1 Introduction

The southern part of the Mariana Trough is an active backarc spreading center that had propagated southward (Fig. 45.1a) (Fryer 1995). Several hydrothermal fields have been mapped around 12°50′N in an alignment perpendicular to the spreading axis (Fig. 45.1b) (Wheat et al. 2003; Ishibashi et al. 2004; Utsumi et al. 2004; Kakegawa et al. 2008; Nakamura et al. 2013). This area is far from continents and large islands, so terrestrial organic matter is absent, and the primary production in surface seawater is poor. On the seafloor, sinking organic matter is very sparse, but flourishing ecosystems are sustained around hydrothermal fields despite high temperatures and pressures (Deming and Baross 1993; Takai and Horikoshi 1999; Kelley et al. 2002). Ecosystems utilizing Fe for energy have been extensively documented, especially around low-temperature hydrothermal vents (e.g., Boyd and Scott 2001; Edwards 2004; Edwards et al. 2003, 2011; Little et al. 2004; Kato et al. 2009; Langley et al. 2009; Toner et al. 2009). Ecosystems based on metabolism of inorganic materials are of interest for their potential insight into the origins of life (e.g., Alt 1988; Juniper and Fouquet 1988; Duhig et al. 1992; Juniper and Tebo 1995; Hofmann and Farmer 2000; Reysenbach and Cady 2001).

Fig. 45.1
figure 1

(a) A topographic map showing the Izu-Bonin-Mariana trench/arc/back-arc system on the eastern edge of the Philippine Sea plate. A box denotes the area of Fig. 45.1b. (b) A topographic map around 12°50′N. The stars represent the positions of hydrothermal fields in this study. S, Y, A, P, and U denote the initial letter of Snail, Yamanaka, Archaean, Pika, and Urashima sites, respectively

Full size image

In the context of the TAIGA project, hydrothermal activity in the Southern Mariana Trough is recognized as Sulfur TAIGA and Iron TAIGA (Urabe et al. Chap. 1), and the project aims to elucidate what kind of ecosystems are prevalent, what kind of metabolisms they employ, and what kind of chemical environments they live in (Urabe et al. 2009). In this tectonic setting, in which hydrothermal activities are extending southward, how ecosystems propagate may shed light on more general questions about the evolution of life processes. Also, a fundamental question about mass flux at Earth’s surface is the influence of subducting materials and magmatic volatiles. Extremely acidic fluids affected by magmatic volatiles raise the solubility of metals and mobilize them from rocks to fluids (Gamo et al. 1997b, 2006; Gena et al. 2001; Yang and Scott 2006). This aspect of the metal supply may have a great influence on Fe-utilizing ecosystems. It was with these topics in mind that we investigated the chemical and isotopic compositions of hydrothermal fluids in the Southern Mariana Trough.

2 Geological Setting

In the Mariana Trough, the Pacific plate subducts beneath the Philippine Sea plate, and the shape of the trough is convex to the east as a result of backarc spreading (Fig. 45.1a). Spreading activity started between 5 and 10 Ma (Hussong and Uyeda 1981; Karig 1971; Bibee et al. 1980; Eguchi 1984; Yamazaki and Stern 1997), and the northern part of the Mariana Trough represents the early stage of rifting of a backarc basin (Yamazaki et al. 1993). The central part of the Mariana Trough, from 16° to 20°N, represents the early stage of backarc basin spreading, and the volcanic arc and backarc are clearly distinguishable (e.g., Hart et al. 1972; Hawkins 1977; Fryer et al. 1981; Natland and Tarney 1981). The backarc is very close to the volcanic arc south of 16°N, and the spreading center of the backarc is ambiguous south of 12.5°N (Hawkins 1977; Smoot 1990). The hydrothermal fields in this study are located around 13°N. Five hydrothermal systems have been discovered around 13°N, which are aligned perpendicular to the axis (Fig. 45.1b). The Snail and Yamanaka sites are on the spreading axis, and the Archaean, Urashima, and Pika sites are off the axis. Around the area, hydrothermal activities have been documented in the TOTO caldera (Gamo et al. 2004), Forecast vent field (Johnson et al. 1993), and Alice Springs field (Craig et al. 1987; Gamo 1993; Gamo et al. 1997a). Descriptions of these hydrothermal systems are found in the InterRidge Vents Database (Beaulieu 2013).

2.1 Snail Site

Snail site is on the spreading axis (12°57.214′N, 143°37.147′E, depth: 2,860 m) (Fig. 45.1b), where hydrothermal fluids as hot as 248 °C were observed in 2003 (Wheat et al. 2003). However, most of the seafloor around the site is covered by snails exposed to low-temperature fluid around 100 °C, and covered by a yellow microbial mat and shimmering fluid below 100 °C (Wheat et al. 2003). The microbial mat may be composed of iron hydroxide and associated with Fe-oxidizing bacteria (Kato et al. 2010). In 2004 and 2010, the Benthic Multicoring System (BMS) penetrated several meters below the seafloor, and basalt was recovered (Urabe et al. 2004). In 2012, an ROV investigation found that hydrothermal fluid of 45 °C was discharging from the casing pipe that was placed in the BMS drilling hole in 2004 (NT12-24 Cruise Report).

2.2 Yamanaka Site

Yamanaka site is on a mound about 30 m high on the spreading axis (12°57.6′N, 143°36.7′E, depth: 2,830 m), 1 km northwest of Snail (Fig. 45.1b). Observations have documented many sea anemones and dead chimneys as well as shimmering fluid below 20 °C seeping from the rocky seafloor (Kakegawa et al. 2008). Magnetic mapping in 2009 showed strong magnetization here, suggesting that hydrothermal alteration is weak (Fujii et al. 2013). Based on these facts, Yamanaka has been considered to be past its peak of hydrothermal activity (Yoshikawa et al. 2012).

2.3 Archaean Site

Archaean site is on a sulfide mound ~100 m high that is 2 km from the spreading axis (12°56.35′N, 143°38.0′E, depth: 2,990 m) (Fig. 45.1b). The ROV ROPOS, during cruise TN167A in 2004, documented black smokers venting fluid at temperatures up to 343 °C (Ishibashi et al. 2004). The BMS drilling campaign in 2010 recovered basalt and andesite from the base of the mound (Nakamura et al., Appendix 3-A1).

2.4 Pika Site

Pika site is on a 75 m-high knoll, with two peaks 5 km from the spreading axis (12°55.1′N, 143°38.9′E, depth: 2,830 m) (Fig. 45.1b). The manned submersible Shinkai 6500, during cruise YK03-09 in 2003, documented a black smoker venting high-temperature fluid up to 330 °C at the top of the western peak of the knoll and no hydrothermal activity on the eastern peak (Utsumi et al. 2004). In 2004 the BMS campaign recovered andesite from borehole APM03 (Kakegawa et al. 2008), which means drilling site No.3 of Archaean Park Project in Mariana (Urabe et al. 2001).

2.5 Urashima Site

Urashima site is on a mound several meters high 500 m north of Pika (12°55.30′N, 143°38.89′E, depth 2,922 m) (Fig. 45.1b), and several gray smokers venting hydrothermal fluids as hot as 280 °C were documented by Shinkai 6500 during cruise YK10-10 in 2010 (Nakamura et al. 2013). A strong negative magnetic anomaly was observed by the autonomous underwater vehicle (AUV) Urashima in 2009 (Seama et al. Chap. 17), suggesting demagnetization by hydrothermal alteration (Rona 1978; Tivey and Johnson 2002). The site includes black dead chimneys, red-brown shimmering chimneys, and areas of yellow microbial mat exposed to low-temperature hydrothermal fluids. The microbial mat is probably composed of iron hydroxide like that at Snail.

3 Materials and Methods

3.1 Sampling Methods

Hydrothermal fluids were collected during YK03-09 (October, 2009), TN167A (March, 2004), YK05-09 (July, 2005), YK10-10 (August, 2010), YK10-13 (October, 2010), and NT12-24 (September, 2012) (Suppl. 45.1a and 45.1b) using the ROCS and the WHATS (Tsunogai et al. 2003; Saegusa et al. 2006). The WHATS is multi-bottle gastight sampling system, which avoids degassing and contamination by tightly closing the valves of sample bottles in situ. During sampling, the bottle inlet was held as close as possible to the hydrothermal vent, but in typical operation the inlet sometimes drops off the vent and draws in seawater. With allowance for a small degree of seawater admixture, the sample represents the average composition of hydrothermal fluid during the sampling period. The fluid temperature was recorded during sampling. The samples were collected with two bottles at each vent, one for gas analyses and the other for fluid analyses.

The sample for fluid analyses was first subsampled into two 5-mL vials for measurement of pH, alkalinity, and H2S concentration. The remaining fluid was then filtered using a 0.45-μm pore-size disk filter attached to a plastic syringe and distributed into 15- and 30-mL plastic bottles. The 15-mL subsample was measured for NH4 + and Si concentrations on board, and Cl− and SO4 2− concentrations on land. The 30-mL subsample was combined with 300 μL of 3N HNO3 on board the tender ship and then measured for major and minor elements on shore.

The sample bottle for gas analyses was processed as soon as possible (Konno et al. 2006); the fluid in the bottle was transferred to a 300-mL evacuated glass container, then acidified with amidosulfuric acid to convert all dissolved carbonate species to CO2 in the headspace of the container. The fluid was treated with sufficient HgCl2 to thoroughly convert H2S to HgS in the vacuum line. The gas phase was transferred to a 50-mL evacuated stainless steel container for gas analyses on shore, and then the residual fluid in the container was filtered for measurement of major elements on shore.

3.2 Analysis

The pH was measured by electrode with a precision of ±0.02 (Gieskes et al. 1991). The calibration was done using buffers of pH 4.01 and pH 6.86 once a day before the measurements. Alkalinity was measured by HCl titration, which endpoint was determined by the Gran’s plot, with a precision of ±2 % (Gieskes et al. 1991). The concentration of H2S was measured by colorimetry using methylene blue with a precision of ±10 %. Ammonia concentration was measured by colorimetry using phenol blue with a precision of ±8 %. The Si concentration was determined by colorimetry using molybdate blue with a precision of ±1 %. Chlorinity was analyzed by the Mohr method with a precision of ±1 %. The SO4 2− concentration was determined by ion chromatography with a precision of ±4 %. Major and minor elements (Na, Ca, Mg, Sr, Ba, Mn, Fe, Si, B, and Li) were analyzed by inductively coupled plasma atomic emission spectroscopy (ICP-AES). The K concentration was analyzed by atomic absorption spectrometry. The Na concentration was also calculated independently on the basis of charge balance.

Gas concentrations of O2, N2, CO2, H2, and He were measured by using gas chromatography with a thermal conductivity detector. CH4 concentration and its stable carbon isotope ratio were determined with a continuous-flow isotope ratio mass spectrometer (CF-irMS) (Tsunogai et al. 2000). The hydrogen isotope ratio (D/H) of CH4 was analyzed by using a CF-irMS with a precision of ±10 ‰. The hydrogen isotope ratio of H2 was analyzed by using a CF-irMS with a precision of ±4 ‰ (Komatsu et al. 2011). Helium isotope ratios (3He/4He) were measured by using noble gas mass spectrometry with a precision of ±2 % (Sano et al. 2006). The ratio of 4He to 20Ne was determined by using quadrupole mass spectrometry with a precision of ±10 % (Sano et al. 2006). Dissolved organic matter was measured with a Shimadzu TOC-5000 total organic carbon analyzer with a high-temperature catalytic method.

4 Results

The analytical results for hydrothermal fluids from the five sites are shown in Suppl. 45.1a for fluid chemistry and Suppl. 45.1b for gas chemistry, together with the maximum and average temperatures recorded during sampling.

4.1 Calculation of End-Member Compositions

Graphs of the concentrations of each chemical component relative to Mg concentration are shown in Fig. 45.2. Under the assumption that pure hydrothermal fluid is free of Mg, its chemical composition was calculated as the y-axis intercept of linear relationship among data plots (Von Damm 1995). The Yamanaka site yielded only low-temperature hydrothermal fluids with compositions that could not be distinguished from that of seawater, and this site is not discussed further. End-member concentrations of the hydrothermal fluids in the remaining four sites are listed in Table 45.1 for aqueous components and Table 45.2 for volatile components, and discussed below.

Fig. 45.2
figure 2figure 2figure 2

A plot of each chemical component vs. Mg concentration in the hydrothermal fluid for each site in the Southern Mariana Trough in this study

Full size image
Table 45.1 End-member concentrations of aqueous components in the hydrothermal fluids for all the sites in the Southern Mariana Trough in this study
Full size table
Table 45.2 End-member compositions of volatile components in the hydrothermal fluids for all the sites in the Southern Mariana Trough in this study
Full size table

The end-member carbon isotopic composition of methane, δ13C(CH4), was estimated as the intercept of the y-axis on a graph of δ13C(CH4) against the reciprocal of CH4 concentration, as shown in Fig. 45.3 (Keeling 1961). Estimates of this type are based on the assumption that the background concentration is negligible. In the case of methane, background CH4 in seawater is several nM, whereas CH4 in hydrothermal fluid is several hundred μM, a difference of five orders of magnitude.

Fig. 45.3
figure 3

Plots of δ13C vs. 1/(CH4 concentration) in the hydrothermal fluids for all the sites in the Southern Mariana Trough in this study

Full size image

However, because the background CO2 concentration is substantial in seawater (2 mM) compared to hydrothermal fluid (several tens mM), we relied on a method used in atmospheric chemistry to evaluate end-member δ13C(CO2) (Miller and Tans 2003). As illustrated in Fig. 45.4, a product that CO2 concentration multiplies δ13C(CO2) is plotted to CO2 concentration, when the slope can be taken as δ13C(CO2) of the source.

Fig. 45.4
figure 4

Relationship between CO2 concentrations and isotopic index from all the sites in the Southern Mariana Trough in this study are plotted, where the index are calculated by multiplying δ13C(CO2) by CO2 concentrations

Full size image

Helium in hydrothermal gases is a mixture of mantle, crust, and seawater contributions. Because 20Ne is negligible in rock, helium isotope ratios can be corrected for seawater admixture using 4He/20Ne ratios, as shown in Fig. 45.5 (Craig et al. 1978).

Fig. 45.5
figure 5

Relationship between 4He/20Ne and 3He/4He in the hydrothermal fluids for all the sites in the Southern Mariana Trough in this study, together with the mixing line between mantle and seawater

Full size image

4.2 End-Member Concentrations of Dissolved Constituents

Maximum fluid temperatures were as follows: Snail, 116 °C; Archaean, 343 °C; Pika, 330 °C; Urashima, 280 °C (Table 45.1).

The lowest pH ranged from 2.8 to 3.5, and only at Snail was it greater than 3. Except for Snail, these values are lower than that of hydrothermal fluids in mid-ocean ridges ranging about 3–3.5 on average (Von Damm 1995).

The end-members for all sites have negative alkalinity; −0.94 for Snail, −1.41 for Archaean, −1.14 for Pika, and −1.44 for Urashima. These data suggests that these hydrothermal fluids have substantial strong acidity compared with hydrothermal fluid in the mid-ocean ridge setting, where the fluid is rather neutralized during high temperature fluid-rock interactions. This view is supported by relatively low pH in Southern Mariana Trough hydrothermal fluids.

The Si concentrations were measured by colorimetry on board and by ICP-AES on land, and both end-member concentrations were consistent with each other. Values lay in a narrow range between 16.4 and 17.0 mM except for Snail, where the Si concentration was greater than 18 mM.

The end-member Cl− concentrations deviated from that of seawater (550 mM) at all sites. At Archaean, the Cl− end-member in 2004 (466 mM) had a higher concentration than those in 2005 and 2010 (407 mM). At Pika, high-Cl− fluids (603 mM) were consistently reported after its discovery in 2003, but a low-Cl− fluid (460 mM) was recovered in 2010 (Table 45.1). Such Cl− deviations from that of seawater have been reported in many hydrothermal systems in the world, which are thought to be due to phase separation beneath the seafloor (Von Damm 1995). Especially, higher-Cl− and lower-Cl− fluids than that of seawater have been sampled, like those at Pika, suggesting that active phase separation beneath the seafloor has occurred there, as mentioned in many hydrothermal systems in the world (Von Damm 1995). A positive Cl− deviation from that of seawater, like those at Urashima, can be due to input of HCl derived from magmatic volatiles, as mentioned at DESMOS hydrothermal field in Manus Basin (Gamo et al. 1997b). In that case, the hydrothermal fluids also have extremely low pH and an excess of SO4 2− (Gamo et al. 1997b), and we found that hydrothermal fluids from Urashima site have an end-member pH slightly lower than 3 as mentioned above (Table 45.1).

The Na concentration was independently calculated by charge balance as well as measured by ICP-AES. The calculated values were more clearly linear relative to Mg concentrations than the measured values. Although the errors of the calculated concentrations are larger than those of the measured concentrations, we took the calculated values as those of the preferred end-members. The end-member value for Snail was close to that of seawater (460 mM). The end-member Na concentration for Archaean (345 mM) was lower than that of seawater. At Pika, the end-member value in 2003 and 2005 was higher than that of seawater, although sample D1219 W-4, taken in 2010, had a lower value than seawater (Table 45.1). For Urashima, the end-member Na concentration (515 mM) was higher than that of seawater.

For Ca concentrations, the end-members for all four sites had higher values than that of seawater (10 mM). At Archaean, the end-member Ca concentration for 2004 was higher than the one for 2005 and 2010, which is similar to our finding for Cl− concentrations.

For K concentrations, end-member values for all sites fell into a narrow range (31.3–32.8 mM), and the values for Pika and Archaean were significantly higher than those for Snail and Urashima.

For Fe concentrations, the end-members for all four sites had higher concentrations than that of seawater (<1 nM). The end-member value was 0.64 mM for Snail, 2.55 mM for Archaean in 2003–2005, 3.00 mM for Archaean in 2010, 7.38 mM for Pika, and 6.37 mM for Urashima (Table 45.1). The end-member Fe concentration at Snail site is as low or lower when compared to those in the hydrothermal fluids in EPR21°N (0.2–2.4 mM; Von Damm et al. 2002), but otherwise those in the Southern Mariana Trough appear higher as compared with those in EPR21°N. Fe leaches from rocks at low pH attributed to an input of magmatic volatiles (Mottl et al. 1979; Hajash and Chandler 1982; Seyfried and Janecky 1985; Seewald and Seyfried 1990). The cause of Fe enrichment in Pika, Urashima, and Archaean sites would be low pH coming from an input of magmatic volatiles.

The end-members of major elements had two concentrations for both Pika and Archaean and unique values for both Snail and Urashima (Table 45.1).

4.3 Distribution of Gas Components

Dissolved oxygen concentrations showed nonlinear relationships to Mg concentrations, so end-member concentrations were not computed (Fig. 45.2). We just explain the concentration for each Mg concentration (Suppl. 45.1b and Fig. 45.2). At Snail, samples contained Mg concentrations higher than 40 mM and had oxygen concentrations higher than that of seawater, ranging around 100–200 μM. At Archaean, samples had Mg concentrations lower than 40 mM and oxygen concentrations lower than 20 μM. The single sample from Pika contained 12 mM Mg and an oxygen concentration of 18 μM. Urashima yielded samples containing from 7 to 54 mM Mg, and oxygen was not detected in samples with less than 40 mM Mg.

Dissolved nitrogen was detected in all samples, but its concentrations showed nonlinear relationships to Mg concentrations (Fig. 45.2). At Snail, all samples contained Mg concentrations above 40 mM, and nitrogen concentrations were higher than that of seawater (500–1,000 μM). At Archaean, two samples had Mg concentrations around 10 and 30 mM, and the N2 concentration of the latter was higher than that of the former. The single sample from Pika had a Mg concentration of 12 mM and a N2 concentration of about 1 mM.

End-member dissolved CO2 concentrations were calculated for each site from their relation to Mg concentrations (data for Archaean did not fit a linear relation). The end-member values were all higher than that of seawater (2.4 mM). At Snail, although the majority of samples had values close to that of seawater, the end-member value, 78.8 mM, was highest among our sites. Although we did not calculate an end-member concentration for Archaean, the range of CO2 concentrations there was 6.4–32.6 mM. For Pika, samples taken in 2005 had an especially high value (69.1 mM); whereas the end-member concentration for 2003, 2004, and 2010 samples was much lower (33.7 mM). For Urashima, the end-member value was 23.2 mM. The reported range of CO2 concentrations in hydrothermal fluids from EPR are lower than 10 mM (Von Damm et al. 2002), and so the observed CO2 concentrations in hydrothermal fluids in the Southern Mariana Trough in this study are relatively higher than those in EPR. Such CO2 enrichments could be influenced by a little input of magmatic volatiles, as mentioned in hydrothermal systems related to arc systems (Sakai et al. 1990; Gamo 1993; Tsunogai et al. 1994; Gamo et al. 2006; Lupton et al. 2008).

The end-member value of CH4 for Snail was 29.6 μM, although the error was large because values for most samples were close to seawater values. For Archaean, the data were too scattered to calculate an end-member concentration, but sample D903 W-3, from a clear smoker vent with a maximum temperature of 117 °C, had the highest concentration of CH4 in this study, 115 μM. For Pika, the 2005 end-member concentration (29.9 μM) was higher than the end-member concentration for the other years (7.2 μM). The end-member value for Urashima was 10.8 μM.

Hydrogen concentrations were not linearly related to Mg concentrations, and we did not calculate H2 end-member concentrations. The maximum H2 concentrations were 549 μM at Snail, 466 μM at Archaean, 301 μM at Pika and 105 μM at Urashima.

Helium was not detected in Snail samples. For Archaean, the relation of He to Mg concentrations was not clearly linear, but the end-member concentration was calculated to be 2.1 μM. Although we have only one sample from Pika, the end-member value was 1.8 μM. For Urashima, given the helium concentration in seawater (2 nM), the relationship between He and Mg concentrations was smoothly linear, and the end-member value was 2.4 μM.

4.4 Isotopic Compositions of End-Members

The end-member δ13C(CO2) ranged from −2.3 to +0.1 ‰ (all values are relative to VPDB): −1.3 ‰ for Snail, −2.0 ‰ for Archaean, +0.1 ‰ for Pika, and −2.4 ‰ for Urashima (Table 45.2), which was the only value higher than that of seawater (0 ‰).

The δ13C(CH4) values showed a nonlinear relationship to 1/CH4. For Snail, the δ13C(CH4) values in CH4-rich samples were around −5 ‰ and −25 ‰ for Snail, −10 ‰ and −50 ‰ for Archaean, around −5 ‰ for Pika, and −5 ‰ and −0 ‰ for Urashima.

The helium isotope ratios in source fluids, corrected for seawater admixture, were 8.1RA for Snail, 8.4RA for Archaean, 8.3RA for Pika, and 8.0RA for Urashima.

5 Summary

Hydrothermal fluids were sampled from several hydrothermal activities around the Southern Mariana backarc spreading center, and analyzed for chemical and isotopic compositions. The fluid chemistry around the Southern Mariana Trough would be influenced by a little input of magmatic volatiles, leading to low pH and high CO2 concentrations, consequently the hydrothermal fluids are characterized by rich in Fe alongside of the MOR hydrothermal fluids.