory,

NOAA/R/PMEL, 7600 Sand Point Way NE/Bldg. 3, Seattle, WA 98115-0070, USA. ([email protected]; [email protected])J. J. Gharib, Department of Geology and Geophysics, 1680 East West

Road, University of Hawaii, Honolulu, HI 96822, USA. ([email protected])J. E. Lupton, NOAA/PMEL VENTS Program, Hatfield Marine Science

Center, 2115 S.E. OSU Dr., Newport, OR 97365, USA. ([email protected])G. J. Massoth, Institute of Geological and Nuclear Sciences, 30

Gracefield Road, Lower Hutt, New Zealand. ([email protected])F. J. Sansone, Department of Oceanography, University of Hawaii, 1000

Pope Road, Honolulu, HI 96822, USA. ([email protected])


16 of 16

B10101

methane dynamics in hydrothermal plumes over a superfast ... et al 2005 jgr.pdfb10101 gharib et al.:...

Documents