HYDROGEN AND OXYGEN ISOTOPIC COMPOSITIONS OF WATERS FROM FUMAROLES AT KILAUEA SUMMIT, HAWAII Todd K. Hinkley Mail Stop 903 U.S. Geological Survey Box 25046 Denver Federal Center Denver, CO 80225 Tel: (303) 236-5858 FAX: (303) 236-1414 email: thinkley@greenwood.cr.usgs.gov James E. Quick U.S. Geological Survey MS 903 Box 25046 Denver Federal Center Denver, CO 80225 Robert T. Gregory Department of Geological Sciences Southern Methodist University Dallas TX 75275 Terrence M. Gerlach Cascades Volcano Observatory U.S. Geological Survey 5400 MacArthur Blvd. Vancouver WA 98661 Abstract Condensate samples were collected in 1992 from a high-temperature (300) fumarole on the floor of the Halemaumau Pit Crater at Kilauea. The emergence about two years earlier of such a hot fumarole was basically unprecedented at such a central location at Kilauea. The condensates have hydrogen and oxygen isotopic compositions that indicate that the waters emitted by the fumarole are composed largely of meteoric water, that any magmatic water component must be minor, and that the precipitation that was the original source to the fumarole fell on a recharge area on the slopes of Mauna Loa Volcano to the west. However, the fumarole has no tritium, indicating that it taps a source of water that has been isolated from atmospheric water for at least 40 years. It is noteworthy, considering the unstable tectonic environment and abundant local rainfall of the Kilauea and Mauna Loa regions, that waters that are sources to the hot fumarole remain uncontaminated from atmospheric sources over such long times, and long transport distances. As for the common, boiling point fumaroles of the Kilauea summit region, their 18O, D, and tritium concentrations indicate that they are dominated by recycling of present day meteoric water. Though the waters of both hot and boiling point fumaroles have dominantly meteoric sources, they seem to be of separate hydrologic regimes. Large concentrations of halogens and sulfur species in the condensates, together with location at the center of the Kilauea summit region and high temperature, initially suggested that much of the total mass of the emissions of the hot fumarole, including the H2O, might have come directly from a magma body. The results of the present study indicate that it is unreliable to infer a magmatic origin of volcanic waters based solely on halogen or sulfur contents, or other aspects of chemical composition of total condensates. INTRODUCTION Little is known about the hydrothermal systematics of Kilauea Volcano, or of volcanic edifices in general. Field and chemical studies of water and other emitted volatiles have been conducted since the nineteenth century (see Shepherd, 1938). Geophysical studies and drilling have located bodies of groundwater in the Kilauea region (Keller et al., 1979; Jackson and Kauahikaua, 1988). Most fumaroles on Hawaii are low temperature (<100°C, Casadevall and Hazlett, 1983), called "boiling point" fumaroles because they are maintained at their boiling temperatures for their respective elevations, by the latent heat release from liquid condensation. They have generally been interpreted as emitting meteoric water that has been cycled through shallow hydrothermal convection systems. Hot fumaroles, with temperatures of about 300°C, were recently described (Gerlach et al., 1991) within Halemaumau Pit Crater, within the Kilauea summit region (Fig. 1). Because of the unprecedented combination of their high temperature and their location at the possible epicenter of the Kilauea magmatic system, we wished to collect samples from them for information about the hydrothermal systematics of this much-studied volcano. We analyzed the samples for isotopes of oxygen and hydrogen to give information about possible sources of the waters emitted by the fumaroles. Tritium was of particular interest, because of its short half- life, for detecting the presence of young meteoric waters. Our results demonstrate that, contrary to preliminary indications from fumarole chemical composition that emissions could have a magmatic source, all of the fumaroles were dominated by meteoric water. SITE, SAMPLE COLLECTION, AND COLLECTION SYSTEMATICS Sample sites of the present study are shown in Fig. 1. Two hot fumaroles, W-2 and W-3, were on the south end of the floor of the Halemaumau Pit Crater, and vented with temperatures of ~300C in 1991 and 1992. We also sampled two boiling point fumaroles, one known as HMM located near the rim of the Halemaumau Pit Crater, the other known as Sulphur Bank, ~4km away near the rim of the Kilauea caldera. The main suite of condensates was collected in latest March and earliest April 1992, after several months of lower-than-average rainfall. We used a sampling apparatus consisting of three serial traps that allows collection of the large volumes (~1 l) necessary for high-precision tritium analysis. Details of sampling are given in the caption of Fig. 2. Tritium data are presented in Table 1 for first collection traps only, because the volumes of condensates collected in subsequent traps were too small to analyze (Ostlund and Dorsey, 1975). A preliminary 1991 collection used a simpler apparatus, and was made following a period of normal rainfall. All of the isotopic data from the traps indicate that stable isotopes are fractionated by the sampling apparatus. Large differences between isotopic compositions of successive traps in the collection apparatus suggest that no single trap of our apparatus gives the true composition of the waters actually emitted by a fumarole. Contrasting D and 18O values between first and second traps were 40 and >4 per mil, respectively. The compositions of the condensates in the first traps provide upper limits on the D and 18O concentrations of the waters emitted by the fumaroles, because condensed water is isotopically heavier than the vapor from which it condenses, in the pressure and temperature ranges of our experiments. However, the original isotopic composition of fumarole water actually emitted by the fumarole can be inferred more precisely, by three different methods that can be used for correcting or restoring theD and 18O values of the condensates to the probable actual compositions of waters originally emitted by the fumaroles. The three methods are used to plot, in Figures 3 and 4, these D and 18O compositions of the original emissions. In the Figures the resulting data points are shown by separate symbols for each method. The methods use the isotopic compositions and masses of the samples in the serial collection traps, and are as follows. First, if it is assumed that insignificant water is lost from the condensation line, the mass-weighted average of the isotopic compositions of the condensates in the separate traps is the original isotopic composition of the vapor from the fumarole. Second, condensation may be modelled as a batch process. The batch model is described by two simultaneous equations: where n refers to the vapor leaving the first measured trap and n+1 refers to the next trap downstream. The fraction of vapor, f, refers to the vapor exiting the trap after condensation and equilibration in the trapping volume. The quantity (1-f) is the fractional amount of liquid condensed in the trap. The quantity a is the liquid-vapor isotopic fractionation factor for 18O. Analogous equations may be written for hydrogen, with a different fractionation factor. As a third alternative, the initial composition of the vapor may be modelled assuming Rayleigh condensation in the traps. The stable isotopic compositions of water (l) condensed from vapor during Rayleigh fractionation are given by: dl = a[dvo + 1000]fa-1 -1000 with one equation for each isotope, 18O and D. vo is the initial isotopic composition of the vapor. The average stable isotopic compositions, , for 18O and D of liquids condensed in collection trap number t can be calculated by integrating the above equations, with the result:
= {[dvo + 1000][ft-ft-1a] - 1000}/[ft-ft-1] where ft-1 and ft are the fractions of initial water in the vapor entering and leaving the trap, respectively. In the special case of the first trap, ft-1 = fo = 1. Because more than one trap was collected, a series of parallel equations may be constructed and solved for 18Ovo and D. For the Rayleigh calculation, a values were specified based on the temperature dependencies described by Bottinga and Craig, 1969, and Majoube, 1971 respectively. We assumed temperatures in the first traps of 97C, the boiling temperature at the altitude of collection, and in subsequent traps of 23C, the ambient temperature. All three models predict compositions closer to those of the first traps, consistent with the largest volumes being retained in the first traps. The compositions specified by the batch and Rayleigh models are indistinguishable from one another. Both are isotopically lighter than the values specified by the simple mass balance model, because the latter does not account for the small amounts of isotopically light H2O that did escape from the collection system. For boiling point fumarole HMM and hot fumarole W-3, all three correction procedures predict emitted water compositions that lie on the lines joining the compositions of the first and second traps (Figure 3). At boiling point fumarole Sulphur Bank, the condensate collected in the third trap was analyzed for D and 18O separately, whereas at W-3 and HMM the condensates from second and third traps were combined. At W-3 and HMM the volumes in first traps were about ten times larger than those in subsequent traps. At Sulphur Bank, the volume in the first trap was only about twice as large as in the subsequent traps, indicating that the collection process was somehow different there. Also, compared to the other two fumaroles, the total amount of water collected per unit time at Sulphur Bank was greater by about 50%, and larger amounts of condensed water were visible in the line leading from the fumarole to the first trap. RESULTS The volumes of condensates collected in traps are presented in Table 1. Also in Table 1 are isotopic compositions of condensates in each trap, and of rain, river water, and ocean water from the island of Hawaii analyzed as part of the present study. The stable isotopic data for the fumarole condensates are presented diagrammatically in Figure 3. The condensates from HMM define a linear array that lies to the left of the meteoric water line (Craig, 1961), and parallel to it. The hot fumarole W-3, despite venting at a temperature 200C hotter than HMM, also defines a linear array that parallels the meteoric water line, but lying to its right. In contrast, the data from the boiling point Sulphur Bank fumarole form a more gently sloping array discordant to the meteoric water line. The boiling point fumaroles are clearly distinguished from the hot fumarole by their much higher tritium concentrations (Table 1), ranging from 0.8 T.U. at Sulphur Bank to 1.5 at HMM. These values, although slightly lower than values for rain and ocean water, are consistent with the presence of young meteoric water (Mazor, 1991; table, chap. 10). In contrast, the high temperature fumarole samples have tritium concentrations that are close to or equal to zero. DISCUSSION Although the ranges in isotopic compositions between separate traps for each of the collections show that there is a large amount of fractionation during condensation, the actual isotopic compositions of water emitted from the fumaroles can be closely estimated in three different ways, using the compositions and masses of the collections in each serial trap from the condensation, as discussed above. These three correction methods lead to restored isotopic composition values that are very close to one another, although there are slight, systematic differences among them, as seen by the distinct symbol types shown in Figs. 3 and 4. The restored compositions define the isotopic compositions of the waters emitted by the fumaroles with accuracy and precision adequate for the discussion of the present study. The data indicate that none of the isotopic ratios of the fumarole waters is close to those of magmatic waters. Estimates for the composition of magmatic water lie within a relatively small range of oxygen isotopic compositions, to the right of the meteoric water line (Sheppard and Epstein, 1970; Dixon, et al., 1991; Figure 4). For purposes of this paper, magmatic water refers to H2O that was dissolved in a silicate magma independent of its origin (e.g. Sheppard, 1986). The water dissolved in a silicate magma can, in principle, come from a variety of sources in the crust and the mantle. Because of the effect of plate tectonics on the recycling of surface fluids into the source regions for basalt, it is virtually impossible to differentiate juvenile water from magmatic water on the basis of oxygen and hydrogen isotopes (e.g. Taylor, 1977; Sheppard, 1986). The most probable sources of magmatic water are from the melting of hydrous minerals in the source regions of magmas and from the assimilation of hydrous wall rocks during the ascent through the crust or from residence in crustal magma chambers where stoping occurs. There is little stable isotopic evidence that crustal pore fluids directly diffuse across plutonic contacts into molten magmas (Taylor and Forester, 1979). However, assimilation and stoping of hydrothermally altered rocks into magma chambers can indirectly transfer surface waters into magmas (Taylor, 1977; Michael and Schilling, 1989). In general, oxygen is a major element in silicate magmas, and magmatic waters are enriched in 18O relative to the magma, i.e. the silicate oxygen reservoir buffers the oxygen isotope composition of magmatic water. Hydrogen is a minor element in silicate magmas. Therefore, the D value of magmatic waters may more directly reflect the immediate source of the fluid or the mixing of hydrogen reservoirs that results from partial melting and assimilation coupled with fractional crystallization. From the definition, the stable isotopic composition of magmatic water is expected to be in isotopic equilibrium with silicate magma. Because silicate-water oxygen isotope fractionations (the difference in the 18O values between magma and water) reverse at magmatic temperatures (change from positive to negative; e.g. see Javoy, 1977), the oxygen isotope composition of magmatic water is enriched relative to the magma. Hence, + 6 basalts exsolve magmatic waters with 18O values >+6. In basaltic systems, these magmatic waters are expected to exsolve at temperatures approaching 1000C. Most fumaroles on Hawaii do not approach these temperatures. If the fumaroles contain any exsolved magmatic water, the water must lose heat to the surrounding country rocks. Because the stable isotope fractionation between fluid and water is a function of temperature, the exsolved fluid will exchange stable isotopes with the surrounding rocks during the ascent to the surface. In a rock-dominated system, the fluid becomes more depleted (i.e., takes smaller 18O and D concentrations) the lower the temperature. At the temperature of the hot fumaroles (300C, the fluid in equilibrium with typical Hawaiian basalt still would have 18O>+4 (e.g. Bowers and Taylor, 1985). At higher ratios of magmatic fluid to rock, the fluid would be even more enriched. The high temperature fumaroles at Kilauea have 18O values inconsistent with their origin as magmatic waters isotopically reequilibrated at lower temperature. The Kilauea hot fumarolic waters, 18O -4, are far too depleted in 18O to be explained by this mechanism. Hydrogen isotopes are more difficult to interpret in magmas because of the low abundance of hydrogen in the crust and mantle. In basaltic and andesitic systems D values range from about -25 to -85 per mil (e.g. Taylor, 1986). Because this range of D values overlaps both the range of surface fluids over a large part of the Earth (see table 1) and the range of D values of many sedimentary and metamorphic rocks, hydrogen isotopes are less diagnostic of the sources of fumarolic fluids. Even though the hydrogen isotopic compositions of the fumaroles are not diagnostic, the 18O-depleted character of the fumarole waters clearly implicates meteoric waters. The term meteoric water refers to any waters passing through the surface water cycle in the form of atmospheric precipitation, rain, snow, ice, etc. These waters are ultimately derived from the ocean. Lakes, rivers, and shallow ground water reservoirs are dominated by meteoric waters. Modern meteoric waters are remarkably regular in their stable isotopic composition, exhibiting a linear relationship between D and 18O values known as the meteoric water line. The proximity of the data of the present study to the meteoric water line suggests that the fumaroles are dominated by meteoric water. The position of HMM to the left of the water line suggests that there is no significant involvement of magmatic water in that system. Although it is possible that W-3 and Sulphur Bank waters may be mixtures of meteoric and magmatic water, Figure 4 indicates that no more than about 15 percent of W-3 and 25 percent of Sulphur Bank could be magmatic. In fact, the stable isotopic compositions of all of these fumaroles can be adequately explained by processes that do not involve magmatic waters. Processes affecting the isotopic compositions of waters of fumaroles are likely to be complex, and different effects may dominate different fumaroles. Fumarolic waters at W-3 and Sulphur Bank may have been shifted away from the water line to heavier oxygen values by exchange with basaltic rocks at high temperature (Figure 4; Craig, 1966; Muehlenbachs, 1986). The composition of HMM may represent a shift away from the meteoric water line by low-temperature hydration of basaltic glass, which moves water compositions toward lighter oxygen compositions and slightly heavier hydrogen compositions (Figure 4; Friedman and Smith, 1958; Frape and Fritz, 1982). Alternatively HMM waters may be vapor products of disequilibrium evaporation. In any case, all three fumaroles are dominantly if not completely meteoric water. This is particularly important in the case of W-3, which had previously been identified as a candidate for presence of magmatic water on the basis of its high temperature, central location, and the presence of halogen and sulfur species (to 1600 ppm HCl and 400 ppm HF, Gerlach et al., 1992). Because of mass and concentration considerations, halogen and sulfur species must necessarily be identified with sources in magma bodies, whereas water components of fumarole emissions have multiple and varied potential sources. At Sulphur Bank the gentle slope of the data array suggests disequilibrium condensation (Craig, 1966). The slope could also be explained by the presence in the fumarole emissions of a mixture of gaseous and liquid H2O in isotopic disequilibrium. Sulphur Bank was the only fumarole where water droplets were abundant in the line leading from the fumarole to the first condensation trap. No point calculated by the Rayleigh model (see above) for isotopic composition of water emitted by Sulphur Bank is shown in Fig. 3. In any case, both batch and mass balance models for calculating the isotopic composition of water emitted by the fumarole indicate that the composition of Sulphur Bank is similar to that of W-3. Tritium concentrations in the fumaroles place important constraints on the residence time of meteoric waters in the subsurface, because of the short half-life of that nuclide (Jenks, 1955). Effectively, all tritium in meteoric water results from atmospheric thermonuclear testing that began in the early 1950's (Ostlund and Mason, 1974, 1985), except for a small amount from decay of cosmogenic 15N, also in the atmosphere (Mazor, 1991). Condensates from boiling point fumaroles have tritium abundances in the range of present-day meteoric water (Ostlund and Mason, 1974). In marked contrast, waters of the hot fumarole system contain virtually no tritium (Table 1). The absence of tritium in waters of the hot fumarole W- 3 indicates that the meteoric component of that fumarole was isolated from the atmosphere for at least 40 years, the length of time that has passed since the beginning of atmospheric thermonuclear bomb testing (Doney et al., 1992). Regardless of whether the composition of waters of hot fumarole W-3 reflects high-temperature exchange with rock, or mixing with a small amount of magmatic water, the relationships in Figures 3 and 4 suggest that the original composition of the meteoric component of that fumarole was approximately 18O = -6 and D = -40. This isotopic composition is bracketed by the compositions of precipitation on the island of Hawaii (Fig. 4 of present paper; also see Scholl et al., 1993). More specifically, it is bracketed by the extremes of the precipitation that falls over a range of elevations on a recharge area on the flanks of Mauna Loa (Fig. 3). The isotopic composition of water in aquifers typically represents that of the local precipitation contributing to groundwater recharge (Joseph et al., 1992; Fontes et al., 1969; Arnason, 1977). Permeability data for Kilauea Summit (Keller et al., 1979) indicate that decades would be required for water to reach the fumaroles at Kilauea from that area, consistent with the requirement from the tritium data that meteoric water be isolated for more than 40 years. The Kilauea region in general is subject to continual volcanic intrusion and eruption and accompanying tectonic dynamic processes, and is an area of high average rainfall. More specifically, the Halemaumau Pit Crater is certainly subject to the same processes. Large fluxes of rainwater cascade over its rim onto its floor during intense rains. It is an accumulation point for a substantial surrounding drainage area. It is noteworthy in light of both tectonic and rainfall considerations that the hot fumarole waters appear uncontaminated by young, tritium-bearing meteoric waters. At a minimum, our results extend the findings of Kroopnik et al. (1978) who reported the presence of low-tritium bodies of groundwater in the more tectonically and thermally stable Puna region of Hawaii. If only from relatively high emission temperature, it is clear that the hot fumarole waters have been in contact with rock of substantially high temperatures. At Kilauea and other sites, long-period earthquakes, commonly preceding phreatic eruptions, may be associated with presence of waters in deep, hot zones of the volcanic system (A. Okamura, individual communication; Latter, 1981) . It is possible that the waters emitted from the hot fumaroles at Kilauea are related or identical to the waters involved in such seismic events. Although the stable isotopic compositions of emissions of neither the hot nor the boiling temperature fumaroles are remarkable in themselves, the details of the hydrogen and oxygen isotopic compositions of the two types of fumaroles (hot, and boiling point) indicate that there are two distinct regimes of water in the hydrothermal system of the Kilauea summit region. A segregation of two components of the hydrothermal system at Larderello was recently proposed by Petrucci et al. (1993). Two decades ago, convincing demonstrations were made that fumaroles may be universally dominated by meteoric water (e.g., Taylor, 1974). Recent work on active volcanoes, however, has proposed that waters of magmatic origin do exist as significant components of hydrothermal and fumarolic fluids (Bolognesi and D'Amore, 1993; Ciodini et al., 1993; Taran et al., 1991; Giggenbach and Soto, 1992; Shevenell and Goff, 1993; Giggenbach, 1992a, b; Blattner, 1993). The initial chemical data on the hot Kilauea fumaroles (high halogen and sulfur content, Gerlach et al., 1992), and their placement at the possible epicenter of the Kilauea magmatic system, suggested that they might prove to have a substantial, identifiable magmatic water component. The finding of the present study based on light stable isotopic data that the hot fumarole waters are dominantly or entirely meteoric in origin shows that it may be entirely inappropriate to assume a magmatic origin of hydrothermal waters on the basis of chemical composition alone. CONCLUSIONS Water condensates from new hot fumaroles at Kilauea have stable isotopic compositions that indicate that they, like the common and long-studied boiling point fumaroles of the Kilauea summit region, are mostly or entirely meteoric water. Any magmatic water component must be small. Absence of tritium in the waters of the hot fumaroles indicates that they, in contrast to the waters of boiling-point fumaroles, have been isolated from communication with tritium-bearing atmospheric waters for more than 40 years. The light stable isotope data indicate that the source of water for the hot fumaroles is the catchment area on the slopes of Mauna Loa. But the tritium data indicate that this water has been isolated from contamination by young atmospheric water for the most recent several decades or longer, during transport over distances of several kilometers to the Kilauea summit area. It is noteworthy, considering the unstable tectonic environment of the Kilauea and Mauna Loa regions, and abundant local rainfall, that these waters remain uncontaminated over such long times, and long transport distances. The overall hydrogen and oxygen isotopic compositions of the waters of the hot and boiling point fumaroles of Kilauea indicate that they are two components of the hydrothermal regime that are not in mutual communication, and may be totally separate and distinct. Initial chemical information from the condensates of the hot fumaroles (large sulfur, and especially halogen concentrations), together with their central placement, seemed to indicate that they might have substantial components of magmatic water. The isotopic data indicate that the water component of the fumarolic gas is still dominated by meteoric water. ACKNOWLEGDMENTS We thank M.T. Colucci for light stable isotopic analysis REFERENCES Arnason B (1977) Hydrothermal systems in Iceland traced by deuteriuum. Geothermics 5:125-151 Blattner P (1993) "Andesitic water": a phantom of the isotopic evolution of water-silicate systems. Comment on "Isotopic shifts in waters from geothermal and volcanic systems along convergent plate boundaries and their origin" by W.F. Giggenbach. Earth Plan. Sci. Lett. 120: 511-518. Bolognesi L, and D'Amore F (1993) Isotopic variation of the hydrothermal system on Vulcano Island, Italy. Geochim et Cosmochim Acta 57:2069-2082 Bottinga Y, and Craig H (1969) Oxygen isotope fractionation between CO2 and water and the isotopic composition of the marine atmosphere. Earth Planet Sci Lett, 5:285-295 Bowers, TE, and Taylor, HP Jr (1985) An integrated chamical and stable-isotope moodel of the origin of midocean ridge hot spring systems, J Geophys Res 90:12583-12606. Casadevall TJ, and Hazlett RW (1983) Thermal areas on Kilauea and Mauna Loa Volcanoes, Hawaii. J Volc Geotherm Res 16:173-188 Cheng P and Lau KH (1978) Steady state free convection in an unconfined geothermal reservoir. J. Geophys. Res. 79, 4425-4431. Chiodini G, Cioni R, and Marini L. (1993) Reactions governing the chemistry of crater fumaroles from Vulcano Island, Italy, and implications for volcanic surveillance. Applied Geochem 8:357-371 Craig H (1961) Isotopic variations in meteoric waters. Science133:1702-1703 Craig H (1966) Isotopic composition and origin of the Red Sea and Salton Sea geothermal brines. Science 154:1544-1548 Dixon JE, Clague DA, and Stolper EM (1991) Degassing history of water, sulfur, and carbon in submarine lavas from Kilauea volcano, Hawaii. J Geol 99:371-394 Doney SC, Glover DM, and Jenkins WJ (1992) A model function of the global bomb tritium distribution in precipitation, 1960-1986. J Geophys Res, 97, C4:5481-5492 Fontes J Ch, Maglione G and Roche MA (1969) Donnees isotopiques preliminaires sur les rapports du lac Tchad avec les nappes de la bordure nord-est. Cah ORSTOM, ser Hydrol, VI,1:17-34 Frape SK, and Fritz P (1982) The chemistry and isotopic composition of saline groundwaters from the Sudbury Basin, Ontario, Canada. Can J Earth Sci 19:645-661 Friedman I, and Reimer GM (1987) Helium at Kilauea Volcano, Part I, Spatial and temporal variations at Sulphur Banks. USGS Professional Paper 1350: 809-813 Friedman I and Smith RL (1958) The deuterium content of water in some volcanic glasses. Geochim Cosmochim Acta 15:218-228 Gerlach TM, Hinkley TK, Briggs P, and Stokes JB (1991) Fumarole gases of Halemaumau Pit Crater and summit magma chamber degassing models for Kilauea Volcano. Trans Am Geophys Un (Eos), 72, 44 Abstract Vol p 562 Giggenbach WF (1987) Redox processes governing the chemistry of fumarolic gas discharges from White Island, New Zealand. Appl Geochem 2:143-161 Giggenbach WF (1992a) Magma degassing and mineral deposition in hydrothermal systems along convergent plate boundaries. Econ Geol 87:1927-1944 Giggenbach WF (1992b) Isotopic shifts in waters from geothermal and volcanic systems along convergent plate boundaries and their origins. Earth. Plan. Sci. Lett. 113: 495-510. Giggenbach WF and Soto CR (1992) Isotopic and chemical composition of water and steam discharges from volcanic-magmatic-hydrothermal systems of the Guanacaste Geothermal Province, Costa Rica. Appl Geochem 7:309-332. Jackson DB, and Kauahikaua J, (1988) Regional self-potential anomalies at Kilauea Volcano. US Geol Survey Prof Paper 1350 2:47-960 Jenks WM (1955) Half-life of tritium. Physical Rev 100:124-125 Joseph A, Frangi JP, and Aranyossy FF (1992) Isotope characteristics of meteoric water and groundwater in the Sahelo-Sudanese Zone. J Geophys Res 97, D7:7543-7551 Javoy M (1977) Stable isotopes and geothermometry, J Geol Soc Lond 133: 609-636. Keller GV, Grose LT, Murray JC and Skokan CK (1979) Results of an experimental drill hole at the summit of Kilauea volcano, Hawaii. J Volc Geotherm Res 5:345-385 Kroopnick PM, Lau LS, Buddemeier RW, and Thomas D (1978) Geochemistry of a Hawaiian geothermal well, HGP-A. Geothermal Resources Council, Transactions 2:375-377 Latter JH (1981) Volcanic earthquakes, and their relationship to eruptions at Ruapehu and Ngauruhoe Volcanoes. J Volcanol Geotherm Res 9:293-309 Majoube M (1971) Fractionnement en oxygene-18 et en deuterium entre l'eau et la vapeur. J Chem Phys 10:1423-1436 Mazor E (1991) Applied chemical and isotopic groundwater hydrology. Halsted Press, J Wiley & Sons, New York, 274 pp Michael PJ and Schilling JG (1989) Chlorine in mid-ocean ridge magmas: Evidence for assimilation of seawater influenced components, Geochim Cosmochim Acta 53: 3131-3143. Muehlenbachs K (1986) Alteration of the oceanic crust and the 18O historyof seawater, 425-444. In JW Valley, HP Taylor, Jr, and JR O'Neil, eds., Stable isotopes in high temperature geological processes, Reviews in Mineralogy, 16 Ostlund HG, and Mason AS (1974) Atmospheric HT and HTO: I, Experimental procedures and troposheric data 1968-72. Tellus 26:91-101 Ostlund HG, and Dorsey HG (1975) Rapid electrolytic enrichment and hydrogen gas proportional counting of tritium, Proc. Conf. on Low Radioactiv. Meas. 1975, 55-60 in Low-radioactivity measurments and applications, Slovenske Pedagogicke Nakladatelstvo, Bratislava, 1977, 549 pp. Ostlund HG, and Mason, AS (1985) Atmospheric tritium 1968-1984, Tritium Laboratory Data Report n.. 14; Univ. of Miami, Rosenstiel Sch. of Mar. and Atm. Sci., Miami, Florida, 134 pp. Petrucci E, Sheppard SMF, and Turi B (1993) Water/rock interaction in the Larderello Geothermal Field (Southern Tuscany, Italy): an 18O/16O and D/H isotope study. J Volcanol and Geotherm Res 59:145-160 Reimer GM (1987) Helium at Kilauea Volcano, distribution in the summit region. USGS Prof Paper 1350 Part IIK,, 815-819 Scholl MA, Janik, C.J., Ingebritsen SE, Kauahikaua, JP and Trusdell, FA (1993) Preliminary results from an isotope hydrology study of the Kilauea Volcano area, Hawaii. Geothermal Resources Council Transactions 17: Oct. 1993, 187-194. Shepherd ES (1938) The gases in rocks and some related problems. Am. Jour. Sci. 5th ser. v.35A 311-351 Sheppard SMF, and Epstein S (1970) D/H and O18/O16 ratios of minerals of possible mantle or lower crustal origin. Earth Planet Sci Lett 9:232-239 Sheppard SMF (1986) Characterization and isotopic variations in natural waters. in Valley JW, Taylor HP Jr, and O'Neil JR, eds., Stable isotopes in high temperature geological processes, Reviews in Mineralogy, 16, Min. Soc. Am.: 165-184 Shevenell L and Goff F (1993) Addition of magmatic volatiles into the hot spring waters of Loowit Canyon, Mount St. Helens, Washington, USA. Bull Volcanol 55:489-503 Taran Yu A, Rozhkov AM, Serafimova EK, and Esikov, AD (1991) Chemical and isotopic composition of magmatic gases from the 1988 eruption of Klyuchevskoy volcano, Kamchatka. J Volcanol and Geotherm Res 46:255-263 Taylor BE (1986) Magmatic volatiles: isotopic variation of C, H, and S, in Valley JW, Taylor HP Jr, and O'Neil JR, eds., Stable isotopes in high temperature geological processes, Reviews in Mineralogy, 16, Min. Soc. Am.: 185-226. Taylor HP Jr (1977) Water/rock interactions and the origin of H2O in granitic batholiths, J Geol Soc Lond 133: 509-558. Taylor HP Jr and Forester RW (1979) An oxygen and hydrogen isotope study of the Skaergard intrusion and its country rocks: a description of a 55-M.Y. old fossil hydrothermal system J Petrol 20: 355-419. Tilling RI, and Dvorak JJ (1993) Kilauea: anatomy of a basaltic volcano. Nature 363:125-133 Tilling RI, and Jones BF (1991) Composition of waters from the research drill hole at summit of Kilauea Volcano and of selected thermal and non-thermal groundwaters, Hawaii. US Geol Surv Open File Report 91-133-A:27 Tilling RI, and Jones BF (in review in Feb 1993) Temporal variations in composition of water above an active basaltic volcano, Kilauea, Hawaii: 1973-1976 FIGURE CAPTIONS Figure 1 Map of the Kilauea Summit Region, showing main Kilauea Crater, Halemaumau Pit Crater with hot fumaroles, and boiling point fumaroles HMM and Sulphur Bank. Inset shows regions of island of Hawaii constituted by distinct volcanoes (K, Kilauea; ML, Mauna Loa; MK, Mauna Kea; H, Hualalai; Ko, Kohala). Figure 2 Schematic drawing of apparatus for condensation of large volumes needed for tritium measurement. Placement of a silica glass tube inside the titanium pipe, with an inner tube jacketed by a larger-diameter reflectively-plated tube and with evacuated void space, was designed to prevent early condensation, fallback and loss. Loops immersed into a water-and-ice bath served as a non-collecting preliminary cooler- condenser. Water condensing there was visibly swept on, and only a small amount (<5 ml) remained at end of collection. The main jar was immersed in the bath only to a level below the lid, to preclude possiblity of access of bath water through threaded seal into the sample. Condensing surfaces inside the preliminary loops and inside the first jar remained far above bath temperature. Some H2O was lost from the train, as condensed drops (<5 percent of collected volume) and as uncondensed vapor in amounts limited by partial pressure of water at ambient temperature at sampling elevation (~0.035 atm). Such lost H2O was inferred to be depleted in the heavy isotopes relative to waters collected in the traps. Apparatus used at cold fumaroles was identical except teflon tube was inserted directly into the fumarole mouth. Gases exiting the train were mostly SO2 and CO2 (together ~50 mole percent of the gases emitted by the fumarole, Gerlach, unpublished data). Hot fumarole samples taken in 1991 used a simpler apparatus than that shown, without the insulated silica tube at the fumarole mouth. Figure 3 Hydrogen and oxygen isotope variation of condensates taken from hot and cold fumaroles of the Kilauea Summit Region. Also shown are corrections, by three different models, to restore data to compositions of waters originally emitted by fumaroles (see text). Periods 1 and 2 were successive collections at the W-3 fumarole. Figure 4 Isotopic compositions of waters emitted by hot and boiling temperature fumaroles, compared to magmatic water and to precipitation on the island of Hawaii (Tilling and Jones; 1991; and in review 1993; International Atomic Energy Agency data for composition of rain at Hilo for 1963, weighted mean, 3.5 meters of precip.; elevation of summit of Mauna Loa is 3959 m). Fumarole compositions shown are restorations, calculated by three different models, to correct for sampling bias due to condensation in collector (see text). Magmatic water composition of lower part of box from Sheppard, 1986 (Fig. 7); other data on magmatic water compositions associated with rocks of specific compositional types can be found in Taylor, 1986 (Figure 15 and Table 6). Slopes and directions are shown for processes that alter the isotopic compositions of waters. October 21, 1994 Dr. W. F. Giggenbach Institute of Geological and Nuclear Sciences P O. Box 31 312, Lower Hutt New Zealand Dear Werner Please find enclosed our paper "Hydrogen and oxygen isotopic composition of waters from fumaroles at Kilauea Summit, Hawaii", revised in accordance with the suggestions of the two reviewers and yourself. We have added to the discussion section of the paper to be more specific about "magmatic water", its possible origins, some basics of the relations between magmas and rocks, and the stable isotopes of hydrogen and oxygen. The additions are in the second through eighth paragraphs of the discussion section, on pages 7, 8 and 9. We hope that this new material satisfactorily responds to the queries of the reviewer who wrote the single page of comments. Five additional references follow from this added discussion, at least one suggested by the reviewer. Elsewhere in the text we inserted a brief comment to clarify that sulfur and chlorine have a magmatic source, whereas H2O may have multiple sources (re: query [5] of the other reviewer, who wrote the longer set of comments). In response to the reviewer's mention of the Ingebritsen and Scholl paper, we took another look at it, and confirmed our earlier impression that it does not bear directly enough on volcanic and high-temperature hydrology to warrant citing it in our paper. Similarly, we decided not to specifically address the matter of the importance of H2S in changing the deuterium content of waters. We think that our comment that our collection apparatus was designed to get the large amounts of condensate needed for high precision tritium analysis should satisfy the other reviewer's question (two sheets of comments, his comment [2]) about why our newly designed apparatus was needed. We think that the zero tritium content of the condensate from the hot fumarole should be enough to satisfy his query about contamination by air, and by the water that might be contained by such air. Comments [3] and [4] by that reviewer, about isotopic and temperature effects from water-rock interaction and about the meaning of "equilibrium", should also be satisfied by the new paragraphs added to the discussion section ( above). In response to this reviewer's comment [6], the 18O and D data pairs in the Table for rain at two locations are not off by an order of magnitude nor mixed up with one another: the isotopic compositions of rain around the Big Island do not in fact always follow textbook systematics. We moved the Appendix into the main body of the paper ("Site, sample collection, and collection systematics" section) as suggested by the final comment of the reviewer who wrote the single page. We added more detail on the chemical composition of the condensates, citing the proper reference. We hope that you will find that we have responded adequately to the comments of the referees, and that the paper is in satisfactory form to proceed with publication in Bull. Volc. If you have any additional concerns, I will be in a position to respond promptly. Sincerely Todd Hinkley tel: 303-236-5850 (voice); 303-236-1414 (fax); 303-355-2361 (dom.) thinkley@usgs.gov enclosures: the revised manuscript (???copies) "laser print" figures and table (blue envelope) copies of reviewers' comment sheets copy of letter from you disc, with ms. in Word for Windows, ASCII, and RTF