Clay Mineral Suites in Submarine Mud Volcanoes in the Kumano Forearc Basin , Nankai Trough : Constraints on the Origin of Mud Volcano Sediments

Clay mineralogy is an important characteristic of mud volcano sediments. This study determined the clay mineral compositions of sediment from two submarine mud volcanoes in the Kumano forearc basin, Nankai Trough, by X-ray diffraction analysis. Similar compositions dominated by smectite in the two mud volcanoes indicate that the mud volcanoes in the basin are rooted in the same source sequence. These clay mineral compositions differed from those in Pleistocene basin sediment, suggesting that the mud volcano sediment originated beneath the Pleistocene sediment. The illite content in the illite–smectite mixed layer averaged 32% in the mud volcano sediment, which implies that the sediment experienced temperatures above 60 ◦C that promoted the smectite-to-illite transformation. However, porewater extracted from the mud volcano sediment had Clconcentrations roughly half that of seawater and proportional enrichment in 18O and depletion in D, indicating that dehydration reactions of clay minerals had previously occurred in a deeply buried sedimentary layer. The smectite and illite contents (<60%) in the clay-size fraction rule out in situ smectite dewatering as the cause of the dilution of Clin porewater. Thus, fluids derived from clay dewatering must have originated from deeper than the source of the mud volcano sediment.


Introduction
Submarine mud volcanoes are remarkable seafloor features that consist of mud breccia derived from sediment layers beneath the seafloor [1].They form by expulsions of mud and associated deep-sourced fluids, predominantly methane gas [2,3].The origin and composition of these fluids have been the subjects of many studies (e.g., [4][5][6][7]).These studies have shown that the fluids commonly originate from dehydration reactions of clay minerals.The fluids also contain hydrocarbon gases derived from the thermocatalytic decomposition of organic matter (e.g., [6][7][8]).These characteristics suggest that the fluids originate at depths where temperatures are high enough to promote these reactions.However, there are fewer lithological studies that include chemical and mineralogical analyses of mud volcano sediments than there are studies of mud volcano fluids (e.g., [9][10][11]), although the lithology of erupted sediment is strongly related to the formation of mud volcanoes and the source of their sediments.Given the large proportion of clay in submarine mud volcanoes, clay mineralogy is an important characteristic of mud volcano sediments, particularly signs of clay mineral dehydration that is an important proxy for the thermal history of sediments [9].This paper reports the clay mineralogy and porewater chemistry of sediment from mud volcanoes in the northern part of the Kumano forearc basin, in the Nankai Trough, and compares these results to previously reported data from sediment sampled during scientific drilling in the Nankai Trough.Although the fluid and gas chemistry at these mud volcanoes has been studied [8,[12][13][14], the clay mineralogy of the mud volcano sediments has not been investigated.

Geologic Setting and Previous Studies
The Kumano Basin, located southeast of the Kii Peninsula, is a forearc basin in the Nankai Trough, where the Philippine Sea plate is subducting beneath the Eurasia plate (Figure 1).The basin floor, at a water depth of approximately 2100 m, overlies a thick sequence (up to 2000 m) of terrigenous and hemipelagic sediment [12].Thirteen mud volcanoes have been mapped on anticlines above landward-dipping megasplay faults related to the subduction zone and have been assigned "Kumano mud volcano numbers" (KMV#) from 1 to 13 [8,[12][13][14] (Figure 1).At KMV#2, KMV#4, KMV#5, and KMV#10, the carbon isotopic compositions of methane and the methane to (ethane + propane) concentration ratios in the hydrate-bound hydrocarbons have been analyzed [8].These data suggest that the hydrocarbon gases are derived from thermal cracking of organic matter in the old accretionary prism underlying the Kumano forearc basin sediment.The oxygen and hydrogen isotopic compositions of porewater retrieved from 10 of these mud volcanoes (KMV#2-11) have suggested that the fluid in the mud volcanoes was derived from dehydration of clay minerals in the old accretionary prism [13].KMV#5 is the most intensively surveyed mud volcano in the Kumano basin.Core drilling in 2009 and 2012 by deep-sea drilling vessel Chikyu reached 200 m below the summit of KMV#5 and revealed the existence of methane hydrate in the erupted sediments [15].Although the oxygen and hydrogen isotopic compositions of porewater were affected by dissociation of methane hydrate, the inferred original isotopic composition was consistent with fluid derived from clay mineral dehydration in the old accretionary prism [13].In addition, the lithium isotopic composition of porewater extracted from the drilled sediment showed that a portion of the fluid originated in the corner of the serpentinized forearc mantle wedge [16].In contrast, nannofossils in the shallow sediment at KMV#5 retrieved by piston coring during 2001-2002 are of younger origin, dating from the late early to early middle Miocene (13.6-18.2Ma), which implies that the mud is from the bottom of the forearc basin, not from the older underlying accretionary prism [12].

Porewater Samples
Sediment samples of about 500 cm 3 (10 cm length of whole round core samples) were collected from the cores at 50-cm intervals, and porewater was extracted from them on board using a stainless steel squeezer [17].
The Cl -concentration in the porewater samples was analyzed by using an ion chromatograph (Dionex DX-500) with an isocratic carbonate/bicarbonate eluent coupled with suppressed conductivity detection and an IonPac AS12A carbonate eluent anion-exchange column (200 mm long, 4 mm i.d.; Dionex, Sunnyvale, CA, USA) equipped with a guard column (IonPac AG12A, 50 mm long, 4 mm i.d.; Dionex).Analytical precision was estimated to be within 0.3% from repeated measurements of the same samples.
For δD measurements, porewater samples were injected into an online chromium reduction analytical system attached to an isotope ratio mass spectrometer (VG SIRA10, VG Isogas Ltd., Middlewich, UK) [18].For each analysis, a 2-µL water sample was injected into a quartz furnace containing chromium metal powder at 800 °C.The external precision (1σ) was ±0.4‰.

Clay Mineral Assemblage
For X-ray diffraction (XRD) analysis, sediment samples of about 5 cm 3 were collected by a tip cut disposable syringe from the top of the same whole round core samples used for porewater extraction.The clay-size fraction (<2 µm) was separated from the sediment samples by centrifugation [19] at 1500 rpm for 1 min.The size of the clay-size fraction was determined by weighing the dried bulk sediment sample and the separated clay-size fraction; note that this is a minimum value, because the recovery rate of the clay-size fraction by centrifugation is not 100%.
The mineralogy of the clay-size fractions was determined by X-ray diffraction (XRD) analysis.Suspensions of clay were dropped onto glass slides and dried in an oven at 50 °C to prepare oriented samples.XRD patterns of these samples were obtained using an X-ray diffractometer (MacScience, MXP3HF) with CuKα radiation at 40 kV and 20 mA, with 1° divergence and antiscattering slits and a 0.15 mm receiving slit.The diffracted X-rays were counted for 2 s at steps of 0.02° in 2θ.The samples were then saturated with ethylene glycol vapor at 50 °C for 5 h and the XRD scanning was repeated.For each sample, ethylene glycol vapor saturation and XRD scanning were performed twice.

Porewater Samples
Sediment samples of about 500 cm 3 (10 cm length of whole round core samples) were collected from the cores at 50-cm intervals, and porewater was extracted from them on board using a stainless steel squeezer [17].
The Cl -concentration in the porewater samples was analyzed by using an ion chromatograph (Dionex DX-500) with an isocratic carbonate/bicarbonate eluent coupled with suppressed conductivity detection and an IonPac AS12A carbonate eluent anion-exchange column (200 mm long, 4 mm i.d.; Dionex, Sunnyvale, CA, USA) equipped with a guard column (IonPac AG12A, 50 mm long, 4 mm i.d.; Dionex).Analytical precision was estimated to be within 0.3% from repeated measurements of the same samples.
For δD measurements, porewater samples were injected into an online chromium reduction analytical system attached to an isotope ratio mass spectrometer (VG SIRA10, VG Isogas Ltd., Middlewich, UK) [18].For each analysis, a 2-µL water sample was injected into a quartz furnace containing chromium metal powder at 800 • C. The external precision (1σ) was ±0.4‰.

Clay Mineral Assemblage
For X-ray diffraction (XRD) analysis, sediment samples of about 5 cm 3 were collected by a tip cut disposable syringe from the top of the same whole round core samples used for porewater extraction.The clay-size fraction (<2 µm) was separated from the sediment samples by centrifugation [19] at 1500 rpm for 1 min.The size of the clay-size fraction was determined by weighing the dried bulk sediment sample and the separated clay-size fraction; note that this is a minimum value, because the recovery rate of the clay-size fraction by centrifugation is not 100%.
The mineralogy of the clay-size fractions was determined by X-ray diffraction (XRD) analysis.Suspensions of clay were dropped onto glass slides and dried in an oven at 50 • C to prepare oriented samples.XRD patterns of these samples were obtained using an X-ray diffractometer (MacScience, MXP3HF) with CuKα radiation at 40 kV and 20 mA, with 1 • divergence and antiscattering slits and a 0.15 mm receiving slit.The diffracted X-rays were counted for 2 s at steps of 0.02 • in 2θ.The samples were then saturated with ethylene glycol vapor at 50 • C for 5 h and the XRD scanning was repeated.For each sample, ethylene glycol vapor saturation and XRD scanning were performed twice.
The proportions of major minerals in the clay-size fraction were semiquantitatively estimated from the XRD data by measurement of peak areas and application of the calibrated weighting factors reported by Underwood et al. [20] (Supplementary Table S1).A matrix of weighting factors was applied to the integrated areas of a smectite(001) peak at 2θ ≈ 5.5 • , the illite(001) peak at 2θ ≈ 9 • , the chlorite(002) + kaolinite(001) peak at 2θ ≈ 12.5 • , and the quartz(100) peak at 2θ ≈ 20.9 • .The proportions of major clay minerals in the clay-size fraction were also semiquantitatively estimated by the methods of Biscaye [21], the most commonly used approach to calculate the relative proportions of smectite, illite, and chlorite in marine geology.However, errors in such data can be substantial, and accuracy changes significantly in response to the abundance by weight of each mineral [19,20].To avoid these errors, Underwood et al. [20] introduced a normalization factor in which the weighting factors are calculated independently for each mineral (Supplementary Table S1).However, the estimates made by the Underwood method are also semiquantitative in light of limitations imposed by differences in indigenous mineral mixtures, X-ray diffractometers (e.g., scanning mode and slits), and the individual instrument performance [20].

Porewater Chemistry
At the reference site (PC12), the Cl -concentration in porewater at all depths was essentially constant at 543-558 mM (Figure 2a).Porewater Cl -concentrations at KMV#5 and KMV#10 (PC13 and PC14, respectively) were much lower than that of seawater (ca.550 mM), decreasing with depth in the cores to minimums of 280 mM at KMV#5 and 471 mM at KMV#10 at the bottom of the cores.
Oxygen isotopic compositions (δ 18 O) of porewater increased with increasing depth at both KMV#5 and KMV#10 (Figure 2b), and hydrogen isotopic ratios (δD) decreased with increasing depth (Figure 2c).The plot of δ 18 O against Cl -concentration shows 18 O enrichment proportional to Cl -depletion.The plot of δD against Cl -concentration shows D depletion proportional to Cl -depletion (Figure 3).
The proportions of major minerals in the clay-size fraction were semiquantitatively estimated from the XRD data by measurement of peak areas and application of the calibrated weighting factors reported by Underwood et al. [20] (Supplementary Table S1).A matrix of weighting factors was applied to the integrated areas of a smectite(001) peak at 2θ ≈ 5.5°, the illite(001) peak at 2θ ≈ 9°, the chlorite(002) + kaolinite(001) peak at 2θ ≈ 12.5°, and the quartz(100) peak at 2θ ≈ 20.9°.The proportions of major clay minerals in the clay-size fraction were also semiquantitatively estimated by the methods of Biscaye [21], the most commonly used approach to calculate the relative proportions of smectite, illite, and chlorite in marine geology.However, errors in such data can be substantial, and accuracy changes significantly in response to the abundance by weight of each mineral [19,20].To avoid these errors, Underwood et al. [20] introduced a normalization factor in which the weighting factors are calculated independently for each mineral (Supplementary Table S1).However, the estimates made by the Underwood method are also semiquantitative in light of limitations imposed by differences in indigenous mineral mixtures, X-ray diffractometers (e.g., scanning mode and slits), and the individual instrument performance [20].

Porewater Chemistry
At the reference site (PC12), the Cl -concentration in porewater at all depths was essentially constant at 543-558 mM (Figure 2a).Porewater Cl -concentrations at KMV#5 and KMV#10 (PC13 and PC14, respectively) were much lower than that of seawater (ca.550 mM), decreasing with depth in the cores to minimums of 280 mM at KMV#5 and 471 mM at KMV#10 at the bottom of the cores.
Oxygen isotopic compositions (δ 18 O) of porewater increased with increasing depth at both KMV#5 and KMV#10 (Figure 2b), and hydrogen isotopic ratios (δD) decreased with increasing depth (Figure 2c).The plot of δ 18 O against Cl -concentration shows 18 O enrichment proportional to Cl - depletion.The plot of δD against Cl -concentration shows D depletion proportional to Cl -depletion (Figure 3).

Clay Mineral Assemblage
A total of 23 samples were analyzed by XRD. Figure 4 shows representative examples of X-ray diffractograms for the samples in the ethylene-glycolated state from KMV#5, KMV#10, and the reference site.
The X-ray diffractograms of the samples from KMV#5 and KMV#10 showed similar patterns.These displayed strong intensity peaks at 2θ ≈ 5.5°, representing a mixture of discrete smectite and I/S mixed-layer.The composite reflection of illite(002)/smectite(003) at 2θ ≈ 16° was detected.The peak intensities of the peaks at 2θ ≈ 9° representing illite(001), and the overlapped peak of chlorite(002) and kaolinite(001) at 2θ ≈ 12.5° were low.We could not identify the kaolinite(002) peak at 24-26° in the samples from KMV#5 and the reference site (Figure 5).In the samples from KMV#10, we identified a small contribution of kaolinite(002) in the peak at 24-26° and decomposed this peak into elementary peaks of kaolinite(002) and chlorite(004) using Profile Fit Software (Philips PROFIT v.1.0).The intensity ratios of kaolinite(002)/chlorite(004) were less than 0.07 (Table 1).These observations indicate small contents of kaolinite in the studied samples.
The peak intensities in the X-ray diffractogram of the reference-site sample were much lower than those of the mud volcano sediments (Figure 4).The smectite peak at 2θ ≈ 5.5° was subtle.
The clay-size fraction (<2 µm) made up 1.8-12.2wt % of the sediment at KMV#5 and KMV#10 and 1.6-3.5 wt % of the reference sediment (Figure 6; Table 1).In the mud volcano sediments, the clay-size fraction increased from 2% to 7% near the surface to 12% at the bottom of the cores.

Clay Mineral Assemblage
A total of 23 samples were analyzed by XRD. Figure 4 shows representative examples of X-ray diffractograms for the samples in the ethylene-glycolated state from KMV#5, KMV#10, and the reference site.
The X-ray diffractograms of the samples from KMV#5 and KMV#10 showed similar patterns.These displayed strong intensity peaks at 2θ ≈ 5.5 • , representing a mixture of discrete smectite and I/S mixed-layer.The composite reflection of illite(002)/smectite(003) at 2θ ≈ 16 • was detected.The peak intensities of the peaks at 2θ ≈ 9 • representing illite(001), and the overlapped peak of chlorite(002) and kaolinite(001) at 2θ ≈ 12.5 • were low.We could not identify the kaolinite(002) peak at 24-26 • in the samples from KMV#5 and the reference site (Figure 5).In the samples from KMV#10, we identified a small contribution of kaolinite(002) in the peak at 24-26 • and decomposed this peak into elementary peaks of kaolinite(002) and chlorite(004) using Profile Fit Software (Philips PROFIT v.1.0).The intensity ratios of kaolinite(002)/chlorite(004) were less than 0.07 (Table 1).These observations indicate small contents of kaolinite in the studied samples.
The peak intensities in the X-ray diffractogram of the reference-site sample were much lower than those of the mud volcano sediments (Figure 4).The smectite peak at 2θ ≈ 5.5 • was subtle.
The clay-size fraction (<2 µm) made up 1.8-12.2wt % of the sediment at KMV#5 and KMV#10 and 1.6-3.5 wt % of the reference sediment (Figure 6; Table 1).In the mud volcano sediments, the clay-size fraction increased from 2% to 7% near the surface to 12% at the bottom of the cores.
The illite content in the I/S mixed layer was estimated only for the mud volcano sediments, because the composite reflections of illite(001)/smectite(002) and illite(002)/smectite(003) were absent in the X-ray diffractograms for the samples from the reference site.At KMV#5 and KMV#10, the inferred illite contents in the I/S mixed layer ranged from 19% to 38% (average 32%) (Figure 7a).
Expandability was estimated only for the mud volcano sediments because the saddle/peak ratio could not be determined in the reference sediment.The expandability values increased from 61% to 70%, with increasing depth in the cores, and averaged 67% in the mud volcano sediments (Figure 7b).The Weaver's Sharpness Ratio ranged from 3.3 to 5.5 for the mud volcano sediments and from 2.8 to 3.8 for the reference-site sample (Figure 7c).
The illite content in the I/S mixed layer was estimated only for the mud volcano sediments, because the composite reflections of illite(001)/smectite(002) and illite(002)/smectite(003) were absent in the X-ray diffractograms for the samples from the reference site.At KMV#5 and KMV#10, the inferred illite contents in the I/S mixed layer ranged from 19% to 38% (average 32%) (Figure 7a).
Expandability was estimated only for the mud volcano sediments because the saddle/peak ratio could not be determined in the reference sediment.The expandability values increased from 61% to 70%, with increasing depth in the cores, and averaged 67% in the mud volcano sediments (Figure 7b).The Weaver's Sharpness Ratio ranged from 3.3 to 5.5 for the mud volcano sediments and from 2.8 to 3.8 for the reference-site sample (Figure 7c).

Discussion
The porewater samples from the Kumano mud volcanoes can be characterized by depletion of Cl -relative to the reference sediment of the Kumano basin, indicating the mixing of a freshwater component.At KMV#5, the Cl -concentration in porewater extracted from the drilled core was constant at ca. 130-150 mM from 4 m depth in the core to the core bottom (120 m) [15].In the drilled core, δ 18 O and δD values generally shifted with depth toward high and low values relative to seawater, respectively (Figure 3).Conversely, δ 18 O and δD values of water obtained from methane hydrate were higher than those of porewater.Ijiri et al. [15] estimated that the original δ 18 O and δD values of porewater before methane hydrate formation were +4.3‰-+5.3‰and −9.2‰-−16.2‰,respectively.These values were attributed to the addition of water from smectite dewatering during the smectite-illite (S-I) transformation, which produces water enriched in 18 O and depleted in D [25,26].Because our porewater samples from KMV#10 show the same trends in δ 18 O and δD values with respect to Cl -concentrations, KMV#10 also was affected by water derived from clay mineral dehydration.Indeed, this process has previously been reported for other mud volcanoes in the Kumano basin (KMV#2-11) [13].
However, there are other possible source fluids beside those from the S-I transformation.Nishio et al. [16] inferred that Li isotope ratios in porewater of the drilled core from KMV#5 indicate temperatures as high as 310 • C, which is significantly higher than the S-I conversion temperature range of 60-150 • C.They suggested that a portion of fluid was derived from the periodic injection of deep-seated fluid accumulated in the corner of the serpentinized forearc mantle wedge, transported upward along a thrust fault [16].Modeling of the water circulation in mud volcanoes of the Kumano basin has suggested that the fluids in the mud volcano sediment are derived by S-I transformation in the accretionary prism and decomposition of basaltic saponite in the subducting slab (Figure 8) [13].In sum, the fluid in the mud volcanoes of the Kumano basin appears to originate in or beneath the accretionary prism underlying the Kumano forearc basin sediment.
The sediment samples from KMV#5 and KMV#10 were enriched in the clay size fraction compared to the reference sediment from the Kumano basin.The average composition of the clay-size minerals in the mud volcano sediments was 49% smectite, 25% illite, 15% quartz, and 11% chlorite, whereas in the reference sediment the clay-size minerals averaged 30% illite, 29% quartz, 23% smectite, and 18% chlorite.Interestingly, the sediment samples from KMV#5 and KMV#10 had very similar X-ray diffractograms, clay mineral compositions, expandability, and illite contents in the I/S mixed layer despite being about 30 km apart (Figure 1).This similarity indicates that these mud volcanoes are rooted in the same source sequence.
Sediments of the Kumano basin have also been sampled at Integrated Ocean Drilling Program (IODP) Site C0002, approximately 30 km south of KMV#5 (Figure 1) [18].Because the clay-size mineral compositions at Site C0002 were estimated by the same methods as in this study [19,20], they can be compared directly with the results of this study.At Site C0002, the basin sediment shallower than 922 m below the seafloor (mbsf) contains Pleistocene nannofossils [27].The sediment below 922 mbsf was interpreted as a boundary between accreted trench turbidites and muddy slope-apron deposits [27].In sediments below 922 mbsf, this unit was assigned a late Miocene age on the basis of a nannofossil event [27].In the basin sediment above 826 mbsf at Site C0002, illite is the dominant mineral in the clay-size fraction (~35-38%), as it is at the reference site of this study (30%).Between 834 and 922 mbsf, the clay-size fraction consists of smectite (36%), illite (35%), chlorite (16%), quartz (9%), and kaolinite (4%).Over the upper 922 m of the core, the expandability of the I/S mixed layer averages 60%.In the upper accretionary prism sediments below 922 mbsf at Site C0002, smectite is the dominant mineral in the clay-size fraction (41% average, 60% maximum), followed by illite (31%), chlorite (15%), quartz (6%), and kaolinite (5%).Smectite is likewise the dominant mineral (average 49%, maximum 59%) at KMV#5 and KMV#10, and the average expandability at KMV#5 and KMV#10 (67%) is close to that at Site C0002 below 922 m (69%).These similarities suggest that the mud volcano sediment is derived from the upper accretionary prism rather than the Kumano basin.However, the mud volcano sediment and the accretionary prism sediment at Site C0002 appear to have conflicting ages.In the accretionary prism sediment of Site C0002, the maximum nannofossil age of 5.90 Ma corresponds to the late Miocene [27], whereas the nannofossils at KMV#5 are of late early to early middle Miocene age (13.6-18.2Ma) [12].The correspondence between this older age and the formation of abandoned forearc basins in the Shimanto accretionary belt on the Kii Peninsula led Morita et al. [12] to suggest that the mud volcano sediment was derived from the bottom of an old forearc basin.They reported that the deep structure of the Kumano Basin is divided into a northern zone consisting of three anticlines and a southern zone consisting of a large sedimentary basin that is bounded by a newly uplifted outer arc high.Site C0002 is in the southern zone, and the mud volcanoes of this study are in the northern zone (Figure 8).Thus, it is conceivable that the forearc basin sediment erupting in the mud volcanoes is older than the upper accretionary prism sediment at Site C0002.In sum, the evidence shows that the clay mineral composition of the mud volcano sediment differs from that of the Pleistocene basin sediments in the Kumano Basin and suggests that the mud volcano sediment originated in older sediments beneath the Pleistocene basin sediments.
Sharpness Ratios suggest that the mud volcano sediments have higher illite crystallinity, and hence higher maturity, than sediment at the reference site [24].The results are consistent with the idea that the mud volcano sediments were derived from deeply buried strata, whereas the reference site represents surface sediment in the Kumano basin.Furthermore, the presence in the mud volcano sediments of the I/S mixed layer, an indicator of diagenetic S-I transformation [28,29], suggests that this material underwent the S-I transformation.This transformation is considered to occur at temperatures above ~60 • C and to be kinetically or thermodynamically controlled [28][29][30][31][32][33].On that basis, the percentage of illite layers in the S/I mixed layer has been used as a geothermometer (e.g., [28,29]).The illite contents in the samples from KMV#5 and KMV#10 were 19-38% (average 32%).
We cannot confirm beyond doubt that the I/S mixed layer is of authigenic origin.However, at Ocean Drilling Program (ODP) Site 808, in the western part of the Nankai Trough (Figure 1), illite first appears at a depth of 555 mbsf, where the estimated temperature was ~65 • C, and increases monotonically with depth, reaching the illite contents of the I/S mixed layer in the mud volcano sediments (19-38%) at the depth range of 700-950 mbsf, where the estimated temperature range was 80-105 • C [34].If the temperature dependence of S-I diagenesis is the same in the mud volcano sediments as it is at Site 808 and the I/S mixed layer is of authigenic origin, the mud volcano sediments at KMV#5 and KMV#10 experienced temperatures above 80 • C. Note that the sediment emitted by a mud volcano might involve the sediments surrounding the diapir conduit via hydraulic fracturing through mud diapir intrusion [35].In that case, the I/S mixed layer in shallow strata, in which the illite content is low, would dilute the illite content in the I/S mixed layer of the diapiric mud volcano sediments, and thus the observed illite content of the I/S mixed layer, as well as the estimated temperature, may be minimum values.
At IODP Site C0009 in the Kumano Basin about 25 km south of KMV#5 (Figure 1), the temperature at the boundary between the basin sediments and the underlying accretionary prism (1288 mbsf) was estimated at ~40 • C [36].To have reached temperatures above 80 • C, then, the mud volcano sediment must have come from beneath the basin sediment in the accretionary prism.However, the fossil evidence from KMV#5 points instead to an origin from the bottom of the basin sediment [12].There are two possible ways to reconcile an origin in the basin with the XRD evidence suggesting that the mud volcano sediments experienced temperatures above 65 • C. One is that some factor other than temperature is different between the mud volcanoes and Site 808.Although temperature is the most important factor in the S-I transformation, the reaction is affected by many other internal and external variables [34].The other possibility is that the source of the mud volcano sediment had a higher temperature in the past.Considering that the mud volcanoes lie along anticlines related to formerly active megasplay faults [12,14], it may be that their source sediment was once more deeply buried and later raised by subduction-related uplift.
Another question concerns the degree to which the depletion of Cl -in porewater at KMV#5 and KMV#10 occurred in situ from the S-I transformation in the mud volcano sediments.This can be addressed by the following rough mass balance calculation: The Cl -concentration of the original porewater before the precipitation of methane hydrate is assumed to be ~150 mM (ca. 70% depletion) [15].To produce the maximum depletion of porewater Cl -of ca.70% in sediment with a porosity of 53% (the minimum porosity measured in cores PC13 and PC14) requires the production of ca.0.37 g of water per cubic centimeter of sediment.Smectite in marine sediment has been found to bind from 17.7 wt % to 27 wt % water [37,38], depending on environmental conditions.Assuming a value of 20% bound water on smectite, it would take 1.9 g smectite/cm 3 sediment to produce 0.37 g water/cm 3 sediment.Given the maximum bulk density (2.1 g/cm 3 ) and the minimum porosity (53%) of the sediments measured in cores PC13 and PC14, 1.9 g smectite/cm 3 sediment exceeds the dry bulk density (1.6 g/cm 3 ).Even if the original porosity of the source layer sediment was 30%, the dry bulk density is 1.8 g/cm 3 , which is still lower than the required smectite volume.Consequently, the smectite contents in the mud volcano sediment are too small to produce the observed dilution of Cl -in porewater by clay diagenesis in situ.This conclusion is consistent with the notion that the fluids in the mud volcano sediment are derived from the accretionary prism and subducting slab at depths below the source layer of the mud volcano sediment (Figure 8) [12].
Geosciences 2018, 8, x FOR PEER REVIEW 11 of 14 porewater before the precipitation of methane hydrate is assumed to be ~150 mM (ca. 70% depletion) [15].To produce the maximum depletion of porewater Cl -of ca.70% in sediment with a porosity of 53% (the minimum porosity measured in cores PC13 and PC14) requires the production of ca.0.37 g of water per cubic centimeter of sediment.Smectite in marine sediment has been found to bind from 17.7 wt % to 27 wt % water [37,38], depending on environmental conditions.Assuming a value of 20% bound water on smectite, it would take 1.9 g smectite/cm 3 sediment to produce 0.37 g water/cm 3 sediment.Given the maximum bulk density (2.1 g/cm 3 ) and the minimum porosity (53%) of the sediments measured in cores PC13 and PC14, 1.9 g smectite/cm 3 sediment exceeds the dry bulk density (1.6 g/cm 3 ).Even if the original porosity of the source layer sediment was 30%, the dry bulk density is 1.8 g/cm 3 , which is still lower than the required smectite volume.Consequently, the smectite contents in the mud volcano sediment are too small to produce the observed dilution of Cl - in porewater by clay diagenesis in situ.This conclusion is consistent with the notion that the fluids in the mud volcano sediment are derived from the accretionary prism and subducting slab at depths below the source layer of the mud volcano sediment (Figure 8) [12].

Conclusions
We determined the clay mineral compositions of sediment from submarine mud volcanoes, MV#5 and MV#10, in the Kumano forearc basin, Nankai Trough, by X-ray diffraction analysis.The sediment samples from KMV#5 and KMV#10 had very similar X-ray diffractograms, clay mineral compositions, expandability, and illite contents in the I/S mixed layer despite being about 30 km apart.The similar compositions in two mud volcanoes in the Kumano basin indicate that these mud volcanoes in the basin are rooted in the same source sequence.
These clay mineral compositions in the mud volcano sediment differed from those in Pleistocene basin sediment at Integrated Ocean Drilling Program Site C0002, in the southern part of the Kumano basin, suggesting that the mud volcano sediment originated beneath the Pleistocene sediment.Because the mud volcanoes are in the northern part of the Kumano basin where sediment deposits are older, it is conceivable that the source sequence of the mud volcano sediments is older than the Pleistocene basin sediments as well as the upper accretionary prism at IODP Site C0002 located in the southern part of the Kumano basin.The illite content in the illite-smectite mixed layer averaged 32% in the mud volcano sediment, which implies that the sediment experienced temperatures above 60 °C that promoted the smectite-to-illite transformation.However, porewater extracted from the mud volcano sediment had Cl -concentrations roughly half that of seawater and proportional enrichment in 18 O and depletion in D in porewater, indicating that dehydration reactions of clay minerals had previously occurred in a deeply buried sedimentary layer.The smectite and illite contents (<60%) of the clay-size fraction in the mud volcano sediment are too small

Conclusions
We determined the clay mineral compositions of sediment from submarine mud volcanoes, MV#5 and MV#10, in the Kumano forearc basin, Nankai Trough, by X-ray diffraction analysis.The sediment samples from KMV#5 and KMV#10 had very similar X-ray diffractograms, clay mineral compositions, expandability, and illite contents in the I/S mixed layer despite being about 30 km apart.The similar compositions in two mud volcanoes in the Kumano basin indicate that these mud volcanoes in the basin are rooted in the same source sequence.
These clay mineral compositions in the mud volcano sediment differed from those in Pleistocene basin sediment at Integrated Ocean Drilling Program Site C0002, in the southern part of the Kumano basin, suggesting that the mud volcano sediment originated beneath the Pleistocene sediment.Because the mud volcanoes are in the northern part of the Kumano basin where sediment deposits are older, it is conceivable that the source sequence of the mud volcano sediments is older than the Pleistocene basin sediments as well as the upper accretionary prism at IODP Site C0002 located in the southern part of the Kumano basin.The illite content in the illite-smectite mixed layer averaged 32% in the

Figure 2 .
Figure 2. Vertical profiles of (a) Cl -concentrations, (b) δ 18 O and (c) δD in porewater samples from this study and the drilled core from KMV#5 [15].VSMOW, Vienna Standard Mean Ocean Water.

Figure 2 .
Figure 2. Vertical profiles of (a) Cl -concentrations, (b) δ 18 O and (c) δD in porewater samples from this study and the drilled core from KMV#5 [15].VSMOW, Vienna Standard Mean Ocean Water.

Figure 3 .
Figure 3. Relationships of δ 18 O vs. Cl -(top) and δD vs. Cl -(bottom) in the samples from this study and in the drilled core from KMV#5 [15].Also shown are isotopic data from water obtained from methane hydrate dissociation.

Figure 3 .
Figure 3. Relationships of δ 18 O vs. Cl -(top) and δD vs. Cl -(bottom) in the samples from this study and in the drilled core from KMV#5 [15].Also shown are isotopic data from water obtained from methane hydrate dissociation.

Figure 6 .
Figure 6.Vertical profiles of (a) clay-size fractions and percentages of (b) smectite, (c) illite, (d) chlorite and (e) quartz from this study.

Figure 7 .
Figure 7. Vertical profiles of (a) illite in the I/S mixed layer; (b) expandability in clay-size fractions from KMV#5 and KMV#10 and (c) Weaver's Sharpness Ratio for the mud-volcano sediments and the reference-site sample.

Figure 6 .
Figure 6.Vertical profiles of (a) clay-size fractions and percentages of (b) smectite, (c) illite, (d) chlorite and (e) quartz from this study.

Figure 7 .
Figure 7. Vertical profiles of (a) illite in the I/S mixed layer; (b) expandability in clay-size fractions from KMV#5 and KMV#10 and (c) Weaver's Sharpness Ratio for the mud-volcano sediments and the reference-site sample.

Figure 6 .
Figure 6.Vertical profiles of (a) clay-size fractions and percentages of (b) smectite, (c) illite, (d) chlorite and (e) quartz from this study.

Figure 7 .
Figure 7. Vertical profiles of (a) illite in the I/S mixed layer; (b) expandability in clay-size fractions from KMV#5 and KMV#10 and (c) Weaver's Sharpness Ratio for the mud-volcano sediments and the reference-site sample.

Figure 7 .
Figure 7. Vertical profiles of (a) illite in the I/S mixed layer; (b) expandability in clay-size fractions from KMV#5 and KMV#10 and (c) Weaver's Sharpness Ratio for the mud-volcano sediments and the reference-site sample.

Figure 8 .
Figure 8. Sectional view of the Kumano forearc basin showing the origin of mud volcano sediments (source layer) and fluid migration from the old accretionary prism.Red bars indicate the locations of IODP Sites C0002 and C0009.

Figure 8 .
Figure 8. Sectional view of the Kumano forearc basin showing the origin of mud volcano sediments (source layer) and fluid migration from the old accretionary prism.Red bars indicate the locations of IODP Sites C0002 and C0009.