Search

The hydrocarbon (n-alkane and n-alkene) distribution in deep seawaters of the Mid-Indian Basin (MIB) and hydrothermal vent sites of the Central (CIR) and Southeast Indian Ridges (SEIR) were investigated to understand the impact of hydrothermal activities in deep seawater organic matter. The distinguishing feature of the detected hydrocarbons was in a bimodal distribution with a strong predominance of even-numbered n-alkenes (C16:1-C34:1) maximizing at n-C16 or n-C18, and a series of n-alkane hydrocarbons ranging from C16–C35 with slightly high content in long chain par but without an even/odd carbon number preference. The mono-unsaturated aliphatics were generally 1–4 times higher than the saturated counterpart and exhibited more complex vertical variabilities. The higher content of aliphatics in deeper water, i.e., the signals at 2500 m from the CIR and the increasing tendency towards the bottom from the SEIR in the near vent field sites, indicate a potential hydrothermal contribution. Sulfate-reducing microorganisms of bacteria and archaeon in seawater and the hydrothermal system may be responsible for the distribution feature of the saturated or unsaturated hydrocarbons, but the possibility of abiogenic sources cannot be excluded. In addition, the hydrocarbon profiles recorded the organic debris mineralization in the subsurface coinciding with dramatically hydrographic variability. The present aliphatic profiles feature reflected the complex bio-geochemical process, a minor impact of hydrothermal activity on the hydrocarbon distributions and deep water circulations in the Indian Ocean.

Sea-floor imagery, volcanic rock, massive sulfide, and hydrothermal plume samples (δ 3 He, pH, dissolved Fe and Mn, and particulate chemistry) have been collected from the Rumble II West volcano, southern Kermadec arc, New Zealand. Rumble II West is a caldera volcano with an ∼3-km-diameter summit depression bounded by ring faults with a resurgent central cone. Rocks recovered to date are predominantly mafic in composition (i.e., basalt to basaltic andesite) with volumetrically lesser intermediate rocks (i.e., andesite). On the basis of its size, geometry, volcanic products, and composition, Rumble II West can be classified as a mafic caldera volcano. Rumble II West has a weak hydrothermal plume signature characterized by a small but detectable δ 3 He anomaly (25%). Time-series light scattering data though, obtained from vertical casts and tow-yos, do show that hydrothermal activity has increased in intensity between 1999 and 2011. Massive sulfides recovered from the eastern caldera wall and eastern flank of the central cone are primarily comprised of barite and chalcopyrite, with lesser sphalerite, pyrite, and traces of galena. The weak hydrothermal plume signal indicates that the volcano is in a volcanic-hydrothermal quiescent stage compared to other volcanoes along the southern Kermadec arc, although the preponderance of barite with massive sulfide mineralization indicates higher temperature venting in the past. Of the volcanoes along the Kermadec-Tonga arc known to host massive sulfides (i.e., Clark, Rumble II West, Brothers, Monowai, Volcano 19, and Volcano 1), the majority (five out of six) are dominantly mafic in composition and all but one of these mafic volcanoes form moderate-size to large calderas. To date, mafic calderas have been largely ignored as hosts to sea-floor massive sulfide deposits. That 75% of the presently known massive sulfide-bearing calderas along the arc are mafic in composition (the dacitic Brothers volcano is the exception) has important implications for sea-floor massive sulfide mineral exploration in the modern oceans and ancient rock record on land.

The Monowai volcanic center is located at the midpoint along the ~2,530-km-long Tonga-Kermadec arc system. The Monowai volcanic center is comprised of a large elongate caldera (Monowai caldera area ~35 km2; depth to caldera floor 1,590 m), which has formed within an older caldera some 84 km2 in area. To the south of this nested caldera system is a large composite volcano, Monowai cone, which rises to within ~100 m of the sea surface and which has been volcanically active for the past several decades. Mafic volcanic rocks dominate the Monowai volcanic center; basalts are the most common rock type recovered from the cone, whereas basaltic andesites are common within the caldera. Hydrothermal plume mapping has shown at least three major hydrothermal systems associated with the caldera and cone: (1) the summit of the cone, (2) low-temperature venting (

Brothers volcano, of the Kermadec intraoceanic arc, is host to a hydrothermal system unique among seafloor hydrothermal systems known anywhere in the world. It has two distinct vent fields, known as the NW Caldera and Cone sites, whose geology, permeability, vent fluid compositions, mineralogy, and ore-forming conditions are in stark contrast to each other. The NW Caldera site strikes for ∼600 m in a SW–NE direction with chimneys occurring over a ∼145-m depth interval, between ∼1,690 and 1,545 m. At least 100 dead and active sulfide chimney spires occur in this field and are typically 2–3 m in height, with some reaching 6–7 m. Their ages (at time of sampling) fall broadly into three groups

Microbial communities in black smoker chimney structures at the NW caldera vent field of the Brothers volcano, Kermadec arc were characterized by using both culture-dependent and -independent techniques. The hydrothermal vent fluid chemistry, as given by end-member salinities and gas contents, differ among the black smoker sites of the NW caldera field, indicating probable phase-separation-controlled variability in the fluid chemistry. Chimney structures collected from typical Cl-depleted and Cl-enriched hydrothermal fluid vents were used for the microbiological investigation. The 16S rRNA gene clone analysis showed that the archaeal rRNA gene communities were similar within interior and exterior substructures of any single chimneys, and even between chimneys having either Cl-depleted or Cl-enriched hydrothermal emissions. By contrast, the bacterial rRNA gene communities varied between chimneys hosting Cl-depleted or Cl-enriched fluids. Cultivation analysis showed significant variation in the viable count...

A comprehensive survey of hydrothermal plumes and their geochemistry revealed 16 hydrothermally active volcanoes along 1,200 km of the Mariana arc from 13.5° N to 22.5° N. Of these 16, one volcano, NW Rota-1, is discharging sulfuric acid-rich fluids producing some of the largest chemical anomalies ever observed in non-buoyant hydrothermal plumes. Two types of venting were observed, one of focused flow rich in Al, S, Si, CO2, Fe, Mn, and 3He, and a second of diffuse flow rich in Fe, Mn, CO2, and 3He but without Al, S, and Si. The plume waters from the acid-rich focused flow showed decreases in pH up to 0.73 pH units compared to ambient seawater and an estimated 80 percent of this change was brought about by volcanic and/or hydrothermal SO2 (and possibly HCl) with the remaining 20 percent by CO2. Conservative estimates suggest that the pH of the venting fluids was

Dissolved methane and high resolution bathymetry surveys were conducted over the Rock Garden region of Ritchie Ridge, along the Hikurangi margin, eastern New Zealand. Multibeam bathymetry reveals two prominent, northeast trending ridges, parallel to subduction along the margin, that are steep sided and extensively slumped. Elevated concentrations of methane (up to 10 nM, 10× background) within the water column are associated with a slump structure at the southern end of Eastern Rock Garden. The anomalous methane concentrations were detected by a methane sensor (METS) attached to a conductivity‐temperature‐depth‐optical backscatter device (CTDO) and are associated with elevated light scattering and flare‐shaped backscatter signals revealed by the ship's echo sounder. Increased particulate matter in the water column, possibly related to the seepage and/or higher rates of erosion near slump structures, is considered to be the cause of the increased light scattering, rather than bubbles in the water ...

Earth’s fastest present seafloor spreading occurs along the East Pacific Rise near 31°–32° S. Two of the major hydrothermal plume areas discovered during a 1998 multidisciplinary geophysical/hydrothermal investigation of these mid-ocean ridge axes were explored during a 1999 Alvin expedition. Both occur in recently eruptive areas where shallow collapse structures mark the neovolcanic axis. The 31° S vent area occurs in a broad linear zone of collapses and fractures coalescing into an axial summit trough. The 32° S vent area has been volcanically repaved by a more recent eruption, with non-linear collapses that have not yet coalesced. Both sites occur in highly inflated areas, near local inflation peaks, which is the best segment-scale predictor of hydrothermal activity at these superfast spreading rates (150 mm/yr).

Brothers volcano, which is part of the active Kermadec arc, northeast of New Zealand, forms an elongate edifice 13 km long by 8 km across that strikes northwest-southeast. The volcano has a caldera with a basal diameter of ~3 km and a floor at 1,850 m below sea level, surrounded by 290- to 530-m-high walls. A volcanic cone of dacite rises 350 m from the caldera floor and partially coalesces with the southern caldera wall. Three hydrothermal sites have been located: on the northwest caldera wall, on the southeast caldera wall, and on the dacite cone. Multiple hydrothermal plumes rise ~750 m through the water column upward from the caldera floor, originating from the northwest caldera walls and atop the cone, itself host to three separate vent fields (summit, upper flank, northeast flank). In 1999, the cone site had plumes with relatively high concentrations of gas with a ΔpH of −0.27 relative to seawater (proxy for CO2 + S gases), dissolved H2S up to 4,250 nM, high concentrations of particulate Cu (up to 3.4 nM), total dissolvable Fe (up to 4,720 nM), total dissolvable Mn (up to 260 nM) and Fe/Mn values of 4.4 to 18.2. By 2002, plumes from the summit vent field had much lower particulate Cu (0.3 nM), total dissolvable Fe (175 nM), and Fe/Mn values of 0.8 but similar ΔpH (−0.22) and higher H2S (7,000 nM). The 1999 plume results are consistent with a magmatic fluid component with the concentration of Fe suggesting direct exsolution of a liquid brine, whereas the much lower concentrations of metals but higher overall gas contents in the 2002 plumes likely reflect subsea-floor phase separation. Plumes above the northwest caldera site are chemically distinct, and their compositions have not changed over the same 3-year interval. They have less CO2 (ΔpH of −0.09), no detectable H2S, total dissolved Fe of 955 nM, total dissolved Mn of 150 nM, and Fe/Mn of 6.4. An overall increase in 3He/4He values in the plumes from R/RA = 6.1 in 1999 to 7.2 in 2002 is further consistent with a magmatic pulse perturbing the system. The northwest caldera site is host to at least two large areas (~600 m by at least 50 m) of chimneys and sub-cropping massive sulfide. One deposit is partially buried by sediment near the caldera rim at ~1,450 m, whereas the other crops out along narrow, fault-bounded ledges between ~1,600 and 1,650 m. Camera tows imaged active 1- to 2-m-high black smoker chimneys in the deeper zone together with numerous 1- to 5-m-high inactive spires, abundant sulfide talus, partially buried massive sulfides, and hydrothermally altered volcanic rocks. 210Pb/226Ra dating of one chimney gives an age of 27 ± 6 years; 226Ra/Ba dating of other mineralization indicates ages up to 1,200 years. Formation temperatures derived from Δ34Ssulfate-sulfide mineral pairs are 245° to 295° for the northwest caldera site, 225° to 260°C for the southeast caldera and ~260° to 305°C for the cone. Fluid inclusion gas data suggest subsea-floor phase separation occurred at the northwest caldera site. Alteration minerals identified include silicates, silica polymorphs, sulfates, sulfides, Fe and Mn oxide and/or oxyhydroxides, and native sulfur, which are consistent with precipitation at a range of temperatures from fluids of different compositions. An advanced argillic assemblage of illite + amorphous silica + natroalunite + pyrite + native S at the cone site, the occurrence of chalcocite + covellite + bornite + iss + chalcopyrite + pyrite in sulfide samples from the southeast caldera site, and veins of enargite in a rhyodacitic sample from the northwest caldera site are indicative of high-sulfidation conditions similar to those of subaerial magmatic-hydrothermal systems. The northwest caldera vent site is a long-lived hydrothermal system that is today dominated by evolved sea-water but has had episodic injections of magmatic fluid. The southeast caldera site represents the main upflow of a relatively well established magmatic-hydrothermal system on the sea floor where sulfide-rich chimneys are extant. The cone site is a nascent magmatic-hydrothermal system where crack zones localize upwelling acidic waters. Each of these different vent sites represents diverse parts of an evolving hydrothermal system, any one of which may be typical of submarine volcanic arcs.

The vestimentiferan tubeworm Ridgeia piscesae is an ecosystem-structuring organism in the hydrothermal vent environments of the Northeast Pacific. During this study, a single representative aggregation of the long-skinny morphotype of R. piscesae from the main endeavor segment was monitored for 3 yr before being collected in its entirety with a hydraulically actuated collection device manipulated in situ by a research vehicle. Vestimentiferan growth rates in this aggregation were determined by staining the exterior of the tubes and measuring newly deposited tube sections. The average growth rate of R. piscesae in this aggregation was very low in both years of the growth study (3.2 mm yr−1). Although the incidence of plume damage from partial predation was very high (>95%), mortality was very low (

The eruption of Axial Volcano in January 1998 produced extensive plumes in the overlying water column with large anomalies in Fe, Mn, pH, light attenuation, and temperature. A strong correlation between total iron and light attenuation (dc) suggests a low abundance of particulate sulfur (PS) in the plumes. Because total carbon dioxide (ΣCO2) samples were not collected, high-precision pH measurements were used to estimate maximum CO2 anomalies (ΔCO2) which, when compared to the other physical and chemical data, suggest that the fluids being vented 3 weeks after eruption were formed by the interaction between an intruded dike and a mixture of interstitial seawater and mature hydrothermal fluids.

Hydrothermal event plumes are unique water-column features observed over mid-ocean ridges, presumably generated by the sudden release of large volumes of hot, buoyant fluid. Although the specifics of event plume generation are unknown, event plumes have been attributed to the rapid emptying of a hydrothermal reservoir or to rapid heat extraction from a recently emplaced dike or seafloor lava flows. The chemical and thermal signatures of event plumes as compared to the underlying steady-state plumes offer important clues to the generation of event plumes. Event plumes have low 3He/heat ratios of ∼0.4 × 10−17 mol J−1, similar to vent fluids from mature hydrothermal systems. In contrast, the steady-state plumes found beneath the event plumes have elevated and variable 3He/heat ratios of 2 to 5 × 10−17 mol J−1. Fluids collected directly over fresh lava flows have even higher 3He/heat ratios of 2 to 8 × 10−17 mol J−1, up to 30 times the event plume values. These disparate 3He/heat ratios place strong constraints on models of event plume generation, especially models which rely on heat extraction from seafloor eruptions. Published by Elsevier Science B.V.

Total heat, salt, and other tracer masses released during a hydrothermal event are shown to be proportional to, but not necessarily equal to, volume integrals of resulting water column anomalies. Proportionality coefficients depend on anomaly definition, on background hydrographic and tracer profiles, on expansion coefficients of the equation of state at an appropriate pressure, and on tracer to heat anomaly ratios at the venting source. For Gorda Ridge event plumes, which are described in other papers of this issue, volume integrals of conventionally defined heat anomalies underestimate actual released heat by a factor of 2.4 if the discharge is not anomalously saline. Under certain combinations of hydrographic and source anomaly conditions, not unlike those found on the East Pacific Rise at 10°N, the apparent total heat released during an event can be deceivingly zero. This analysis also establishes a linear relationship between the ratio of tracer anomaly to heat anomaly at any point in the plume to the same ratio at the source. One consequence is that the ratio of anomalous 3 He to heat in Gorda Ridge event plumes is approximately 2.4 larger in the water column than it is at the source. Results are independent of the entrainment process involved in event plume formation, and they are shown to hold true even for background hydrographic profiles that do not vary linearly with depth.

We sampled hydrothermal plumes over the N. Gorda Ridge four times between March and August 1996 to document Mn and Fe discharge resulting from a magmatic intrusion/seafloor eruption. Two separate event plumes, EP96A and B, and chronic hydrothermal emissions lasting ∼6 months were characterized. Shipboard time-series measurements of an event plume sample were used to calibrate an Fe phase clock useful for estimating sample age up to ∼6 days after fluid discharge. Samples collected from EP96A and B had Mn/heat ( 2 mol mol-1) ratios similar to historical event plume observations. We suggest these “signature” ratio values are generally characteristic of event plumes and hypothesize that Mn and Fe may be supplied to event plumes by different processes: Mn by entrainment of fluids from an extant shallow subseafloor reservoir, and Fe by short-lived, high-temperature water-rock reaction coincident with dike emplacement. Calculations based on the Fe phase clock indicate that the two event plumes were released more than a month apart. The largest event plume, EP96A (∼2.3×106 M Mn and 13×106 M Fe), formed ∼7 March soon after seismic activity began. The smaller EP96B (∼0.49×106 M Mn and 3.5×106 M Fe) was not discharged until ∼11 April, 3 weeks after the cessation of seismic activity detectable by SOSUS T-phase monitoring. We hypothesize that the subseafloor disturbance that triggered EP96B also resulted in the episodic flushing of a reservoir of chronic-plume-like fluids. Total event plume inventories of Mn and Fe at N. Gorda Ridge are much smaller than those associated with the 1986 event at N. Cleft segment of the Juan de Fuca Ridge, but comparable to event plume inventories at N. Cleft segment in 1987 and CoAxial segment in 1993. Mn/heat values for chronic plumes over the eruption site underlying EP96A evolved from moderate (∼0.25 oonmol J-1, reflecting probable admixture with event plume formation fluids) to high (∼0.7 nmol J-1, typical of chronic plumes) to low (∼0.1 nmol J-1, similar to diffuse vent fluid values), marking a complete episode of intrusion/eruption-induced hydrothermal discharge.

In response to the 1996 magmatic intrusion on the Gorda Ridge near 42.68°N, 126.78°W in late February, we conducted three cruises to the region in March, April, and June. On all three cruises particulate samples were collected, along with CTD/nephelometer data and total dissolved Fe and Mn samples. During each cruise, we collected samples from large oblate spheroid-shaped event plumes. These event plumes had long axis diameters of about 10–15 km and ranged in depth from about 1800 to 2700 m. A strong linear correlation between nephelometer voltage and particulate Fe allowed us to estimate the total amount of particulate Fe in the event plumes at approximately 20×10 6 moles of Fe, or ∼28% of the Fe in the 1986 megaplume on the Cleft Segment of the Juan de Fuca Ridge. We observed significant decreases in particulate Cu and Zn concentrations (>100% decrease in Cu/Fe and Zn/Fe ratios) between the Gorda Ridge event plumes. These results suggest that each of the two event plumes originated from a chemically distinct source fluid. Fe ferrihydrite particles maintained a constant ratio of coprecipitated oxyanion species in the two event plumes. Based upon the chemical inventories for particulate Fe, P, and V, we suggest that event plumes might play a small role in the geochemical budgets for these elements.

Event plumes form as episodic discharges of large volumes of hydrothermal solutions in response to magmatic diking/eruptive events. In consequence, event plumes represent the sudden injection of exploitable reduced chemical substrates, as well as inhibitory constituents, and likely induce successional changes in the microbial community structure and activity within event plume waters. In response to a major seismic event detected beginning 28 February 1996 at the northern Gorda Ridge, a series of three rapid response and follow-up cruises (GREAT 1, 2 and 3) were mounted over a period of three months. This report focuses on time-series measurements of manganese geomicrobial parameters in the two event plumes found in association with this seismic event. Scanning transmission electron microscopy, elemental microanalysis, and radioisotope (54Mn) uptake experiments were employed on samples collected from vertical and tow-yo casts from the three cruises. Numbers of bacteria and ratios of metal precipitating capsuled bacteria to total bacteria were greatest in the youngest (days old) plume, EP96A, found during GREAT 1; however, when normalized to the hydrothermal temperature anomaly, the greatest values were found in a second event plume, EP96B, discovered during GREAT 2 (up to 1 month old). Early capsule bacteria and particulate Mn distributions may have been influenced by entrainment of resuspended sediment, while those of the oldest (2–3 months) plume sample may have been subjected to preferential aggregation and particle settling.

The superfast-spreading portion of the East Pacific Rise at 17.5°S is a magmatically robust section of mid-ocean ridge. Plumes characterized by high volatile/metal ratios, which has been hypothesized to be indicative of recent magmatism, were found between 17°22′S and 17°35′S in November–December 1993. Dives with the French submersible Nautile in December 1993 observed young sheet flows and widespread venting centered on the segment's shoalest point at 17°26′S. Eight submersible dives made one year later with the Japanese submersible Shinkai 6500 in September–November 1994 found glassy, unsedimented lavas, a range of diffuse and high-temperature vents, and several types of biologic communities on the ridge crest beneath the volatile-rich 1993 plumes. The diffuse vents range from small areas in the youngest lavas characterized by very low-temperature flow (

In June 1993, a seafloor dike intrusion along the CoAxial segment of the Juan de Fuca Ridge was acoustically detected. A near-immediate field response and repeated plume mapping and sampling surveys during eight cruises over the next 3 years have provided a unique opportunity to estimate heat and mass fluxes from both event and chronic discharge during the life cycle of a newly created hydrothermal system. The intrusion triggered the release of at least three event plumes followed by chronic discharge focused at two sites: Flow, the site of a lava eruption at the distal end of the intrusion; and Floe, 30 km back along the trend of the dike. We have combined measurements of plume temperature anomalies, plume areal extent, and year-long averages of current flow at both sites to estimate the chronic hydrothermal heat flux Hu. Initial values of Hu at both sites were of order 104 MW, declining over time t as Hu = at−k, with k ≈ 1. Significant plumes were no longer detectable at Flow by June 1995, or at Floe by June 1996. Elemental fluxes from the CoAxial system have been derived from Hu and measurements of the ratios Mn/heat, Fe/heat, and particulate S/heat in chronic plumes. While Mn and Fe fluxes mirrored the power curve decline of heat, the combined regional particulate S (PS) flux experienced a second pronounced maximum some months after the eruption owing to a sharp increase in the S/heat ratio at Floc. Integrated inventories from chronic discharge were ∼4×1017 J for heat, ∼3×108 mol for Mn, ∼2×108 mol for Fe, and ∼1×108 mol for PS. Realistic uncertainties for all species are roughly a factor of 2. The three event plumes accounted for

Dissolved H2S is a major environmental factor in hydrothermal vent ecosystems. In a study of adaptations to sulfide by alvinellid polychaetes, the sulfide-binding capacity of body fluids was examined in Paralvinella palmiformis from northeast Pacific ridges and Alvinella species from the East Pacific Rise. Sulfide concentrations in vascular blood and coelomic fluid of freshly collected animals were notably variable. Separation of P. palmiformis body-fluid components revealed that most sulfide (ca. 77%) was accumulated in the dissolved fraction. In P. palmiformis, both vascular blood and coelomic fluid could reversibly bind sulfide in vitro with a low affinity, saturating only at high dialysate concentrations (ca. 2 mmol L-1). No sulfide-binding activity was observed in the vascular blood from Alvinella species. A dissolved protein component of greater than 90 kDa appears to be involved in sulfide binding in Paralvinella, probably a vascular extracellular high-molecular-weight hemoglobin. Some sulfide may also adsorb onto a 15.38-kDa intracellular hemoglobin present in the coelomic erythrocyte fraction. In the absence of epibiotic bacteria, Paralvinella body fluids may function as a sulfide buffer to protect tissues from deleterious effects of sulfide exposure.

A recent eruption on CoAxial Segment of Juan de Fuca Ridge initiated hydrothermal conditions with rapid changes in water chemistry and growth of microbial communities. Vent animals recruited from distal sources within a year. One site with newly erupted lava attracted no animals to high-iron and low-sulphide conditions. However, sustained release of flocculent material at a second site suggests extensive subterranean microbial production; here, the dissolved sulphide/heat ratio peaked during the first year. The first larval recruits included vestimentiferans, alvinellid polychaetes and nemerteans; despite the small areal extent of venting, one-third of the regional vent species pool had arrived by 2 years. Near-optimal growth conditions and recruitment by many species continued in the centre of the system but several habitats went extinct within 2 years. Rapid response and exploitation by vent animals must be an important adaptation to such ephemeral conditions.

A major challenge confronting geochemists is to relate the chemistry of vented hydrothermal fluids to the local or regional tectonic and volcanic state of midocean ridges. After more than 15 years ...

A survey of hydrothermal activity along the superfast-spreading (approximately 150 millimeters per year) East Pacific Rise shows that hydrothermal plumes overlay approximately 60 percent of the ridge crest between 13°509 and 18°409S, a plume abundance nearly twice that known from any other rige portion of comparable length. Plumes were most abundant where the axial cross section is inflated and an axial magma chamber is present. Plumes with high ratios of volatile ( 3 He, CH 4 , and H 2 S) to nonvolatile (Mn and Fe) species marked where hydrothermal circulation has been perturbed by recent magmatic activity. The high proportion of volatile-rich plumes observed implies that such episodes are more frequent here than on slower spreading ridges.

The episodic magmatic intrusions that accompany seafloor spreading can profoundly affect the discharge of hydrothermal heat and fluid. A rapid field response to the acoustic detection of seismic activity on the CoAxial segment of the Juan de Fuca Ridge, June/July 1993, provided the opportunity to conduct a unique series of observations of hydrothermal event plumes. Between July 3 and 26 we detected at least three event plumes ranging in plume volume from 1.3 to 4 × 1010 m³, and in excess heat from 2.2 to 12.4 × 1015 J. Two were discovered directly above a new lava extrusion. One of these was first sampled within days of its release, as evidenced by its asymmetric shape, complex structure, and a steadily increasing light attenuation anomaly apparently generated by ongoing precipitation of hydrothermal Fe. We hypothesize that event plumes were produced intermittently in response to successive volcano-tectonic events that produced pulses in seismic activity between June 28 and July 14.

Magmatic injection during June/July 1993 on the CoAxial segment of the Juan de Fuca Ridge resulted in an axial lava eruption, an overlying chronic-style hydrothermal plume, and at least three megaplume-like event plumes. A comprehensive view of Mn and Fe liberated as hydrothermal discharge during this volcano-hydrothermal event was obtained as part of a rapid field response that included continuous in situ measurements of Mn and Fe. Distribution maps for Mn, Fe and temperature anomaly within the chronic and event plumes reveal significant thermochemical diversity that we ascribe to variable extraction efficiencies for heat and metals by source fluids, and to plume age differences. The Mn/heat data suggest that diffuse fluids could be the primary source of the chronic plume at northern CoAxial segment but make only a minor contribution to event plumes. Measurements of ephemeral Fe(II) were used to estimate that the event plumes were ∼2, ∼4, and >6–8 days old when first sampled, corresponding to formation times during peak seismic activity. The inventories of Mn and Fe within the three CoAxial event plumes were similar to values previously reported for megaplumes.

The 1993 volcanic/hydrothermal event on the CoAxial segment, northern Juan de Fuca Ridge, generated a complex set of event plumes and steady-state type plumes with a pattern of 3 He/heat ratios very similar to that observed during the 1986 megaplume event on the southern JdFR. The event plumes residing -600 m above the seafloor had normal He/heat ratios averaging 0.34 x 10 -12 cm 3 STP cal -1 , while the deeper, steady-state plumes had elevated 3 He/heat values ranging from 0.6 to 4.5 x 10 -12 cm 3 STP cal -1 . These results generally confirm that elevated 3 He/heat ratios in some of the water-column plumes are a reliable indicator of a recent magmatic/tectonic event. However, in contrast to the 1986 JdFR event, there was no evidence for hydrothermal activity on the CoAxial segment prior to 1993, thereby requiring revision of our models for event plume generation.

[1] The Northwest Lau Backarc Basin, consisting of the Northwest Lau Spreading Center (NWLSC) and the Rochambeau Rifts (RR), is unique in having elevated 3He/4He ratios (up to 28 Ra) in the erupted lavas, clearly indicating a hot spot or ocean island basalt (OIB)-type signature. This OIB-type helium signature does not appear in any other part of the Lau Basin. Water column plume surveys conducted in 2008 and 2010 identified several sites of active hydrothermal discharge along the NWLSC-RR and showed that the incidence of hydrothermal activity is high, consistent with the high spreading rate of ∼100 mm/year. Hydrocasts into the Central Caldera and Southern Caldera of the NWLSC detected elevated3He/4He (δ3He = 55% and 100%, respectively), trace metals (TMn, TFe), and suspended particles, indicating localized hydrothermal venting at these two sites. Hydrocasts along the northern rift zone of the NWLSC also had excess δ3He, TMn, and suspended particles suggesting additional sites of hydrothermal activity. The RR are dominated by Lobster Caldera, a large volcano with four radiating rift zones. Hydrocasts into Lobster Caldera in 2008 detected high δ3He (up to 239%) and suspended particle and TMn signals, indicating active venting within the caldera. A repeat survey of Lobster in 2010 confirmed the site was still active two years later. Plumes at Lobster Caldera and Central Caldera have end-member3He/4He ratios of 19 Ra and 11 Ra, respectively, confirming that hot spot-type helium is also present in the hydrothermal fluids.

Hydrothermal fluid samples collected in 1984, 1987, and 1988 from a large vent field near 47°57′N on the Endeavour segment of the Juan de Fuca Ridge (JFR) have been analyzed for major and minor elements and gases. There are of the order of 100 individual smoker vents on ∼10 large sulfide structures, which are localized along faults and fault intersections across the vent field. Each sulfide structure has a characteristic fluid composition, which varies very little from one vent orifice to the next, or from year to year, on a given structure. However, there are large gradients in fluid composition across the vent field, with endmember chlorinity increasing from ∼255 mmol/kg in the SW to 505 mmol/kg in the NE. End-member concentrations of major elements are well correlated with chlorinity, and endmember volatile concentrations in the lowest chlorinity fluids are approximately twice as high as in the highest chlorinity fluids. The gradients in composition across the vent field and measured vent fluid temperatures >400°C are consistent with supercritical phase separation and loss of brine phase below the seafloor. The factor-of-2 variation in CO2 (and H2S) is larger than expected for loss of a very high-chlorinity brine. Concentrations of iron and manganese are not positively correlated with chlorinity, suggesting that temperature and pH are more important in controlling metal solubility. Elevated ammonia and bromide/chloride ratios indicate that there has been subseafloor interaction between the hydrothermal fluids and organic matter, and high boron concentrations point to a sedimentary source.

In situ measurements of dissolved manganese and total dissolved iron were conducted in hydrothermal plumes over the Juan de Fuca Ridge using a submersible chemical analyzer (Scanner). The Scanner was deployed as part of a CTD/transmissometer/rosette instrument package on both tow-yos and vertical casts during the VENTS Leg I cruise in 1989. Dissolved manganese and total dissolved iron concentrations, along with temperature and light attenuation anomalies, were determined over the ridge crest every 5 s. Discrete samples for laboratory analyses of dissolved iron II, total dissolved iron II+III and manganese were also collected. Metal to heat ratios (Me:Q) measured in situ were extremely variable in one steady state plume, while an event plume had constant Me:Q. Uniform values of Mn:Q in the event plume demonstrate that Mn behaves conservatively in the near-field plume. Variability in the Mn: Q ratios in a steady state plume indicated the presence of at least two hydrothermal sources with distinct Me:Q values. A simple mixing model shows that the contribution of Mn from high Me:Q sources, with a composition characteristic of black smoker vents, varies between 1% and 99% within the core of the steady state plume with an average value of 55%. On average, over 50% of the excess heat within the plume originates from low Me:Q ratio sources, with a composition characteristic of low-temperature, diffuse flow vent fluids. Less than 4% of the volume of hydrothermal fluids in the plume originates from black smokers. The Fe II concentrations were used to provide an estimate of plume age on a transect across the ridge axis. Plume ages were about 2.5 days on axis and >12 days off axis. These plume ages were modeled to provide estimates of plume transport and horizontal diffusion and show excellent agreement with ages determined using 222Rn.

Concentrations of trace metals in Fe- and Cl-rich hydrothermal solutions from the southern Juan de Fuca Ridge (SJFR) have been determined and corrected for residual precipitates formed in the sampler. Precipitate corrections for Ag, Cd, Cu, Mo, Pb, Sb, and Zn were variable, contributing an average of 20% to total solution concentrations whereas corrections for Co averaged less than 5% and essentially no corrections were required for Fe and Mn. Values for Cu, Co, and Mo in these solutions showed a strong dependence on temperature with sharp decreases in concentrations as temperatures decreased to less than 320°C. In addition, and unlike most other metals studied, all vent fluids from the SJFR were almost completely depleted in Mo relative to seawater values of about 110 nmol kg−1. In contrast to the Cu-Co-Mo group, concentrations of Ag, Cd, Sb, and Pb correlated well with those for Zn and are presumed to follow a distribution that is influenced less by temperature over the 246 to 332°C range encountered in this study and more by the combined chemical processes that control Zn levels along the transport pathway from the deep reaction zone to the vent orifice.

Dissolved and total 234Th (t1/2 = 24.1 days) profiles were measured over the hydrothermal field located within the North Cleft segment of the Juan de Fuca Ridge. Above the hydrothermal effluent plume, which rose approximately 200 m from the seafloor, ≥90% of the 234Th was dissolved, and the total 234Th activity was in equilibrium with its parent uranium. This is typical of deep water for which the residence time of dissolved 234Th with respect to particle uptake is greater than 200 days. Within the effluent plume, however, as little as 30% of the 234Th was in the dissolved state and the residence time of dissolved 234Th with respect to particle uptake was then of the order of only a few weeks. The flux (JTh) of dissolved 234Th to the particle phase was about 300 disintegrations per minute (dpm) cm−2 yr−1. The total 234Th was still approximately in equilibrium with U, indicating that the residence time of the 234Th-bearing particles with respect to sinking was greater than 100 days. Particulate P, V, and As, while insignificant above the plume, attained concentrations as high as 60, 0.65, and 0.45 nmol/L within the plume. The particulate [P]/[234Th] ratio and the 234Th flux (JTh) are used to calculate a removal flux of P resulting from hydrothermal scavenging of ∼50 μg cm−2 yr−1 or approximately 5.0× 106 g P yr−1 per kilometer of active ridge crest. Similarly, 1.0×105 g km−1 of V and As are removed per year from the water column by hydrothermal scavenging over this region. Extrapolation to a global scale suggests that ridge crest scavenging removes ≤10% of these elements from the ocean.

A unique data set for hydrothermal Mn and Fe was collected at Cleft segment on the Juan de Fuca Ridge between 1983 and 1991. The data set includes observations of focused and diffuse venting fluids and neutrally buoyant plumes formed by chronic and episodic venting. Manganese/heat and iron/heat ratios for plumes from the north end of the Cleft segment were combined with independently determined estimates of plume heat flux to yield annually averaged chronic venting fluxes for Mn of 0.36±0.17 mol s−1 and for Fe of 0.61±0.34 mol s−1. Over 6 years of plume measurements at North Cleft segment, observed episodic hydrothermal discharge accounted for −15% of the total vented Mn and −35% of vented Fe. The chronic fluxes for Mn and Fe at a second venting center located at the south end of the Cleft segment were estimated to be approximately equal to the fluxes at North Cleft segment. Chronic plumes at North Cleft segment are mixtures of focused and diffuse discharge that contribute heat, Mn, and Fe in variable proportions. Similar examination of South Cleft segment data strongly suggests the presence of an as yet unobserved venting source relatively depleted in Mn and Fe but contributing substantially to the overall heat. Temporal and spatial variations in the concentrations of Mn and Fe and in Mn/heat and Fe/heat ratios for focused seafloor vents were difficult to resolve within complex chronic plumes. Manganese/heat and iron/heat ratios of megaplumes suggest they may have derived from reservoirs of diffuse fluids while smaller event plumes may have formed by different processes and have properties similar to chronic plumes. The accurate assessment of segment-scale hydrothermal fluxes of Mn and Fe requires coordinated measurements of representative seafloor sources and the neutrally buoyant plume that integrates all seafloor discharge.

In 1990 and 1991, particles from buoyant and neutrally buoyant hydrothermal plumes above hydrothermal vents at the North Cleft segment of the Juan de Fuca Ridge were sampled to study their changing composition and fluxes away from the vent field. In the rising buoyant plume, >75% of the P, V, Cr, and As scavenging from seawater by hydrothermal precipitates occurs in the first 50 m above the vent. Cu and Zn are most enriched in buoyant plume particles collected from the first few meters above the vent. However, the degree of enrichment decreases very rapidly with increased height above the vents due to sedimentation of the more dense Cu- and Zn-rich sulfide phases. Using the plume data, coupled with the results of our analysis of sediment trap samples, we estimated that more than 99% and 99.9%, respectively, of the total hydrothermal Fe and Mn produced at the vent field are transported beyond the vent field and dispersed in the open ocean.

Hydrothermal vent fluids from the North Cleft segment of the Juan de Fuca Ridge between 44°54′ and 45°00′N were sampled in 1988, 1990, 1991, and 1992. In 1988, chloride-depleted (relative to seawater) diffuse fluids with low dissolved metal concentrations (relative to other mid-ocean ridge (MOR) fluids) were sampled over a distance of ∼10 km along axis. In 1990 and 1991, both high-temperature (>200°C) and diffuse vents were consistently metal- and chloride-enriched. The end-member compositions of high-temperature vents and nearby diffuse vents were very similar. There were small but significant correlated changes in diffuse and high-temperature composition from 1990 to 1991. The large-scale but temporary venting of low-chlorinity fluids over the entire North Cleft segment provides a resolution to the apparent mass imbalance implied by observations of continuous venting of chloride-enriched fluids. It is possible that a volcanic event along the North Cleft segment prior to 1987, for which there is firm geologic evidence, initially caused a boiling event which resulted in the preferential venting of vapor-enriched fluids through 1988, followed by a transition to brine-enriched fluids by 1990. High iron, low sodium, and low Sr/Ca ratios in the high-chlorinity fluids suggest that the brine phase has continued to react and approach reequilibration with an alteration mineral assemblage after the phase separation event. The absence of chloride-depleted fluids from 1990 onward, and the systematics of lithium, boron, and manganese with chloride in the high-temperature fluids from North Cleft suggest that the evolution toward lower chlorinity at Monolith vent from 1990 to 1992 is caused by progressive dilution of a brine with hydrothermal seawater.

We conducted a survey of water-column hydrothermal plumes along a 350-km long section of the East Pacific Rise (EPR) axis stretching from 8°45′N to 11°50′N, including the region at 9°45′–54′N where 1991 ALVIN dives found evidence for recent seafloor volcanic eruptions. Our survey included measurements of temperature anomaly, light attenuation (suspended particles), methane, hydrogen, iron, manganese, helium, and aluminum. We detected strong light attenuation plumes in two main regions: south of the Clipperton Transform Fault (CTF) from 9°27′N to 9°57′N and north of the CTF from 11°05′N to 11°35′N. However, the plumes at these two regions had very different physical and chemical characteristics. South of the CTF, where the ridge is thought to be magmatically robust, the plumes had very high ratios of ³He/heat, methane/Mn and S/Fe. Plumes north of the CTF had much lower ratios of ³He/heat, methane/Mn, and S/Fe. These striking differences in volatile/heat and volatile/metal ratios suggest that the plumes in the vicinity of 9°50′N were derived from a young, evolving hydrothermal system, while the activity at ∼11°10′N is from an older, stable system. This survey demonstrated that water-column plume measurements can be used not only to accurately locate regions of hydrothermal activity, but also to determine the first-order chemical and physical characteristics of the hydrothermal fluids exiting from the seafloor.

Dissolved silica (Si) and inorganic germanium (Ge) concentrations were measured in hydrothermal fluids from black smoker vents on the East Pacific Rise (21°N EPR) and the Southern Juan de Fuca Ridge (45°N SJdFR: North and South Cleft Sites, Axial Volcano). These typically display end-member concentrations ranging from 16 to 23 mM (Si) and 150 to 280 nM (Ge), and end-member Ge/Si ratios clustering between 8 and 14 × 10−6, more than 10-fold greater than the ratio entering the ocean via rivers (0.54 × 10−6) and being recycled in seawater (0.7 × 10−6). ‘Excess’ concentrations of dissolved Si and Ge above oceanic background are observed in mid-water hydrothermal plumes over mid-ocean ridge (MOR) spreading centers on the Southern EPR (SEPR) (10°–20°S) and the SJdFR. The largest Si and Ge concentration anomalies occur over the North Cleft Segment of the SJdFR. These are a factor of three greater than anomalies over the SEPR (10°–20°S). Excess Ge correlates with excess3He in plumes at a Ge/3He molar ratio of about 1 × 104, approximately the same ratio as in black smokers. These observations, combined with low abundances of Ge in FeMn-rich metalliferous sediments, suggest that Ge (and Si) behave conservatively in mid-ocean ridge hydrothermal plumes. A simple ocean Si and Ge balance, constrained by the global river silica flux and Ge/Si ratios in hydrothermal vents, rivers and biogenic silica, suggests that the global hydrothermal silica flux is about 1–4 × 1011 mole yr−1, much lower than that estimated from3He. Either (1) 70–80% of the Ge flux to the ocean is removed in as-yet undiscovered sinks (not opal), or (2) only 10% of the mantle to ocean3He and heat fluxes is associated with MOR hydrothermal convection through the 350°C isotherm (90% is off-ridge), or (3) the oceanic Ge/Si,3He/ (and87Sr86Sr) balances today are far from steady-state.

In 1988 and 1989, particles from on- and off-axis neutrally buoyant plumes over the Juan de Fuca Ridge were sampled to study their changing composition with distance from the ridge crest. Plume depth and particle composition in off-axis hydrothermal plumes were used to identify plume sources and trajectories. Iron and phosphorus concentrations and, in particular, P/Fe molar ratios were found to be sensitive indicators of hydrothermal phases in the suspended matter. Phosphorus, vanadium, and arsenic are scavenged from solution where dissolved iron from hydrothermal vents oxidizes in seawater and forms submicrometer-sized particles of Fe oxyhydroxide. The P/Fe molar ratio (∼0.23) is essentially the same for vent fields along Cleft and Endeavour segments, indicating that the scavenging processes are the same for both regions. High P/Fe ratios are observed over the South and North Cleft Segment vent fields and to the west along the Vance Seamount Chain. The seamounts evidently function as a barrier, bathymetrically steering the hydrothermal plumes to the west away from the ridge crest. Since particles from the Endeavour vent field are greatly enriched in copper relative to hydrothermal plumes from other vent fields along the Juan de Fuca Ridge, copper concentrations in the plumes can be used to distinguish plume sources. Based on water column depth and elemental composition, hydrothermal plume particles from specific vent fields on the Juan de Fuca Ridge may be tracked as much as several tens of kilometers from their source using particle compositions.

[1] Fifty submarine volcanoes of the Mariana Arc, covering 1200 km from 13.5°N to 23.2°N, were surveyed for hydrothermal activity. Of these 50 volcanoes, eight showed limited evidence of hydrothermal activity, while another 10 volcanoes displayed intense, chemically rich, hydrothermal plumes that allow detailed chemical characterization and insight into the hydrothermal activity forming the fluids that generate the plumes. The most active volcanoes exhibit a wide range of CO2 to pH relationships, from the venting of acid-rich fluids, to CO2-rich fluids, to fluids rich in alkalinity and CO2. These pH-CO2-alkalinity relationships are partially responsible for the wide range of Fe:Mn (3 to 32) observed at the different volcanoes. This chemical heterogeneity is further manifest in a wide range of CO2:3He (3 × 109 to 55 × 109), indicating that for all but one, >80% of the CO2 venting from these volcanoes has a slab source. The helium, by contrast, has an upper mantle isotopic signature. The range of chemical conditions suggests that these volcanoes occupy various states of eruptive evolution ranging from ongoing magmatic activity to highly evolved hydrothermal systems.

HYDROTHERMAL vents along mid-ocean ridges are an important source of elements such as lithium, silicon, manganese and iron to the world's oceans1. The venting produces both episodic and steady-state hydrothermal plumes with unique thermochemical signatures in the mid-water column. The particulate phases in these plumes (predominantly iron oxides and hydroxides) also scavenge phosphorus, vanadium, arsenic, lead, polonium and several rare-earth elements from sea water2–5. Thus, on a global scale, hydrothermal plumes are both a source for some elements and a sink for others. Ultimately, the particulate metals precipitated from plumes form extensive regions of metalliferous sediments over the crests and flanks of mid-ocean ridges6, 7. Although the metalliferous sediment coverage is vast and well documented, only a tiny fraction of the vents responsible for these sediments have been located (Fig. la). To date, both the number and location of hydrothermal vents and the detailed distribution of chemical constituents within the resultant plumes are poorly understood because of under-sampling of the mid-ocean ridges and the overlying waters. Here we present the results of high-resolution mapping of the chemical and thermal characteristics of hydrothermal plumes in near real time using a novel submersible chemical analyser (Scanner)8, 9 and a conductivity/temperature/depth/trans-missometer instrument package (CTDT)10. We show that the kinetics of iron oxidation in the plume can be used to constrain estimates of the plume's age, and that variation in the ratio of manganese content to excess heat can be explained by the mixing of several different vent fluids.

Studies of hydrothermal plumes on the Juan de Fuca Ridge and the Mid-Atlantic Ridge indicate that newly-formed Fe oxyhydroxides, formed as a consequence of hydrothermal venting, readily scavenge P and As from seawater in proportion to the dissolved concentrations of these elements. The Fe content of the suspended matter decreases from ∼40% near the vents to

Submersible dives on 22 active submarine volcanoes on the Mariana and Tonga-Kermadec arcs have discovered systems on six of these volcanoes that, in addition to discharging hot vent fluid, are also venting a separate CO2-rich phase either in the form of gas bubbles or liquid CO2 droplets. One of the most impressive is the Champagne vent site on NW Eifuku in the northern Mariana Arc, which is discharging cold droplets of liquid CO2 at an estimated rate of 23 mol CO2/s, about 0.1% of the global mid-ocean ridge (MOR) carbon flux. Three other Mariana Arc submarine volcanoes (NW Rota-1, Nikko, and Daikoku), and two volcanoes on the Tonga-Kermadec Arc (Giggenbach and Volcano-1) also have vent fields discharging CO2-rich gas bubbles. The vent fluids at these volcanoes have very high CO2 concentrations and elevated C/3He and δ 13C (CO2) ratios compared to MOR systems, indicating a contribution to the carbon flux from subducted marine carbonates and organic material. Analysis of the CO2 concentrations shows that most of the fluids are undersaturated with CO2. This deviation from equilibrium would not be expected for pressure release degassing of an ascending fluid saturated with CO2. Mechanisms to produce a separate CO2-rich gas phase at the seafloor require direct injection of magmatic CO2-rich gas. The ascending CO2-rich gas could then partially dissolve into seawater circulating within the volcano edifice without reaching equilibrium. Alternatively, an ascending high-temperature, CO2-rich aqueous fluid could boil to produce a CO2-rich gas phase and a CO2-depleted liquid. These findings indicate that carbon fluxes from submarine arcs may be higher than previously estimated, and that experiments to estimate carbon fluxes at submarine arc volcanoes are merited. Hydrothermal sites such as these with a separate gas phase are valuable natural laboratories for studying the effects of high CO2 concentrations on marine ecosystems.

[1] Submarine volcanic arcs extend for fully a third of the length of the mid-ocean ridge system, but systematic investigation of hydrothermal venting on those arcs is sparse. Here we report on the distribution of hydrothermal activity along the intraoceanic Mariana arc. We identify 76 volcanic edifices along 1370 km of arc, grouped into 60 “volcanic centers,” of which at least 26 (20 submarine) are hydrothermally or volcanically active. The overall volcanic center density is 4.4/100 km of arc, and that of active centers is 1.9/100 km. An equal length of the Tonga-Kermadec arc holds fewer centers, 2.9/100 km, but a similar density of active centers, 1.8/100 km. The lower percentage of active centers on the Mariana arc results from its having both fewer caldera volcanoes, since on both arcs calderas are twice as likely as cones to be active, and a lower percentage of active calderas and cones. Active centers are found from 80 to 230 km above the subducting Pacific slab, and a quarter lie behind the arc front. The frequency distribution of center spacing along the arc front peaks between 20 and 30 km and shows the asymmetric, long-tail shape typical for many other arcs, especially those with volcano populations >∼20. There is no evidence for a regular spacing of volcanic centers. Using the new Mariana data and recent data from the Tonga-Kermadec arc, we estimate that all intraoceanic arcs combined may contribute hydrothermal emissions equal to ∼10% of that from the global mid-ocean ridge.

The biogeochemical scavenging of dissolved manganese (Mn) from hydrothermal plumes was investigated using radiotracer ( 54 Mn) techniques. The measured scavenging rate constant, k 1 , was lowest in the buoyant plume ( −1 ), increasing to ∼2 y −1 in the non-buoyant plume at distances of 20 km from the ridge valley axis. The direct biological contribution to the dissolved Mn scavenging rate (i.e the fraction suppressed by the addition of a metabolic poison) also increased over the same distances, being minor or absent at plume depths in the proximal plume, yet the major component at distal plume stations. These and other data suggest that the capacity for scavenging dissolved Mn onto particles evolves with increasing age of the plume, suggestive of a microbial response to changing conditions within the plume. Estimated maximum scavenging rates of dissolved Mn onto particles ( R DMn = k 1 [DMn])were noted at plume depths for all stations, a function of very high dissolved Mn concentrations in the case of the buoyant plume and proximal non-buoyant plume. R DMn values, integrated over plume depths, ranged from 3.4 to 1.7 mM m −2 y −1 for the non-buoyant plume at on-axis and off-axis stations, respectively. The application of the data to the dispersal of hydrothermal constituents and to plume aging is discussed.

The distributions of dissolved, particulate and sedimentary phosphorus were measured in the region of the Juan de Fuca Ridge to determine the impacts of hydrothermal processes on the phosphorus cycle in the oceans. Significant negative dissolved phosphate anomalies, ranging from 0 to 60 nmol/l, were observed in the water column at depths between 1900 and 2300 db. The largest anomalies (

[1] Quantifying hydrothermal venting at the boundaries of tectonic plates is an outstanding geoscience problem. Considerable progress has been made by detailed surveys along mid-ocean ridges (MORs), but until recently little was known about fluid venting along volcanic arcs. We present the first systematic survey for hydrothermal venting along the 425-km-long south Tonga arc and new chemistry data for particle and thermal plumes previously reported along an adjacent 88-km-long section of the back-arc Valu Fa Ridge (VFR). Eleven hydrothermal plumes, recognized by their anomalous light backscattering, Eh, temperature, pH, dissolved He-3, CH4, and total dissolvable Fe and Mn, were identified arising from seven volcanic centers along the arc. Five plumes on the VFR were characterized chemically. Vent field density for the south Tonga arc was 2.6 sites/100 km of arc front, comparable to that found by surveys of the Kermadec arc (1.9 to 3.8 sites/100 km) and to MORs in the eastern Pacific (average value for 2280 km of surveyed ridgecrest: 3.2 sites/100 km). A "vent gap'' occurs along a 190 km section of the arc closest to the VFR, and a site density twice the average for MORs on the eastern edge of the Pacific plate was found on this part of the VFR (6.6 sites/100 km). We suggest magmas ascending under the adjacent south Tonga arc have been captured by the VFR. While chemical enrichments of plumes on the south Tonga arc were, in general, slightly less than those on the Kermadec arc, several instances of excessive anomalies in pH suggest a similar presence of fluids enriched in magmatic volatiles (CO2-SO2-H2S). Locally, venting on the VFR has contributed to accumulations of 3 He, Fe, and Mn within the southern Lau basin. On a broader scale, our results provide considerable support for the notion that venting from intraoceanic arcs on the convergent margin of the Pacific plate adds significantly to the total hydrothermal input into the Pacific Ocean.

[1] The 2,500-km Kermadec-Tonga arc is the longest submarine arc on the planet. Here, we report on the second of a series of cruises designed to investigate large-scale controls on active hydrothermal venting on this arc. The 2002 NZAPLUME II cruise surveyed 12 submarine volcanic centers along ∼580 km of the middle Kermadec arc (MKA), extending a 1999 cruise that surveyed 260 km of the southern Kermadec arc (SKA). Average spacing between volcanic centers increases northward from 30 km on backarc crust along the SKA, to 45 km on backarc crust along the southern MKA, to 58 km where the MKA joins the Kermadec Ridge. Volcanic cones dominate in the backarc, and calderas dominate the Kermadec Ridge. The incidence of venting is higher along the MKA (83%, 10 of 12 volcanic centers) than the SKA (67%, 8 of 12), but the relative intensity of venting, as given by plume thickness, areal extent, and concentration of dissolved gases and ionic species, is generally weaker in the MKA. This pattern may reflect subduction of the ∼17-km-thick oceanic Hikurangi Plateau beneath the SKA. Subduction of this basaltic mass should greatly increase fluid loss from the downgoing slab, initiating extensive melting in the upper mantle wedge and invigorating the hydrothermal systems of the SKA. Conversely, volcanic centers in the southern MKA are starved of magma replenishment and so their hydrothermal systems are waning. Farther north, where the MKA centers merge with the Kermadec Ridge, fewer but larger magma bodies accumulate in the thicker (older) crust, ensuring more widely separated, caldera-dominated volcanic centers.