Mud volcanoes (MVs) are structures formed by the eruption of mud breccia from below the surface to the surface owing to increased pressure and buoyancy (Mazzini and Etiope, 2017). The number of MVs on land is probably approximately 10³ (Etiope and Milkov, 2004); conversely, that on the seafloor was estimated to be 10³-10⁵ as the sedimentation rate is high (Milkov, 2000). The ejected materials comprise gas, mainly CH₄ and CO₂, mud, rock fragments, and water (Kopf, 2002), which are thought to be related to natural gases, petroleum, and/or gas hydrates (Etiope et al., 2009; Tinivella and Giustiniani, 2012). Some studies have indicated that earthquakes trigger MV activities (e.g., Tsunogai et al., 2012), but not all earthquakes are associated with MVs (Manga et al., 2009; Tingay et al., 2018; Wang and Manga, 2021). To investigate causal relationships, it is essential to monitor MV activity. On land, detailed topographical changes can be monitored using satellites (Kopanina and Shvidskaya, 2021). However, seafloor monitoring is difficult. Feseker et al. (2014) installed a submarine observation station at the Håkon Mosby MV off the coast of Norway, monitored it for 431 days, recorded four major events, and observed a rapid sediment uplift of 1 m (Feseker et al., 2014). Topographical changes on the seafloor can cause submarine landslides, which can lead to accidents involving the disconnection of communication cables (Carter et al., 2012; Carter et al., 2014; Gavey et al., 2017). Therefore, understanding the activities and distribution of submarine MVs and identifying the factors that determine the effects of earthquakes and subducting seamounts on MVs are vital (Kioka, 2020).

The indexes of the activity of submarine MVs can be estimated from the CH₄ flux directly in surface water (Tsunogai et al., 2012), sedimentary activity analysis using microfossil age (biostratigraphy) and ¹⁴C age (Ujiié, 2000), and determination of the depth of origin using various geothermometers (Hensen et al., 2004; Nishio et al., 2015; Lalk et al., 2022; Mitsutome et al., 2023). The greater the depth of origin, the greater was the amount of energy that migrated through the material, suggesting the existence of large-scale mud volcanic activity. Hensen et al. (2004) estimated the temperature at which water reaches equilibrium with clay from water isotope ratios. Nishio et al. (2015) estimated the temperature at which lithium equilibrates with clay using lithium isotope ratios. Lalk et al. (2022) used clamp isotopes of methane to estimate the temperature of methane formation. Mitsutome et al. (2023) used the solubilities of Ar, Kr, and Xe to estimate the depths at which those heavy noble gases were present immediately before. Boron isotope ratio also has the potential to be used as a geothermometer (Palmer, 1991; James et al., 1999; Williams et al., 2001; You and Gieskes, 2001).

Boron (B) has two isotopes with mass numbers of 10 and 11. The B isotope ratio, which is the abundance ratio, indicates that B in fluids released at equilibrium from silicates undergoes smaller isotope fractionation at higher temperatures (Deyhle and Kopf, 2005). In contrast, MVs represent the progression of clay mineral alteration (Kopf and Deyhle, 2002). Deyhle and Kopf (2005) investigated B isotope fractionation in fluid and solid phases in nature and found that the B isotope ratio in the fluid was influenced by multiple factors, such as the weight ratio of the fluid to solid phases in the reaction field, B isotope ratio in the solid phase, reaction temperature, and pH in the reaction field. Chao et al. (2011) investigated the B isotope ratios in the pore water of MVs in Taiwan and found that after high-temperature release from the solid phase at depth, B re-adsorption during the ascent of the fluid resulted in an extremely high B isotope ratio. However, they did not specify on its usefulness as a geothermometer (Chao et al., 2011).

In this study, to verify the method of evaluating the activity of each MV using the B isotope ratio, we did the following. First, we collected pore water from the surface sediments of the MVs off Tanegashima Island and investigated their chemical and isotopic compositions. Next, we estimated the advection velocity using the vertical distribution of the chloride (Cl^-) concentration. Subsequently, we estimated the depth of origin from the equilibrium temperature with clay minerals using sodium (Na) and lithium (Li) concentrations. Finally, we discussed the relationship between the origin temperature and isotopic fractionation of the B isotope ratios in the pore water and sediments.

Tanegashima is located in southwest Japan ( Figure 1A ) in the plate convergence zone where the Philippine Sea plate subducts below the Eurasian plate at a rate of 4.6 cm per year (Seno and Maruyama, 1984). The plate boundary off Tanegashima is at the northern end of the Ryukyu Trench, which is roughly classified as an erosion type, and the Nankai Trough, which extends eastward from the Ryukyu Trench into the accretion type (Clift and Vannucchi, 2004). A wedge of approximately 8 km was observed in the forearc region of the plate boundary off Tanegashima (Ujiié, 2000). The basement around Tanegashima comprises the Eocene Kumage Group, corresponding to the Shimanto Belt and exhibiting thrust faults and folds associated with subduction (Okada and Whitaker, 1979). Although the Kumage Group is not always distributed under the seafloor in the MV area, a layer of the same age as the Kumage Group was distributed below because Eocene foraminifera were obtained from the surface layer of the MV (Ujiié, 2000). The side-scan sonar survey conducted in 1990 covered an area with a width and depth of 20-30 km and 1,200-3,500 m, respectively, from the northern tip of the Ryukyu Trench slope to the southern tip of the central Ryukyu Trench slope. More than 30 mud diapirs were found on the northern trench slope (Ujiié, 2000). Currently, the presence of MV1–MV15 off Tanegashima have been confirmed ( Figure 1B ). Additionally, Nakayama et al. (2010), who investigated the pore water in MV1, suggested that water originating from the dehydration of clay minerals was supplied from the deep subseafloor (Nakayama et al., 2010). Furthermore, water masses containing high concentrations of methane (CH₄) and characteristic microbial plumes in the surface sediments were observed several tens of meters above the summits of MV1 and MV14, indicating that these MVs were still active (Hoshino et al., 2017). Moreover, Mitsutome et al. (2023) investigated the noble gases in the surface sediments of MV1 and MV14, and used their solubility at 80 to 230°C to estimate that their depth of origin was 3-9 km below the seafloor (Mitsutome et al., 2023).

Pore water was extracted from the surface sediments collected during the KH-15-2 cruise from August 19 to September 1, 2015, and the KH-19-5 cruise from August 9 to August 18, 2019 ( Table 1 ). For surface sediment sampling, we used the Navigable Sampling System (NSS; Ashi and Tokuyama, 2005) with a piston corer (length: 4-6 m) and a multiple corer (length: 60 cm) mounted on the R/V Hakuho-Maru, and randomly selected the following MVs: MV1, MV2, MV3, and MV14 ( Supplementary Figures 1A–D ). The lengths of the surface deposits obtained ranged from 3.11 to 5.21 m where piston corers were used and from 0.2 to 0.27 m where multiple corers were used ( Table 1 ). The lithofacies of all the MVs mainly comprise 2-3-cmΦ breccia with a maximum size of approximately 5 cmΦ and are covered by semi-pelagic sediments. Mud was sub-sampled from the sediment on board at approximately 25-cm intervals for samples by a piston corer (Type PC in Table 2 ) and 5-cm intervals for samples by a multiple corer (Type MC in Table 2 ).

In the case of sediments collected by the piston corer, the mud in contact with the inner tube was removed because it may have been contaminated with seawater. The collected mud was placed in a squeezer set with a mesh and filter paper (Advantec; Filter paper No. 1), and pressure was applied with a hydraulic press to extract pore water into a 50-mL syringe (Manheim, 1966). The pore water extracted into the syringe was passed through a disc filter (Advantec; disposable syringe filter unit, 0.45 μm) and divided to analyze the: 1) anion concentration, ammonium ion concentration (NH₄ ⁺), pH, and alkalinity; and 2) cation concentration. It was placed in a high-density polyethylene narrow mouth bottle (Nalgene; HDPE Bottle) and stored in a refrigerator at 4 °C. After pore water extraction, the residual sediments were frozen and transported to a land-based laboratory. In the case of sediments taken with the multiple corer, pore water was extracted with a Rhizon sampler (Seeberg-Elverfeldt et al., 2005; Dickens et al., 2007), distributed, stored, and brought back to the land-based laboratory in the same manner as pore water extracted from sediments taken with a piston corer.

The pore water samples were analyzed by ion chromatography for anions, inductively coupled plasma-atomic emission spectroscopy (ICP-AES) for cations, titration with hydrochloric acid for alkalinity, and the phenol blue colorimetric method for ammonium ions. The accuracies were less than 0.2, 3, 5, and 1%, respectively. The HCO₃ ^- concentration was calculated from the pH and alkalinity, assuming equilibrium for carbonic acid. The dissociation constants used in the calculation were K ₁ = 10^-5.847, K ₂ = 10^-8.966, and K _w = 10^-13.127, and the water temperature was 2°C.

B in the pore water was separated using a 5-cm³ PFA conical vial by the micro-sublimation method (Wang et al., 2010) and final solution was adjusted to 50 ppb B concentration in 0.3 N nitrate matrix. Measurements were conducted using a multicollector inductively coupled plasma mass spectrometer (MC-ICP-MS; Neptune Plus) manufactured by Thermo Fisher Scientific, and standard samples of NBS SRM 951 (¹¹B/¹⁰B = 4.0056 ± 0.5) were used. The accuracy of analysis was +39.39 ± 0.33‰ for five measurements of standard seawater NASS-5. Analytical errors (2 S.E.) for individual data are presented in Table 2 .

A portion of the sediment sample was dried at 60°C immediately before use, and a portion of the dried sample was ground to a powder using a mortar and pestle. Approximately 0.6 g of the powdered sample was weighed out and place into a vinyl chloride ring (Rigaku; PVC ring for powder samples, RS100-10) on the WC die pellet (Rigaku; WC die pellet, RS2550). Then WC die pellet was placed on top and pressurized at 50 MPa for 5 min using an electric pump (RIKEN; SMP-3012SK) and a press (RIKEN; CDM-10M). The pellets were stored in a desiccator until further analysis. The mortar, pestle, PVC ring, spatula, and WC die pellets were washed with acetone before and after use. The B concentration and isotope ratio analyses were conducted using MC-ICP-MS (the same machine as described above), with the laser ablation system (ESI NWR-Femto, United States). Data processing was conducted using IOLITE ver. 3.7 software (e.g., Woodhead et al., 2005; Hellstrom et al., 2008; Paton et al., 2011). GSD, supplied by the U.S. Geological Survey, was used as the standard material, and all B isotope ratios were denoted by δ as the deviation (‰) from the international standard material NIST SRM951 ( Table 3 ). Additionally, for the analysis, a laser was applied to three different locations on the same sample to calculate the average; the standard deviations (2 S.D.) were calculated and are presented with each data point ( Table 3 ).

Table 2 summarizes the chemical and isotopic compositions of the pore water and Table 3 presents the B concentrations and isotopic ratios of the sediments. Figure 2 shows the vertical distributions of Cl^-, Na, Li, magnesium (Mg), calcium (Ca), and B concentrations and isotope ratios (δ¹¹B) in the pore water, and B concentrations and isotope ratios in the sediments. The concentrations of Cl^-, Na, Mg, and Ca in all the MVs were similar to those in the seawater immediately below the seafloor and decreased with depth ( Figures 2A–E ). The B and Li concentrations increased with depth just below the seafloor ( Figures 2C, I ). At MV14, the change in the concentration was not large; however, at MV3, it was remarkable. The magnitude of the change in concentration was greatest at MV3, followed by MV2, MV1, and MV14, which gradually decreased ( Figure 2 ).

The B concentration in the sediment was almost 100 ppm, and the concentration seemed to increase with depth ( Figure 2H ). There was almost no difference in concentration between the mud volcanoes; however, the concentrations of MV14 (84 ± 14 ppm) and MV3 (101 ± 34 ppm) were slightly lower. The isotope ratio tended to increase with depth; however, it slightly varied between mountains ( Figure 2I ). The comparison between each MV showed that MV2 had the highest δ¹¹B value (+4.5 ± 2.0‰), and MV1 (–2.9 ± 2.7‰) and MV14 (–2.8 ± 1.6‰) had low δ¹¹B values ( Table 3 ). Our data (B concentration: 84–160 ppm, B isotope ratio: –2.9 to +4.5 ‰; Table 3 ) are similar to those (B = 80 to 160 ppm; δ¹¹B = –6.6 to +4.8‰) of Ishikawa and Nakamura (1993).

We investigated the chemical composition of pore water in sediments collected from the MVs off Tanegashima and discussed the origin depth, advection velocity, and MV activity. The advection velocity was estimated from the Cl⁻ concentrations, and it was found to be different for each MV. The equilibrium temperature with clay minerals was estimated from Na and Li concentrations in the fresh water source, and we found that it differed for each MV, ranging from 93 to 134°C for MV1 to MV3. The Li concentration in the pore water of MV14 is almost the same as that of seawater, and we presumed that water originating from the dehydration of clay minerals is not supplied. Considering a geothermal gradient of 25°C/km, this indicates that the origin depth is approximately 3.7 to 5.5 km, suggesting that the depth corresponds to the upper crust under the seafloor. Additionally, the B isotope ratio in pore water was very high, suggesting that adsorption had a strong effect. As MV3, which is considered the most active, showed the most adsorption, we assumed that the clay minerals that migrated from deeper depths may have a stronger adsorption capacity. The trend between the isotopic fractionation of B isotope ratios in the solid and liquid phases and the equilibrium temperature with clay minerals estimated from the Na/Li ratios in the pore water is different from the usual trend between the isotopic fractionation of silicate and water and the equilibrium temperature, which is unique to mud volcanoes. Rather than functioning as a geothermometer, the pore water of an active mud volcano may be used as an indicator of extremely high B isotope ratios.

The original contributions presented in the study are included in the article/ Supplementary Material . Further inquiries can be directed to the corresponding author.

AI, TT, and RS: Conception or design of the study. YM, KA, TT, K-HS, RS, and AI: Data collection. TT: Analysis and interpretation of data. YM, KA, TT, and RS: Writing the paper. YM, KA, TT, K-HS, RS, and AI: Approval of final draft. All authors contributed to the article and approved the submitted version.