Layout 6 Chemical and isotopic composition of fumarolic gases at Iwate volcano, Japan, during and after seismic activity in 1998: implications for the modification of ascending volcanic gases Takeshi Ohba1, Kenji Nogami2, Jun-ichi Hirabayashi2, Takeshi Sawa2, Kohei Kazahaya3, Noritoshi Morikawa3, Michiko Ohwada3 1 Tokai University, School of Science, Department of Chemistry, Kanagawa, Japan 2 Volcanic Fluid Research Center, Tokyo Institute of Technology, Gunma, Japan 3 Geological Survey of Japan, National Institute of Advanced Industrial Science and Technology, Tsukuba, Japan ANNALS OF GEOPHYSICS, 54, 2, 2011; doi: 10.4401/ag-5182 ABSTRACT In 1998, there were many volcanic earthquakes recorded at Iwate volcano, Japan. Although an eruption was anticipated, it never occurred. Fumarolic gases were sampled at the volcano on six occasions during 1998 and 1999, and were analyzed for their chemical compositions and isotope ratios. The fumarolic gases were again sampled in 2004 and 2006, after the period of seismic activity. The HCl concentrations and isotope ratios of the H2O in the gas samples collected from an active geothermal area at the volcano were high in 1998 and 1999, but decreased significantly after 2004, irrespective of the relatively stable concentrations of CO2 and sulfur- bearing gases. A notable feature of the fumarolic gases is the high isotope ratio of H2O, similar to that of pure magmatic gas, which is typical of andesitic volcanoes, although the temperature at the outlet was less than 146°C in the present case. These features can be explained if the enthalpy of the magmatic gas was equivalent to that of water vapor at 252°C to 370°C under a pressure of 1 bar. The correlations observed among CO2/H2O and HCl/H2O, and the δ 18O values of the fumarolic gases suggest partial condensation of H2O vapor during the ascent of the volcanic gas to the surface. The estimated CO2/H2O molar ratio of the magmatic gas was 0.008, which is less than the 0.03 reported previously for magmatic gas sampled during an effusive eruption at Unzen volcano, Japan. These data suggest that the magma at Iwate volcano is depleted in volatiles, because CO2 is preferentially degassed from the silicate melt relative to H2O. Such depletion in CO2 might explain the failed eruption at Iwate volcano in 1998. 1. Introduction Fumarolic gases generally include volatiles that are emitted from degassing magma; consequently, the chemical and isotopic compositions of such gases can provide information on the degassing magma. However, the chemical and isotopic compositions of the gases can be modified during the ascent to the surface, due to, for example, contamination by ground water, condensation of water vapor, and removal of acidic components, such as HCl. In 1998, many volcanic earthquakes occurred at Iwate volcano, Japan. However, although an eruption was anticipated, it never occurred. During and after this period of volcanic seismicity, we sampled fumarolic gases at the volcano and measured their chemical and isotopic compositions. In this study, we investigate the processes that acted to modify the chemical compositions and isotope ratios of these fumarolic gases, and we estimate the CO2/H2O ratios of the magmatic gases. The modification process is quantitatively modeled assuming simple physicochemical processes, which successfully reproduce the observed variations in the chemical and isotopic compositions of these fumarolic gases. 1.1. Iwate volcano Iwate volcano is an active volcano in northern Honshu, Japan (Figure 1). The volcano has two topographic peaks and it is divided into its western and eastern parts, which are centered around Onawashiro crater (Figure 1, O) and Yakushidake crater (Figure 1, Y), respectively. Itoh and Doi [2005] investigated the development of Iwate volcano. The eastern part is younger than the western part, and its surface layer was deposited over the past 7,000 years. An eruption occurred at the outer rim of the Yakushidake crater in 1732, which produced lava flow. The eastern part contains limited geothermal activity. A weak and diffusive steam discharge occurred at point H in Figure 1, without any apparent sulfur- bearing gases. The western part is relatively old, with the most recent lava from the Onawashiro crater estimated to be older than 30 ka [Itoh and Doi 2005]. The western part is characterized by intensive geothermal activity that is Article history Received December 9, 2010; accepted May 16, 2011. Subject classification: Gases, Magmas, Volcano monitoring, Hydrothermal systems. 187 concentrated in four areas (Figure 1, J, K, U and A). At site J, in particular, there was a strong smell of sulfur-bearing gas. As the most recent volcanic activity, a steam driven explosive eruption was observed at site J in 1919. About 300,000 people live in Morioka city, which is located 20 km southeast of Iwate volcano, and a north–south express highway is located 9 km east of Iwate volcano. During an eruption at Yakushidake crater in 1686, houses near Iwate volcano were affected by ash deposits and a lahar. Any future eruption at Iwate volcano poses serious risk to nearby residents and infrastructure. Iwate volcano is monitored continuously by the Japan Meteorological Agency, in terms of its seismicity and crustal deformation (tiltmeter and GPS), but its fumarolic gases are not monitored. In terms of the geochemistry of the fumarolic gases at Iwate volcano, only Noguchi et al. [1961] have reported on the chemical composition of the gas and sublimate within the Yakushidake crater. Miura et al. [2000] reported deformation of the volcanic body at Iwate volcano in 1998, which was attributed to the intrusion of magma at a depth of 3 km beneath the western part. From May to September 1998, many volcanic earthquakes were recorded beneath the region covered by points O, J, K and U in Figure 1 [Miura et al. 2000]. The earthquake hypocenters were concentrated in a zone that trended E–W. Tanaka et al. [2002] reported a region of anomalously high seismic velocity beneath the Onawashiro crater (Figure 1), which was attributed to a body of solidified magma. A similar model was proposed by Aizawa et al. [2009] based on a magnetotelluric survey. 2. Sampling and analysis of gas and water Most of the fumarolic gases at sites H, J, K, U and A (Figure 1) were sampled following the method proposed by Ozawa [1968], using a glass syringe in which 20 ml of 5 M KOH solution was filled, to absorb acidic gases. This method results in only a small amount of air contamination in the syringe. The KOH solution obtained was subjected to wet chemical analysis, according to Ozawa [1968]. The gas phase collected in the syringe, which included the «residual» composition of the gas (e.g., N2, O2, H2, CH4, Ar, He and Ne), was analyzed using two gas chromatographs (GCs) with Ar and O2 as the carrier gases, through which the components other than Ar and O2 were analyzed. The GC data were then combined to yield the complete composition of the residual gas. Fumarolic gases at site J were sampled in 2004 and 2006 for He isotope analysis, using an evacuated glass bottle [Giggenbach and Goguel 1989] that included 20 ml 5 M KOH, to prevent air contamination. At site J, hot spring water was also sampled in 1998 and 1999. Two fumaroles at site J were selected for repeated sampling. One series of samples (Table 1, 188 OHBA ET AL. Figure 1. Topographic map of Iwate volcano (contour interval, 100 m; altitude, meters above sea level). Inset, location of the volcano. Dashed curves O and Y, outer rims of the Onawashiro and Yakushidake craters, respectively. Shaded areas Lo and Ly, distribution of the lava flows from the Onawashiro and Yakushidake craters, respectively. Letters H (Higashi iwate), J (Ojigokudani), K (Kurokura), U (Ubakura), and A (Amihari), locations of geothermal areas from which the fumarolic gases were collected. At site J (39°51¢05ʺN, 140°58¢21ʺE), hot spring water was also collected. The area with altered soil at site J measured about 100 m (north–south) by 75 m (east–west). Intense fumarolic activity was noted at J, including a strong smell of sulfur-bearing gases (also detected at site A). At sites H, K and U, in contrast, sulfur-bearing gases were not smelt when sampling the fumarolic gases. 189 FUMAROLIC GASES AT IWATE VOLCANO, JAPAN L o ca ti o n C o d e D at e T em p °C H 2O C O 2 H 2S SO 2 H C l H e H 2 O 2 N 2 C H 4 A r δ D SM O W ‰ δ 18 O SM O W ‰ δ 13 C P D B ‰ 3 H e/ 4 H e R / R a 4 H e/ 20 N e J JA 1 19 98 /6 /2 13 5 98 02 00 13 60 0 52 40 26 8 95 8 0. 03 2 94 ** 10 7 0. 32 0. 29 -3 1 3. 2 -3 .2 J JB 1 19 98 /6 /2 11 0 98 41 00 11 90 0 32 10 16 6 61 2 0. 02 8 1. 2 ** 82 .7 0. 08 1 0. 25 4 -2 7 3. 1 -3 .5 J JA 2 19 98 /7 /2 4 13 4 98 65 00 92 80 34 10 35 7 11 70 0. 01 5 2. 3 ** 45 .8 0. 29 0. 17 -2 6 4. 4 -3 .3 J JB 2 19 98 /7 /2 4 10 6 98 69 00 95 40 23 50 27 9 92 6 0. 01 9 0. 41 ** 52 .8 0. 05 2 0. 21 -2 5 4. 4 -3 .2 J JA 3 19 98 /9 /1 0 13 5. 6 98 49 00 96 20 38 00 54 5 22 80 0. 01 4 2. 5 ** 52 .2 0. 02 5 0. 18 -2 2 5. 4 -3 .0 J JB 3 19 98 /9 /1 0 10 9 98 62 00 10 50 0 26 10 46 5 44 7 0. 02 3 2. 4 ** 67 .8 0. 05 7 0. 26 -2 5 4. 4 -3 .0 J JA 4 19 99 /5 /1 2 14 3 98 33 00 84 20 31 50 74 3 39 90 * 34 5 ** 89 .0 0. 37 * -2 2 5. 2 -3 .5 J JB 4 19 99 /5 /1 2 10 2. 6 98 79 00 91 70 20 00 52 0 39 0 0. 01 5 0. 61 ** 51 .6 0. 05 5 0. 34 -2 5 4. 8 -3 .2 J JA 5 19 99 /8 /1 7 13 9 98 17 00 10 11 0 35 70 96 6 34 90 * 67 .2 ** 62 .3 0. 17 * -2 4 4. 8 J JB 5 19 99 /8 /1 7 98 98 85 00 88 00 19 30 40 6 29 8 0. 00 65 0. 27 ** 25 .1 0. 02 3 0. 14 -2 2 4. 8 J JA 6 19 99 /1 0/ 14 14 6 98 15 00 11 00 0 50 00 12 40 11 10 * 72 .3 ** 79 .7 0. 24 0. 43 -3 0 3. 4 J JB 6 19 99 /1 0/ 14 10 1 98 63 00 10 40 0 27 30 55 2 u d 0. 02 3 1. 03 ** 64 .9 0. 11 0. 23 -2 9 3. 5 J JA 7 20 04 /9 /2 2 11 1. 3 97 91 00 15 60 0 41 10 50 8 66 3 0. 00 59 1. 1 1. 8 35 .3 0. 25 0. 16 -3 4 0. 0 -3 .9 5. 01 46 .4 J JB 7 20 04 /9 /2 2 96 .3 96 72 00 28 90 0 37 20 91 u d 0. 00 66 0. 8 1. 4 40 .1 0. 29 0. 22 -4 6 -2 .7 -3 .3 4. 91 31 3 J JA 8 20 06 /8 /1 11 6 25 41 00 14 20 0 58 30 17 30 u d u d 4. 0 12 70 00 59 00 00 u d 68 00 -5 3 -3 .5 J JB 8 20 06 /8 /1 96 .6 98 41 00 11 20 0 44 60 10 5 u d 0. 02 1 1. 6 u d 74 .7 0. 00 01 2 1. 09 -8 1 -1 0. 3 -2 .7 4. 82 31 .6 U U 1 19 98 /9 /1 0 94 .6 87 07 00 12 93 00 -1 48 -2 6. 0 -3 .5 U U 2 20 06 /8 /2 95 .6 47 51 00 13 30 0 3. 9 10 10 00 40 60 00 u d 46 60 -1 35 -2 5. 5 -4 .9 1. 08 0. 34 K K 20 06 /8 /2 92 .4 84 51 00 44 00 u d 29 50 0 12 00 00 u d 13 90 -1 21 -2 0. 8 -4 .2 1. 12 0. 33 A A 20 05 /9 /5 12 1. 3 98 81 00 78 50 36 70 28 .0 u d * 95 21 20 6 50 .7 1. 70 -6 1 -4 .3 -4 .4 4. 32 82 .9 H H 1 20 05 /9 /1 83 .8 31 20 00 66 80 3. 8 13 70 00 53 80 00 u d 62 70 -1 12 -1 7. 1 -7 .6 1. 04 0. 34 H H 2 20 05 /9 /1 83 .6 49 13 00 13 90 0 9. 8 97 80 0 39 20 00 u d 45 90 -1 39 -2 2. 0 -5 .9 1. 11 0. 33 Y Y 1 19 60 /9 /2 9 14 4 98 30 00 14 60 0 10 90 81 7 37 6 Y Y 2 19 60 /9 /2 9 18 3 98 20 00 18 40 0 11 70 15 20 18 5 T ab le 1 .C he m ic al a nd is ot op ic c om po si ti on s of t he f u m ar ol ic g as es a t Iw at e vo lc an o. C he m ic al c on ce nt ra ti on s as μ m ol /m ol . u d, b el ow d et ec ti on li m it . U nt il O ct ob er 1 99 9, t he g as es w er e sa m pl ed u si ng a g la ss s yr in ge a s th is m et ho d w as e as y to a pp ly a nd s u ffi ci en tl y ac cu ra te in t er m s of t he m aj or c he m ic al c om po ne nt s, e xc ep t H e an d A r. A ft er 2 00 4, t he g as es w er e sa m pl ed u si ng a n ev ac u at ed g la ss b ot tl e to m ea su re t he H e is ot op e ra ti o. * H e co nc en tr at io n no t m ea su re d du e to in te rf er en ce fr om a la rg e pe ak o f H 2 in th e ga s ch ro m at og ra ph . * *O 2 su bt ra ct ed fr om th e ra w d at a as su m in g th e O 2 de te ct ed w as c on ta m in at ed b y ai r tr ap pe d in th e gl as s sy ri ng e u se d fo r sa m pl in g. S u bt ra ct io n of th e at m os ph er ic co m po ne nt w as p er fo rm ed fo r N 2 an d A r, b as ed o n th e co nc en tr at io n of O 2. E m pt y ce lls , n o an al ys is . R , 3 H e/ 4 H e m ol ar r at io . R a, R v al u e of n or m al a ir a s 1. 4 × 10 –6 . T he c he m ic al c om po si ti on s of t he g as es w it hi n th e Ya ku sh id ak e cr at er ( Y 1, Y 2) a re fr om N og u ch i e t al . [ 19 61 ]. JA1–JA8) was collected at the fumarole with the highest outlet temperature. A second series (Table 1, JB1–JB8) was sampled at a fumarole located 32 m from the first. In addition, hot spring waters were sampled at site J for analysis of their anionic compositions by ion chromatography. To measure the δD values of H2O in the fumarolic gases, we sampled the gas condensates. Part of the condensates was converted to H2 gas with metallic Zn [Coleman et al. 1982] for analysis using mass spectrometry (Finnigan MAT 252). The same procedure was used for measuring the δD value of H2O in hot spring water. The δ18O values of H2O in the condensate and water sample were measured using the conventional CO2-H2O equilibration method [Epstein and Mayeda 1953], using mass spectrometry. To measure the δ13C value of CO2 in the fumarolic gases, the CO2 in the gases was fixed as a BaCO3 precipitate at the sampling site [Kusakabe 2001]. The precipitate was converted to CO2 in the laboratory for analysis by mass spectrometry. The analytical errors were 1.0‰, 0.1‰, and 0.1‰ for δD, δ18O and δ13C, respectively. The gas collected in the headspace of the evacuated bottle was analyzed noble mass spectrometer (MM5400), to obtain the rare gas composition and 3He/4He ratio. The analytical error of the 3He/4He measurements was less than ±0.1 Ra. Analyses of chemical and stable isotopes were performed at the Kusatsu–Shirane Volcano Observatory at Gunma Prefecture, Japan, and those of He isotopes were performed at the Geological Survey of Japan at Tsukuba city, Japan. 3. Results 3.1. Major components of fumarolic gases Tables 1 and 2 list the chemical and isotopic data for the fumarolic gases and hot spring waters, respectively. The major components of the fumarolic gases are H2O, CO2, H2S, SO2 and HCl. Figure 2 shows a ternary diagram of the H2O, CO2 and S concentrations in the fumarolic gases (where S = H2S + + SO2). Except for the samples JB7 and JA8, the gases at site J are clustered in the center of the diagram, along with the sole sample from site A. The gases at sites U, K and H are depleted in S; as such, they lie along the right-hand edge of the ternary diagram in Figure 2. The CO2/H2O ratios of samples from sites H and U are distinctly (although only slightly) higher than those of the cluster (except for sample U1). Sample U1 is strongly depleted in H2O, whereas the sample from site H is enriched in H2O. Gases at Yakushidake crater, as sampled in 1960, were enriched in S, similar to the composition of gases at site J. 3.2. Temporal variations in the compositions of the fumarolic gases at site J Figure 3 shows the temporal variations in the chemical compositions and isotope ratios of the gases from site J. The CO2 and H2S concentrations were relatively low in 1999, and they showed slight increases in 2004 and 2006. The SO2 concentrations in the JA samples increased from June 1988 to October 1999, followed by a decrease in 2004 and an increase in 2006. Except for 2006, the temporal variations observed for the JB samples are similar to those for the JA samples. The HCl concentrations in the JA samples show an increase from June 1998 to May 1999, followed by a decrease. The HCl concentrations in the JB samples were relatively high in July 1998, but decreased gradually thereafter. The H2 concentrations in the JA samples were high in May 1999, moderate in late 1999, and low in 2004 and 2006. The CH4 concentrations in the JA and JB samples show no clear temporal trends. There is no correlation between the values of the JA and JB series. The δD and δ18O values were high until August 1999, but decreased thereafter. OHBA ET AL. 190 Code Date Temp °C pH δDSMOW ‰ δ18OSMOW ‰ Cl– mg/liter SO2–4 mg/liter W1 1998/7/24 75.7 2.4 -69 -10.4 290 825 W2 1998/7/24 30.6 2.5 -73 -11.5 26 1820 W3 1998/7/24 70.1 1.0 -21 4.9 9300 4050 W4 1998/7/24 95.3 2.1 -64 -8.9 711 2520 W5 1998/9/10 94.8 1.2 -31 0.2 4309 2230 W6 1998/9/10 78.2 2.3 -68 -9.8 472 800 W7 1998/9/10 96.7 2.0 -53 -6.2 489 2211 W8 1998/9/10 90.9 2.1 -64 -9.3 442 730 W9 1999/5/12 74.2 1.9 -73 -10.9 14 1700 W10 1999/5/12 94.4 2.0 -60 -7.5 160 997 W11 1999/5/12 95.1 0.9 -18 4.8 8330 4790 Table 2. Chemical and isotopic compositions of the hot spring waters sampled at site J. Figure 2. H2O-CO2-S ternary diagram (S = H2S + SO2) showing the compositions of the fumarolic gases (molar basis). 191 3.3. Isotope ratios of H2O in the gases and water In 1999, the δD and δ18O values of H2O in the gases at site J were high, similar to those of magmatic water (typical of andesitic volcanoes) [Taran et al. 1989], as shown in Figure 4. The gases with high isotope ratios contained high concentrations of HCl (e.g., JA4). In Figure 4, the gases from site J are distributed along the line connecting the general magmatic water (Figure 4, square at top right) and local meteoric water, as reported by Kazahaya et al. [2000], who analyzed seven water samples from altitudes of 1,300 m to 1,800 m on the Iwate volcano. The average δD and δ18O values of the seven samples were -73‰ and -11.5‰, respectively. Some of the hot spring waters (Table 2, W3 and W11) sampled at site J yielded high isotope ratios, close to those of magmatic water. The concentration of Cl– in samples W3 and W11 was 9,300 mg/L and 8,330 mg/L, respectively, which is equivalent to 4,720 mmol/mol and 4,230 mmol/mol as the Cl/H2O molar ratio (similar to the HCl/H2O molar ratio for sample JA4: 4,060 mmol/mol). In Figure 4, the isotope ratios of the hot spring waters are also distributed along the line connecting the magmatic vapor (MV) and the local meteoric water (LMW). The δD and δ18O values of H2O in the gases from sites U, K and H are much lower than those of the LMW, and the isotope ratios of the gases are much lower than those of the vapors equilibrated with the LMW at 100 °C. 3.4. Residual gas composition Figure 5 shows the compositional data for the fumarolic gases at site J plotted on a N2-He-Ar ternary diagram. The data FUMAROLIC GASES AT IWATE VOLCANO, JAPAN Figure 3. Temporal variations in the compositions (μmol/mol) and isotope ratios (‰ relative to SMOW) of the water in the fumarolic gases from site J (as indicated). Solid and dotted lines, JA and JB series, respectively. Figure 4. The δD and δ18O values of water in the fumarolic gases and hot spring waters. Square, upper right, general range of the values for the typical MV of an andesitic volcano above a subduction zone. Elliptical region, local meteoric water (LMW) at the summit area of Iwate volcano [Kazahaya et al. 2000]. Small filled dots, isotope ratios of the hot spring water. Dashed horizontal and vertical lines, bottom left, δD and δ18O, respectively, of the water vapor equilibrated with the LMW at 100 °C. points are distributed in the region of mixing between the end member with a high N2/Ar ratio and the atmospheric end members. The end member with a high N2/Ar ratio can be represented by the gas of the Circum-Pacific volcanoes (Figure 5, CPVs) [Giggenbach 1997]. The 3He/4He ratios in the fumarolic gases are plotted in Figure 6 ( JA7, JB7, JB8 and A), where the 3He/4He ratio is normalized by the ratio of air (= R/Ra). The R/Ra and 4He/20Ne ratios of U, K and H are almost identical to the atmospheric values. The data points for JA7, JB7 and JB8 plot along the mixing line, the end members of which are air and a component with a high 4He/20Ne ratio. The R/Ra ratio of the end member is estimated to have a value of 5, which is slightly lower than the value of 6 reported for He in magmatic gas from the Japanese islands [Poreda and Craig 1989]. Fumarolic gas from site A lies beneath the mixing line in Figure 6, suggesting an additional contribution of a 4He-enriched component that probably originated in the crustal rock. Figure 7 shows the relationships between the δ13C value of CO2 and the CO2/ 3He molar ratios, indicating that the origin of the CO2 is the fumarolic gases [e.g., Marty et al. 1989, Sano and Marty 1995]. CO2 in volcanic gases can originate from the upper mantle (e.g., volatiles in the mid- oceanic ridge basalt; Figure 7, MORB), the sea-floor sediment (Figure 7, S), or limestone (Figure 7, L). The CO2 in fumarolic gases at site J originated mainly from limestone, with a minor contribution of MORB-type CO2. 4. Discussion 4.1. Apparent equilibrium temperature The apparent equilibrium temperature (AET) is an index of volcanic activity [Giggenbach 1987], and it can be regarded as the quenched temperature of volcanic gas. As H2 gas is the most reactive component in fumarolic gases, and because H2 included in the following reaction has a stoichiometric coefficient of 3, the AET for the reaction: (1) is expected to be a proxy of gas temperature in the conduit connecting the gas source and the outlet of the fumarole. The temperature of the gas conduit is initially high, as it is similar to the temperature of the magma at depth, but it decreases with the ascent of the gas. At high temperatures, the equilibrium of Equation (1) shifts to the left side of the reaction. During the ascent of the gas within the conduit, the temperature decreases, resulting in a shift in the 3 2SO H H S H O2 3 2 2+ = + OHBA ET AL. 192 Figure 5. N2-He-Ar ternary diagram showing the compositions of the fumarolic gases at site J (molar basis). Circum-Pacific volcanoes (CPV) and M, end members typical of the CPV gases and gases originating in the mantle, respectively. ASW , composition of the water saturated with air. The data are plotted from Table 1. Figure 6. 3He/4He versus 4He/20Ne ratios of the fumarolic gases. R and Ra, 3He/4He ratios of fumarolic gas and air (1.4 × 10–6), respectively. Dashed curve, mixing line between air and an end member with an R/Ra of 5 and a 4He/20Ne of 1000. Figure 7. CO2/ 3He ratios and δ13C values of CO2 within the fumarolic gases from site J. L, S, end members typical of oceanic sediment (S) and oceanic limestone (L), respectively. Dashed curves, mixing between S and MORB and between L and MORB. 193 equilibrium of Equation (1) to the right side. However, the reaction takes some time to reach completion, and if the gas ascends rapidly, the shift in the reaction is incomplete, resulting in a high AET. In this case, the increase in the AET is interpreted as reflecting a rapid ascent of the gas. AET was calculated for the composition of the JA and JB series, as follows: (2) where LQ is the logarithm of the quotient, (3) where Ci is the concentration of gas species i. Equation (2) was obtained from the temperature dependence of the standard equilibrium constant of formation, as compiled by Giggenbach [1987]. The total pressure of gas is assumed to be the saturated pressure of water vapor at the calculated AET (see Table 3). Figure 8 shows the temporal variations in the AET and the outlet temperature of the gas. The AET of the JA series showed a significant increase in 1999, over 1998. The outlet temperature of JA increased slightly in 1999 compared with 1998. Miura et al. [2000] reported that the seismic activity beneath Onawashiro crater was high until September 1998, and decreased thereafter. The increase in AET for the JA series lags behind the period of active seismicity. The AET values for the JB series show no clear temporal trends. The short distance between fumaroles JA and JB indicates that the processes governing the temporal variations in the AET of JA was active at shallow depths. 4.2. Magmatic vapor at site J The H2O isotope ratios of the JA and JB series fall along a line connecting the point for the LMW and the water vapor with a magmatic origin (Figure 4, MV). The δD and δ18O values of the MV are similar to those for the JA and JB series; consequently, the line connecting the LMW and the MV passes through the cluster of data points near the MV. The δ18O value of the MV was tentatively assumed to be +7‰ relative to the SMOW, which is in the range of typical values for andesitic volcanoes above a subduction zone [Giggenbach 1992]. The δ18O value of the MV is limited to less than +8.5‰, because the upper limit for δD of water in andesitic magma is -10‰ [Giggenbach 1992]. The δ18O value of JA3 is +5.4‰, close to that of the MV. JA3 contains high concentrations of HCl. If liquid water encounters a gas similar to that of sample JA3, most of the HCl in the gas would be absorbed by the liquid phase, making the vapor phase depleted in HCl. During September 1998 at Iwate volcano, there was relatively little contamination by water, and consequently there was a higher HCl content in the gas phase. After contamination, no liquid phase remains because it absorbs HCl in the gas phase. For the above case, the temperature of the MV was estimated according to the following equations: (4) ( ) 1.242 273.15 8924 AET °C LQ = - - LQ log C C C C C C1 3 H O H S H O SO H O H 2 2 2 2 2 2 = - - a a ak k k% / f f1F MV LMWd d d= + -^ h FUMAROLIC GASES AT IWATE VOLCANO, JAPAN Code Temp (outlet) °C AET °C JA1 135 464 JB1 110 229 JA2 134 263 JB2 106 201 JA3 136 271 JB3 109 273 JA4 143 642 JB4 103 223 JA5 139 482 JB5 98 193 JA6 146 486 JB6 101 240 JA7 111 236 JB7 96 206 JA8 116 379 JB8 97 230 Table 3. Apparent equilibrium temperatures (AETs) for the chemical compositions of JA and JB. Figure 8. Temporal variations in the apparent equilibrium temperature (AET) calculated for the chemical compositions of the JA (solid lines) and JB (dashed lines) series (upper), and for the outlet temperatures of the gases (lower). (5) where δ is the δ18O value of the gas and the water; the subscripts F, MV and LMW indicate the fumarolic gas, MV and LMW, respectively; f is the molar fraction of H2O derived from the MV; H is the enthalpy of gas and water; δMV is assumed to be +7‰ or +8.5‰; δLMW is assumed to be -11.5‰; and δF gives f according to Equation (4). The f obtained is substituted into Equation (5). HF was assumed to be the enthalpy of saturated H2O vapor at the temperature of the fumarolic gas. HLMW was 42 kJ/kg, which is the enthalpy of liquid water at 10 °C. Solving Equation (5), HMV can be obtained. The temperature of the MV was estimated assuming that the MV is a super-heated water vapor under a pressure of 1 bar (the gas pressure beneath the surface is expected to exceed 1 bar). The temperature of the MV obtained is the potential temperature at the surface when the volcanic gas expands, with the enthalpy of water vapor kept constant. The temperature of the MV depends on δF and on the temperature of the fumarolic gas, although the effects of temperature are small relative to those of δF. The calculated temperature is shown in Figure 9. The temperature is lowest for JA3, for which the δF value is highest. As discussed below, the JA series is affected by the partial condensation of H2O vapor, resulting in reduced δF values. The calculated temperature of the MV at JA3 is 252 °C and 370 °C for the δMV values of +7‰ and +8.5‰, respectively. These temperatures are much lower than the general temperatures of degassing magma. It is therefore necessary to invoke a cooling process during the ascent of the MV from the source to the surface. 4.3. Correlation between δ18O values of H2O and CO2/H2O The correlation between the chemical and isotopic compositions of fumarolic gases can be explained by a simple model [Ohwada et al. 2003, Ohba 2007, Ohba et al. 2010] in the case of the Tatun volcanoes (Taiwan), the Izu- Oshima volcano ( Japan) and the Kusatsu–Shirane volcano ( Japan), for which the fumarolic gas is the vapor phase that is generated between high-temperature magmatic gas and low-temperature groundwater after mixing. As shown in the following equations, a simple mixing between fluids is the basis of the calculations of the composition of the vapor phase, assuming the conservation of total enthalpy with the amount of substance, and the partitioning equilibrium between the vapor and liquid phases. (6) (7) (8) (9) (10) where C is the molar concentration of CO2 or HCl; the subscripts PS and PW indicate the vapor phases (primary steam; PS) and the co-existing liquid phase (primary thermal water; PW), respectively, generated from the mixing of the MV and the LMW; and f and g denote the mixing fraction of the MV and the generating fraction of the PS, respectively. In general, g is not equal to f. Equations (6), (7) and (8) describe the conservation of enthalpy, the isotope ratio, and the chemical composition (CO2/H2O or Cl/H2O ratio) for the generation of the PS and the PW, respectively. Equations (9) and (10) describe the equilibrium between PS and PW in terms of stable isotopes, CO2, or the distribution of HCl. α is the isotope fractionation factor of H2O [Horita and Wesolowski 1994], and b is the distribution coefficient of CO2 [Giggenbach 1980]. The partitioning of HCl between the vapor phase and the liquid phase is complicated relative to that of CO2, because HCl is ionized into H+ and Cl- in the liquid phase. The calculated composition of PS is shown in Figure 10, along with data points for the fumarolic gases. The X axis in Figure 10 is the scaled fraction (ScF) of CO2 in the CO2-H2O binary system, defined by: . (11) The ScF is convenient in terms of plotting a wide range of CO2/H2O ratios on a single calibration axis. Given the degree of freedom in f, the composition of PS is expressed as a curve. Smaller values of f give lower δ18O values on the PS curve. The PS curve originates from the mixing curve C C Clor CO PW PS 2b = 1000 1000 PS PWa d d = + + f f g gC C C C1 1LMW PWMV PS+ - = + -^ ^h h f f g g1 1LMW PWMV PSd d d d+ - = + -^ ^h h 1fH H H Hf g g1LMW PWMV PS+ - + -=^ ^h h ( ) 100 ScF CO C C C100 2 CO H O CO 2 2 2 = + f f1H H HF MV LMW= + -^ h OHBA ET AL. 194 Figure 9. Estimated temperatures of the MV. Filled, open circles, estimates assuming δ18O for magmatic vapor of +7‰ and +8.5‰, respectively. 195 between the MV and the LMW, and its location depends on the enthalpy of MV. In Figure 10, the enthalpy of MV was assumed to be equivalent to that of vapor heated to 400 °C at 1 bar. In constructing the PS curve, the CO2/H2O ratio of the MV was assumed to be 0.008, which was chosen to explain the composition of the gas at site A. The correlation between δ18O and CO2/H2O in gases at site J is not compatible with the composition of PS. As shown above, the gas at site J with a high δ18O value could be a simple mixture of the MV and the LMW. The generation of a liquid phase is unlikely. because of the high HCl concentration in the gas at site J. Figure 10 shows a wide range of CO2/H2O ratios. Gases with a high CO2/H2O ratio have a lower δ 18O value, which can be explained if the gas undergoes partial condensation of the water vapor, modeled by the Rayleigh process. The composition of the gas phase after the partial condensation of the water vapor is expressed as follows: (12) (13) where F is the fraction of water vapor remaining after the condensation of the water vapor, r is the CO2/H2O ratio of the gas phase (molar ratio), and the subscript i represents the initial value of the gas before condensation. The gases from site J lie around the trend of condensation at 100 °C or 120 °C, except JB8 (Figure 10). The gas samples from U, K and H lie far from the PS curve in Figure 10. The δ18O values of the gases are much lower than that of PS. Even if partial condensation is assumed for PS, the δ18O values would still be too high; however, the δ18O value can be reduced by introducing a vapor phase that originated from the LMW. If groundwater is heated in the shallow crust within a geothermal area, some of the water would be vaporized. The isotope ratio of the vapor generated (the local meteoric vapor; LMV) is shown on the Y-axis in Figure 10. Given the freedom of the fraction of vaporized water, the δ18O value of the LMV is expressed as a line. The addition of the LMV to a gas phase (e.g., the gas at A) produces a gas phase with a low CO2/H2O ratio, as shown by the shaded region in Figure 10. If such a gas undergoes partial condensation, the composition of the remaining gas phase would approach the composition of the gas from sites U, K and H. 4.4. Correlation between δ18O values of H2O and HCl/H2O The HCl concentrations in the vapor phase (after partial condensation) can also be calculated under the assumption of a Rayleigh process. Two different cases are assumed in terms of condensation. In the first case, the condensed liquid phase is a pure HCl solution. The concentration of H+ in the condensed liquid is equal to that of the Cl- ions. In this case, the HCl concentration in the vapor phase is given by: (14) where m is the molarity of HCl in the vapor phase, and K is the partitioning factor of HCl between the vapor and liquid phases, as defined by: r rFi 1 1 = - b a k F1000 1000i 1d d= + -a-^ ^h h Km F F m k 1 1 1 i 2 = - +c m FUMAROLIC GASES AT IWATE VOLCANO, JAPAN Figure 10. Correlation between the δ18O values of H2O and the CO2/H2O ratios in the fumarolic gases. Bold dashed curve, mixing line between the MV and the LMW. The LMV represents the vapor generated from the LMW at 100 °C. Shaded region, mixture between the LMV and a gas similar to that from site A. Solid curve (labeled PS), composition of primary steam. Thin dashed curves and solid curves, trends of the partial condensation of the water vapor phase, for which F is the fraction of water vapor remaining after partial condensation. Figure 11. Correlation between the δ18O values of H2O and the HCl/H2O ratios in the fumarolic gases from site J. Bold horizontal line, trend during the partial condensation of the vapor, where the condensed liquid contains H+ and Cl– ions with a common molarity. Thin dashed curves, trends of partial condensation, for which the pH of the condensed liquid phase is fixed at –0.3 or –0.9. Bold dashed curve, mixing line between the MV and the LMW. (15) where mH+ and mCl– are the molarities of the H + and Cl- ions in the condensed liquid phase, respectively. The temperature dependence of K was given by Simonson and Palmer (1993). The temperature during condensation was assumed to be 120 °C. The changes in the δ18O value of H2O and in the HCl/H2O ratio of the vapor phase are shown as a sub- horizontal line in Figure 11, in which the X-axis is the ScF of HCl in the HCl-H2O binary system, which is defined by: (16) HCl in the vapor phase is effectively removed by the condensed liquid phase. Most of the HCl was lost by the time that 20% of the water vapor had condensed. The above compositional trend does not always explain the distribution of data points for the gases from site J. We assume the following condition for the second case of condensation. Within a conduit filled with volcanic gas, a condensed liquid phase is likely to be dominant. The liquid phase is likely to be strongly acidic due to the absorption of HCl from the vapor phase; consequently, the removal of HCl from the vapor phase is suppressed because K is the equilibrium constant of the following reaction: . (17) Two cases of condensation were assumed, where the pH of the coexisting liquid phase is -0.3 or -0.9. When the pH of the condensed liquid phase is fixed, the change in the HCl/H2O ratio of the vapor phase is given by: (18) where, . (19) The trend in the composition with progressive condensation is given in Figure 11. Most of the data points for the gases from site J are distributed within the two curves. The composition of the hot spring waters at site J is also shown. Most of the data points for the hot spring water are distributed along the direct mixing curve between the MV and the LMW. 5. Conclusions Based on the composition of the gases from site J, the AET reached a peak in May 1999. This period with high AET followed the period of intense earthquakes, which suggests that magma movement was followed by magma degassing, which may have increased the flux and speed of the gas ascent through the conduit to the surface. The temperature of magmatic vapor at site J was estimated at 252 °C to 370 °C, based on the enthalpy and isotope ratio of the gas. During the ascent, contamination by groundwater was limited, which suggests the existence of a well- developed inner wall of the gas conduit, which prevents the intrusion of groundwater. During the ascent of the magmatic gas to the surface, the enthalpy of the gas would have been lost by conduction through the wall of the conduit. The magmatic gas would have ascended a large distance from the deep source to the surface, which is consistent with the findings of a magnetotelluric survey [Aizawa et al. 2009] that revealed an electrically resistive body beneath the geothermal area at site J, at depths of 0.5 km to 3.0 km. The resistive body was interpreted to be a solidified intrusive magma. The gas was probably transported through the solidified magma. The correlations among the δ18O values of H2O, CO2/H2O and HCl/H2O suggest that the gases at site J experienced partial condensation of the water vapor. During condensation, a strongly acidic liquid phase would have coexisted with the vapor phase, preventing the absorption of HCl in the vapor phase by the liquid phase. The strongly acidic liquid phase might have originated from the condensed liquid phase. The CO2/H2O ratio of the magmatic gas was estimated as 0.008, which is much lower than the 0.028 observed for magmatic gas sampled from the erupting Unzen volcano in 1992 [Ohba et al. 2008]; this suggests that the degassing magma beneath Onawashiro crater is depleted in CO2 relative to the magma of Unzen volcano. The fumarolic gas sampled at site A might be a steam phase generated by the mixing of magmatic gas and groundwater. The fumarolic gases at sites U, K and H showed significant condensation of water vapor. The addition of water vapor derived from groundwater would have occurred prior to the condensation. The N2-He-Ar compositions of the gases from site J suggests that the magmatic component is identical to that of andesitic volcanoes above a subduction zone. The correlation between the 3He/4He and 4He/20Ne ratios in the fumarolic gases suggests that the gases at site J contain an end member for which the 3He/4He ratio is as high as 5 (R/Ra). The correlation between the δ13C value of CO2 and the CO2/ 3He ratio suggests that the origin of the CO2 in the gases from site J was mainly limestone, with a minor component of MORB- type CO2. The hot spring waters discharged at site J comprise a direct mixture of MV and LMW. Acknowledgements. We sincerely thank Dr. Mauro Martelli for his constructive review of the manuscript, and Dr. Salvo Inguaggiato for the editorial handling of the manuscript. K 10 1 pH c = - Fm mi= c ( )H Cl HCl gas+ = + - ( ) 1000 ScF HCl C C C1000 HCl H O HCl 2 = + K m m m H Cl = + - OHBA ET AL. 196 197 References Aizawa, K., Y. Ogawa, M. Mishina, K. Takahashi, S. Na- gaoka, N. Takagi, S. Sakanaka and T. Miura (2009). Structural controls on the 1998 volcanic unrest at Iwate volcano: Relationship between a shallow, electrically re- sistive body and the possible ascent route of magmatic fluid, J. Volcanol. Geotherm. Res., 187, 131-139. Coleman, M.L., T.J. Shepherd, J.J. Durham, J.E. Rouse and G.R. Moore (1982). Reduction of water with zinc for hydrogen isotope analysis, Anal. Chem., 54, 993-995. Epstein, S. and T.K. Mayeda (1953). Variation of 18O con- tent of waters from natural sources, Geochim. Cosmo- chim. Acta, 4, 213-224. Giggenbach, W.F. (1980). Geothermal gas equilibria, Geo- chim. Cosmochim. Acta, 44, 2021-2032. Giggenbach, W.F. (1987). Redox processes governing the chemistry of fumarolic gas discharges from White Is- land, New Zealand, Appl. Geochem., 2, 143-161. Giggenbach, W.F. and R.L. Goguel (1989). Collection and analysis of geothermal and volcanic water and gas di- scharges, CD 2401, Chemistry Division, Department of Scientific and Industrial Research (DSIR), Petone, New Zealand. Giggenbach, W.F. (1992). Isotopic shifts in waters from geo- thermal and volcanic systems along convergent plate boundaries and their origin, Earth Planet. Sci. Lett., 113, 495-510. Giggenbach, W.F. (1997). The origin and evolution of fluids in magmatic-hydrothermal systems, In: H.L. Barnes (Editor), Geochemistry of Hydrothermal Ore Deposit, pp. 737-796. Horita, J. and D.J. Wesolowski (1994). Liquid-vapor frac- tionation of oxygen and hydrogen isotopes of water from the freezing to the critical temperature. Geochim. Cosmochim. Acta, 58, 3425-3437. Itoh, J. and N. Doi (2005). Geological map of Iwate vol- cano, Geological survey of Japan, AIST. Kazahaya, K., M. Yasuhara, T. Sato, M. Takahashi and A. Marui (2000). Isotope altitude effect for groundwaters in the Iwate volcano and Kakkonda areas, Rept. Geol. Surv. Japan, 284, 57-67. ( Japanese with English abstract) Kusakabe, M. (2001). A simple method for sampling total dissolved carbonate in carbonate-rich natural waters and CO2 preparation for δ 13C determination, Geochem. J., 35, 459-464. Marty, B., A. Jambon and Y. Sano (1989). Helium isotopes and CO2 in volcanic gases of Japan, Chem. Geol., 76, 25-40. Miura, S., S. Ueki, T. Sato, K. Tachibana and H. Hamagu- chi (2000). Crustal deformation associated with the 1998 seismo-volcanic crisis of Iwate volcano, northea- stern Japan, as observed by a dense GPS network, Earth Planet Space, 52, 1003-1008. Noguchi, K., S. Ueno, M. Ichikuni and T. Goto (1961). Che- mical composition of fumarolic gases obtained from volcano Iwate, Bull. Volcanol. Soc. Japan, 5, 293-298. Ohba, T. (2007). Formation process of recent fumarolic gases at the Mt. Mihara summit peak of the Izu-Oshima volcano, Japan, Earth Planet Space, 59, 1127-1133. Ohba, T., J. Hirabayashi, K. Nogami, M. Kusakabe and M. Yoshida (2008). Magma degassing process during the eruption of Mt. Unzen, Japan in 1991 to 1995: Mode- ling with the chemical composition of volcanic gas, J. Volcanol. Geotherm. Res., 175, 120-132. Ohba, T., T. Sawa, N. Taira, T. F. Yang, H.F. Lee, T.F. Lan, M. Ohwada, N. Morikawa and K. Kazahaya (2010). Magmatic fluids of Tatun volcanic group, Taiwan, Appl. Geochem., 25, 513-523. Ohwada, M., T. Ohba, J. Hirabayashi, K. Nogami, K. Na- kamura and K. Nagao (2003). Interaction between mag- matic fluid and meteoric water, inferred from 18O/16O and 36Ar/H2O ratios of fumarolic gases at the Kusatsu- Shirane volcano, Japan, Earth Planet Space, 55, 105-110. Ozawa, T. (1968). Chemical analysis of volcanic gases: I. Chemical analysis of volcanic gases containing water vapor, hydrogen chloride, sulfur dioxide, hydrogen sul- fide, carbon dioxide, etc, Geochem. Int., 5, 939-947. Poreda, R. and H. Craig (1989). Helium isotope ratios in circum-Pacific volcanic arcs, Nature, 338, 473-478. Sano, Y. and B. Marty (1995). Origin of carbon in fumaro- lic gas from island arcs, Chem. Geol., 119, 265-274. Simonson, J. M. and D. A. Palmer (1993). Liquid-vapor par- titioning of HCl(aq) to 350 °C, Geochim. Cosmochim. Acta, 57, 1-7. Tanaka, S., H. Hamaguchi, T. Nishimura, T. Yamawaki, S. Ueki, H. Nakamichi, T. Tsutsui, H. Miyamachi, N. Mat- suwa, J. Oikawa, T. Ohminato, K. Miyaoka, S. Onizawa, T. Mori and K. Aizawa (2002). Three-dimensional P- wave velocity structure of Iwate volcano, Japan from active seismic survey, Geophys. Res. Lett., 29, 10.1029/2002GL014983. Taran, Y.A., B.G. Pokrovsky and Y.M. Doubik (1989). Iso- topic composition and origin of water from andesitic magmas, Dokl. Acad. Sci. USSR, 304, 440-443. (in Rus- sian). *Corresponding author: Takeshi Ohba, Tokai University, School of Science, Department of Chemistry, Kanagawa, Japan; e-mail: takeshi_ohba@tokai-u.jp. © 2011 by the Istituto Nazionale di Geofisica e Vulcanologia. All rights reserved. FUMAROLIC GASES AT IWATE VOLCANO, JAPAN