Gas discharges from four remote volcanoes in northern Chile (Putana, Olca, Irruputuncu and Alitar): a geochemical survey

We analyzed gas samples collected from fumaroles and bubbling pools at Irruputuncu, Putana, Olca and Alitar volcanoes located in the central Andes volcanic zone (northern Chile). The Irruputuncu and Putana fumarolic discharges showed outlet temperatures ranging from 83 ̊C to 240 ̊C and from 82 ̊C to 88 ̊C, respectively. The chemical and isotopic (3He/4He, dC-CO2, d O-H2O and dD-H2O) compositions of these discharges were similar to medium-to-high temperature volcanic gases from other active volcanoes in this sector of the Andean volcanic chain (e.g. Lascar volcano). Inorganic and organic gas geothermometers for the H2O-CO2-CO-H2, CO2-CH4 and C2-C3 alkenes-alkanes systems indicated equilibrium temperatures that exceed 500 ̊C at the gas sources. These relatively high temperatures are in agreement with the presence of relevantly high concentrations of magmatic gas emissions, including SO2. Olca and Alitar volcano fluid chemistries indicated lower amounts of magmatic-derived gas species, while both the helium and the water isotopic compositions suggested significant fractions of shallow, crustal/meteoric-originated fluids. These indicate contributions from a hydrothermal environment with temperatures <400 ̊C. The geochemical and isotopic features derived from the present study show that the Irruputuncu, Putana, Olca and Alitar volcanoes should be considered as active and thus warrant periodic geochemical monitoring to determine the evolution of these systems and their potential hazards.

We present the first analytical data of the chemical and isotopic compositions of the gas discharges (fumaroles and bubbling pools) from Irruputuncu, Olca, Putana and Alitar volcanoes, which were collected during two sampling campaigns in March and May, 2007.The main aims are: (i) to provide the first geochemical and isotopic dataset for the fluids discharged from these volcanoes; and (ii) to investigate the physical-chemical processes that occur between the magmatic source and the shallower hydrothermal aquifers.We also use these geochemical proxies to define a monitoring strategy for these volcanic systems.

Geological setting and volcanic activity (historical and present)
2.1.Irruputuncu volcano Irruputuncu (20˚45´S; 68˚34´W; 5,165 m a.s.l.) is a composite stratovolcano that is located at the border between Chile and Bolivia (Figure 1) [de Silva and Francis 1991].Irruputuncu lies within the SW portion of the collapse scarp from a pre-Holocene volcano [González-Ferrán 1995] the products of which cover Upper Miocene-Pleistocene dacitic ignimbrites (Ujina and Pastillos ignimbrites) [Vergara 1978, Vergara andThomas 1984].The Irruputuncu cone edifice contains two coalescent summit craters of ~200 m in diameter (Figure 2) [de Silva andFrancis 1991, Wörner et al. 2000].The cone is primarily composed of andesitic domes with an age of 0.14 ±0.04 Ma, as indicated by K-Ar dating, and pyroclastic deposits [Wörner et al. 2000].The most recent historical activity consisted of a series of small phreatic eruptions that produced a 1,000-m-high ash/vapor plume that was observed on November 26, 1995[Global Volcanism Program 1997].The current fumarolic activity along the inner SW flank of the southern summit crater produces an ~200-m-high plume.Recently, local inhabitants reported that a thermal spring located 13 km west of the volcano [Hauser 1997] disappeared after an earthquake (Mw 7.9) occurred in this area on June 13, 2005.

Olca volcano
Olca (20˚57´S; 68˚30´W; 5,450 m a.s.l.) is a stratovolcano that forms part of a 20-km-long, EW-oriented volcanic chain that includes Paruma and Michincha volcanoes [de Silva andFrancis 1991, González-Ferrán 1995]; it presents strong evidence of past glacial activity along the southern sector (Figure 3).Olca lies above Upper Miocene-Pliocene dacitic and andesitic volcanic products that form the Ujina ignimbrite [Vergara and Thomas 1984].Andeciticdacitic lava flows extend up to 7 km north from the active Olca crater [de Silva and Francis 1991, González-Ferrán 1995, Wörner et al. 2000].Unconfirmed historical eruptions are suspected to have occurred in 1865-1867[González-Ferrán 1995], while more recently, anomalous seismic activity hat was most likely related to intense degassing from the volcano summit was recorded in November 1989and March 1990[Global Volcanism Program 1990].The present fumarolic activity of Olca began within the last ~60 years [Casertano 1963, Global Volcanism Program 1990, Clavero et al. 2006].At present, the main fumarolic field is restricted to the dome within the summit crater (Figure 3) [Aguilera 2008].

Alitar volcano
Alítar (23˚09´S, 67˚38´W; 5,346 m a.s.l.) is a stratovolcano that has a 500-m-wide, 50-m-deep maar (broad, low relief ) crater at the base of its SW flank [González-Ferrán 1995] (Figure 5).Alítar has no documented historical activity.The basement of the Alitar volcanic structure consists of Upper Miocene rhyolitic La Pacana ignimbrite and Pliocene dacitic Atana ignimbrite [Ramirez and Gardeweg 1982], and in comparison, the volcanic edifice consists of Pliocene-Pleistocene andesitic-to-dacitic lavas [Ramirez and Gardeweg 1982].Fumarolic activity has been recognized in the northern sector of the Alitar maar and in a small area 400 m NW of the maar (Figure 5).Its current thermal activity also includes six pools that discharge thermal water and gas along a small, NS-oriented creek that is located ~200 m west of the maar (Figure 5) [Aguilera 2008].

Sampling and analytical methods
We collected gas samples using a 1-m-long titanium tube (Ø, 2.5 cm) that was inserted into the fumarolic vent and connected through quartz-glass dewar tubes to a preevacuated, 60-mL, glass thorion-tapped flask filled with 20 mL 4 N NaOH and a 0.15 M Cd(OH) 2 suspension [Giggenbach and Goguel 1989, Montegrossi et al. 2001, Vaselli et al. 2006].We collected aliquots of fumarolic vapor condensates in-line with the main sampling apparatus using an ice-cooled glass tube.The condensate was used to determine HF concentrations by ion chromatography (Metrohm Basic761) and the d 18 O and dD isotopic ratios.
We determined the inorganic gases (N 2 , O 2 , H 2 , He, Ar and Ne) in the flask headspace using a Shimadzu 15A gas chromatograph equipped with a 10-m-long 5A molecularsieve column and a thermal conductivity detector.We analyzed hydrocarbons, including CH 4 , using a Shimadzu 14A gas chromatograph equipped with a 10-m-long stainless steel column (z, 2 mm) packed with Chromosorb PAW 80/100 mesh coated with 23% SP 1700 and a flame ionization detector.Carbon monoxide was analyzed by gas chromatography-flame ionization detection after conversion of CO to CH 4 at 400 ˚C using a Shimadzu MTN-1 methanizer.
To determine the 13 C/ 12 C isotopic ratio of CO 2 (hereafter expressed as d 13 C-CO2 ‰ V-PDB), we added ~5 mL anhydrous phosphoric acid to 2 mL of the soda solution under vacuum, and allowed the CO 2 emanated to equilibrate at 25 ±0.1 ˚C in a thermal bath overnight.The extracted CO 2 was then purified using liquid N 2 and N 2 -trichloroethylene cryogenic traps, and analyzed using a Finningan Delta S mass spectrometer.We used internal (Carrara and San Vincenzo marbles) and international (NBS18, limestone, and NBS19, carbonate) standards to estimate the external precision.The analytical error and the reproducibility were ±0.05‰ and ±0.1‰, respectively.
We determined the 18 O/ 16 O and 2 H/ 1 H isotopic ratios of the water vapor (steam) condensates (hereafter expressed as d 18 O-H 2 O and dD-H 2 O ‰ V-SMOW, respectively) using a Finningan Delta Plus XL mass spectrometer at the Geokarst Engineering Laboratory (Trieste, Italy).Oxygen isotopes were analyzed according to the method described by Epstein and Mayeda [1953].The hydrogen isotopic measurements were conducted on H 2 generated by the reaction of 10 µL water with metallic zinc at 500 ˚C, according to the analytical procedure described by Coleman et al. [1982].We used V-SMOW and SLAP as analytical standards and AR-1 as an internal standard.The analytical error was ±0.1‰ and ±0.1‰, for d 18 O and dD values, respectively.
The 3 He/ 4 He (hereafter expressed as R/R a ratios, where R is the 3 He/ 4 He measured ratio and R a is the 3 He/ 4 He ratio of the air: 1.39 ×10 −6 ) [Mamyrin and Tolstikhin 1984] and 40 Ar/ 36 Ar ratios were determined at the University of Rochester Rare Gas Facility using a VG 5400 rare-gas mass spectrometer fitted with a Faraday cup (resolution, 200) and a Johnston electron multiplier (resolution, 600) for sequential analyses of the 4 He (F-cup) and 3 He (multiplier) by the method described by Poreda and Farley [1992].For noble gas isotope analysis, the gas samples were purified in a highvacuum line constructed of stainless steel and Corning-1724 glass to minimize helium diffusion.Water vapor and CO 2 were cryogenically trapped at −90 ˚C and −195 ˚C, respectively.Nitrogen and O 2 were removed with a Zr-Al alloy chemical getter (SAES-ST707), whereas Ar and Ne were adsorbed onto activated charcoal at −196 ˚C and at −233 ˚C, respectively.SAES-ST-101 getters reduced the HD+ background to ~1,000 ions/s.We corrected the measured 3 He/ 4 He ratios for the addition of air (monitoring 4 He/ 22 Ne), by assuming that the fumarolic Ne is low and of atmospheric origin [Craig andLupton 1976, Sano andWakita 1988].Analytical error for the R/R a determination was ≤0.3%.After completing the He isotopic analysis, we analyzed the 40 Ar/ 36 Ar ratios with a VG 5400 mass spectrometer.Sensitivity for the Ar concentrations was about 4 × 10 −4 Amps/torr on the Faraday cup.Precision for the 40 Ar/ 36 Ar ratios averaged at 0.2%.
Irruputuncu volcano has fumarole outlet temperatures between 83 ˚C and 240 ˚C, while the other three volcanoes have temperatures between 82 ˚C and 91 ˚C.All of the emission temperatures are above or similar to the boiling point for water at 5,000 m a.s.l.(~83 ˚C).The temperatures of the Alitar bubbling pools (Al1, Al2 and Al3; Table 1) are between 54 ˚C and 57 ˚C.The fumarolic gas composition is dominated by water vapor (X gas from 1.45% to 7.24%), while the bubbling gases show X gas values between 91.9% and 96.4%.The dry gas fractions are mainly composed of CO 2 (from 764 to 983 mmol/mol).
At the Irruputuncu and Putana volcanoes, the fumarolic gas species are dominated by relatively high SO 2 and H 2 S concentrations (up to 170 and 52 mmol/mol, respectively), followed in order of abundance by HCl, N 2 , H 2 and HF (up to 14, 14, 4.0 and 1.8 mmol/mol, respectively) and then O 2 , CO, and CH 4 (up to 0.41, 0.039, and 0.052 mmol/mol, respectively).The trace concentrations of Ar, He and Ne are lower than 0.039, 0.007 and 0.000021 mmol/mol, respectively.Hydrocarbons are all present at trace levels, with the sum of the seven most-abundant species <0.0002 mmol/mol (C 2 H 6 , C 2 H 4 , C 3 H 8 , C 3 H 6 , C 4 H 4 O, C 6 H 6 and C 4 H 4 S; Table 2).
The Olca and Alitar gas compositions vary significantly from the fumaroles of the Irruputuncu and Putana volcanoes.The main differences include: 1) significantly lower concentrations of H 2 , CO and acidic gases (at least one order of magnitude), with the exception of H 2 S (up to 26 mmol/mol); 2) significantly higher concentrations of hydrocarbons (including CH 4 ) and atmospheric-related (i.e.O 2 , Ar, Ne) compounds (up to one order of magnitude).The N 2 and He concentrations are between 1.8 to 6.1 mmol/mol and 0.004 to 0.017 mmol/mol, respectively, and they do not show significant differences from the Irruputuncu and Putana gases.Of note, the CO 2 , CH 4 , H 2 , CO, He and light hydrocarbon concentrations in the dry gas fraction of the Alitar bubbling gases are similar to the other "dry" fumarolic gases at Alitar (Tables 1 and 2), whereas the acidic gases (i.e.HF and SO 2 ) are significantly lower, or even non-detectable.

Isotopic compositions of steam ( d 18 O and D) and gases
(R/R a , 40 Ar/ 36 Ar and d 13 C-CO 2 ) Each of the four volcanoes can be clearly distinguished according to the stable isotopic composition (d 18 O and dD) of their condensed steam emissions (Table 3).The Irruputuncu gas stable isotopic composition is relatively heavy, and it ranges from d 18 O: +6.8‰ to +10.4‰ and dD: −35.1‰ to −44.1‰ V-SMOW, respectively.From Irruputuncu, the stable isotopic composition of the condensed steam gases are progressively lighter in composition to Putana (d 18 O and dD: 1.7‰ and −57.2‰V-SMOW, respectively), Alitar (d 18 O: −1.6‰ to −1.0‰ and dD: −73.4‰ to −67.8‰ V-SMOW, respectively) and Olca (d 18 O: −8.4‰ and −8.0‰ and dD: −97.5‰ to −95.6‰ V-SMOW, respectively).Conversely, our study found no significant differences in the d 13 C-CO 2 values for each of the volcanic systems, which ranged from −8.54‰ to −6.53‰ V-PDB.
Fumarolic isotopic gas measurements including: noble gases (i.e. 3 He/ 4 He [R/R A ] and 40 Ar/ 36 Ar) and CO2 (d 13 C-CO 2 ); selected gases are given in Table 3.The Irruputuncu fumaroles have the highest 3 He/ 4 He (R/R A ) value (7.27), followed by Putana (Pu1 and Pu4 gases with 3 He/ 4 He [R/R A ] of 6.15 and 7.14, respectively).The Olca (Ol1 3 He/ 4 He [R/R A ] of 6.11) helium isotopic composition is similar to that of Pu1, while Alitar is characterized by a significantly lower R/R a value (4.80).The argon isotopic composition ( 40 Ar/ 36 Ar) for the Irruputuncu, Olca and Putana gases range between 312 and 396 and contain a moderately radiogenic component, whereas that of Alitar (Al5, 297) is similar to that of air (295.5)[Ozima and Podosek 2002], and reflects significant influence from air or air-saturated water in this system.

Origin of fumarolic fluids
Volcanic fluids are discharged from volcanic systems along subducting convergent plate boundaries that typically originate from one of the following sources (and may mix to varying extents amongst these sources): 1) mantle or mantle-wedge magmatic degassing; 2) fluid-rock interactions at the various temperatures along the arc; and 3) air or air-saturated water (dissolved in permeating meteoric   20 Ne, CO 2 / 3 He and CH 4 / 3 He ratios of fumaroles from the Irruputuncu (Ir), Olca (Ol), Putana (Pu) and Alitar (Al) volcanoes.L, S and M are the limestone, organic sediments and mantle end-members (%), as defined by Sano and Marty [1995].
We determined the relative contributions of each of these sources for this region in the CAVZ.We used the Ar-CO 2 -(SO 2 + H 2 S) ternary diagram (Figure 6) to demonstrate that the fumarolic gas composition is accounted for by a mixture between hydrothermal and magmatic sources, with minimal atmospheric (air or air-saturated water) contributions.The Irruputuncu gases are characterized by C/St ratios <10 (where St is the sum of SO 2 + H 2 S), which is typical of magmatic gases from andesitic volcanoes [Symonds et al. 1994] and is similar to those of other active volcanoes located along subducting convergent plate boundaries, i.e.Avachinsky (Kamchatka), Galeras Cumbal and Galeras (Colombia) [Fischer et al. 1997, Taran et al. 1997, Lewicki et al. 2000].Interestingly, this low C/St ratio is less than those observed at medium-high-temperature (between 150 ˚C and 385 ˚C) fumaroles from other CAVZ systems, including Lascar (La-HT) [Tassi et al. 2009], a frequently eruptive stratovolcano located few dozen kilometers south of Putana [Gardeweg and Medina 1994, Aguilera et al. 2003, Aguilera et al. 2006, Clavero et al. 2006].Compared to Irruputuncu, the Olca and Alitar fumaroles show significantly higher C/St ratios (ranging from 33 to 65), that we suggest is related to a relatively high contribution from hydrothermal fluids, while the gases at Putana have intermediate C/St ratios (between 15 and 27).Several fumaroles included in this study have anonymously high Ar contributions than described above (and cannot be explained by a simple two-component magmatic/hydrothermal mixture) including: 1) the Pu1 fumarole, which we attribute to air contamination because of the concomitant increase in O 2 concentrations (0.41 mmol/mol; Table 1); and 2) the three bubbling gas pools from Alitar volcano, which were also characterized by relatively low H 2 S concentrations (low St), which we attribute to liquid-gas interactions at shallow depths and increased interactions with shallow air-rich groundwater.These types of secondary processes might, at least partially, explain the high C/St ratios of the Olca and Alitar fumaroles, as the solubility in water of the sulfur gas compounds is significantly higher than that of CO 2 .To better define this system, we constructed the (HCl + H 2 )-SO 2 -CH 4 ternary diagram (Figure 7), assuming: (i) gas-water-rock interactions produce H 2 and HCl at relatively high temperatures [Martini 1993]; (ii) SO 2 is a direct proxy for magmatic contributions [Giggenbach 1980[Giggenbach , 1996]]; and (iii) CH 4 in volcanic environments originates from thermal decomposition of organic material and/or abiogenic processes (i.e.Sabatier reaction involving CO 2 and H 2 that occurs under hydrothermal conditions) [Schoell 1988, Whiticar and Suess 1990, Giggenbach 1997, Darling 1998, Horita and Berndt 1999, Sherwood-Lollar et al. 2002, Fiebig et al. 2004, Fiebig et al. 2007].Our findings here are in agreement with those indicated by the C/St ratios (Figure 6), and they suggest that the fumarolic gases from these volcanoes can be defined according to only magmatic and hydrothermal contributions.We find that the Irruputuncu fumarolic gases, and to a lesser extent those of Putana, have a distinct magmatic signature that is more prominent than the highly active Lascar (La-HT) fumaroles.Conversely, the Alitar and Olca gases show a significant enrichment in CH 4 , and are thus strongly affected by fluid contributions from a hydrothermal source (Figure 7).[Tassi et al. 2009], Avachinski (Kamchatka) [Taran et al. 1997], Galeras (Colombia) [Fischer et al. 1997] and Cumbal (Colombia) [Lewicki et al. 2000] gases are also shown.By evaluating the stable isotopic compositions (d 18 O and dD) of the steam condensates (Figure 8), we show that all four of the Chilean volcanoes included in this study, and including Lascar, plot along a binary mixing line between the northern Chile local meteoric water line [Aravena 1995, Chaffaut et al. 1998, Aravena et al. 1999] and the suggested "andesitic water magmatic" field [Taran et al. 1989, Giggenbach 1992b].In relative terms, the trend shown in Figure 8 supports our findings for the magmatic, hydrothermal, and air-saturated water contributions shown in Figures 6 and 7. We find that an andesite water magmatic end-member dominates the Irruputuncu gases (Figure 8), which agrees with our findings for the gas composition (Figures 6 and 7) and the high helium isotopic composition (R/R A , 7.27) at Irruputuncu (Table 3).Helium isotopic compositions between 7-9 Ra indicate the presence of a strong contribution from upper mantle or mantle-wedge fluids without the significant addition of crustal fluids [Craig and Lupton 1976, Javoy et al. 1982, Poreda et al. 1988, Rollinson 1993].Conversely, the d 18 O and dD isotopic signatures of the Putana, Alitar and Olca fumaroles can be regarded as derived by mixing processes between magmatic fluids and meteoric water in comparable amounts (Figure 8).
Aside from water vapor, CO 2 is the most prominent gas present in all of the samples included in the present study.We determined the source of CO 2 using carbon isotopic composition and the relative proportion of CO 2 to 3 He, a conservative, primordial mantle-gas tracer.All of the samples included in this study have d 13 C-CO 2 values in the range of mid-ocean ridge basalt (MORB) gases (−9‰ to −4‰ V-PDB) [Marty and Jambon 1987, Javoy and Pineau 1991, Sano and Marty 1995], and do not vary systematically in the volcanoes included in the present study.The CO 2 / 3 He ratio is a sensitive tracer that can distinguish the removal ("scrubbing") and/or addition of CO 2 in fumarole gases, and specifically the addition of CO 2 to magmatic sources [e.g.Tedesco et al. 2010].We find that the CO 2 / 3 He ratios from all the investigated volcanic systems range between 1.73 × 10 10 and 6.05 × 10 10 .This range is consistent with typical of fumarole gases from an arc setting and are approximately one order of magnitude higher than those of MORB and Ocean Island basalt gases [Des Marais and Moore 1984, Marty and Jambon 1987, Marty and Tostikhin 1998], probably due to the addition of CO 2 from carbonate-rich sediments released from the subducting slab [Giggenbach and Poreda 1993, Sano and Williams 1996, Patino et al. 2000, Snyder et al. 2001].
In subduction zones, CO 2 primarily originates from one of three sources: 1) sedimentary organic (S); 2) limestone (L); and 3) MORB (a representative for upper mantle gas) (M), with minimal crustal-derived contributions [e.g.Hilton et al. 2002].We evaluated the relative contributions of each source in the CAVZ using the following equations [Sano and Marty 1995]: (1) (2) (3) where subscripts obs, MORB, Lim and Sed refer to the sample, MORB (a proxy for upper mantle contributions), limestone and sediment, respectively.We assume that the end members have the following compositions: The results for each contribution are presented in Table 3.We find that the calculated relative amounts for each of the three potential carbon sources (i.e.M, S and L) result in a maximum contribution of 8.7% magmatic (M) CO 2 with a ratio of limestone/sedimentary carbon (L/S) ~3 at Irruputuncu, Olca and Putana.These data are consistent with previous findings for Lascar volcano [Tassi et al. 2009], as well as for other arc volcanoes in Salvador and Costa Rica [Shaw et al. 2003, de Leeuw et al. 2007].Alitar gas contains FLUID GEOCHEMISTRY OF CHILEAN VOLCANOES The "andesitic water" field [Taran et al. 1989, Giggenbach 1992b], and the local meteoric water line (LMWL: d 2 H = 8.15d 18 O + 15.3) [Chaffaut et al. 1998] are also shown.Symbols as in legend to Figure 6.For comparison, the isotopic composition of Lascar gases (La-HT) [Tassi et al. 2009] is also shown.the lowest magmatic contribution (M: 3.7%), which is more consistent with geothermal systems in northern Chile than these CAVZ volcanic systems [Tassi et al. 2010a].These results suggest that the majority (>90%) of CO 2 at these volcanoes in the CAVZ is derived from nonmagmatic (non-mantle) sources.Instead, the CO 2 is primarily derived from the devolatilization of subducting carbonates and to a lesser extent from thermogenic interactions with carbon in rocks and sediments, which will be discussed further below.
To further constrain the origins of the gases from potential thermogenic interactions of hot fluids with rocks/sediments and slab devolatilization, we evaluated the relative abundances of N 2 , Ar and He (Figure 9).The N 2 /Ar ratio can be used to distinguish the relative contributions to a gas mixture from air, air-saturated fluids, thermogenic interactions with subducting and surficial rocks/sediments, and mantle gases [e.g.Snyder et al. 2001Snyder et al. , 2003].The gas composition of the volcanoes in the present study is marked by high N 2 /Ar ratios (up to 1,509) relative to air (82).These high N 2 /Ar ratios are typical of subduction-zone andesitic gas emissions [Matsuo et al. 1978, Giggenbach 1992b], where N 2 is extracted from the subducting slab and slab sediments as the temperatures increase during the subduction process [Snyder et al. 2003].Note that in the present study, the N 2 /Ar ratio identifies two distinct trends (Irruputuncu and Putana vs. Olca and Alitar) for the source of the thermogenic N 2 , including: 1) subducting-slab-generated N 2 ; and 2) "crustal-like" N 2 that suggests the addition of N 2 from crustal (near surface) sources, and most likely from the interactions of heated groundwater with crustal volcanic sequences closer to the surface.The addition of crustal 4 He is supported by the decrease in R/R air for Olca and Alitar (6.11 and 4.80), as compared to Irruputuncu (7.27) and Putana (6.15-7.14)(Table 3).
The Irruputuncu and Putana gases are similar to those of arc-type andesitic volcanism and they contain relatively low He and Ar, while the Olca and Alitar fumaroles show significant He enrichment (specifically 4 He enrichment, as inferred from the lower R/R a ; Table 3), with minimal addition of Ar in most samples.Note that the Alitar bubbling pools and the one Putana sample (Pu1) suspected of mixing with air, as discussed above, each plot along a mixing line between other gases from the respective volcano and the air composition (Figure 9).Interestingly, these samples do not appear to mix with air-saturated water, indicating minimal interactions with noble gases dissolved in groundwater for these samples.
The origins of light hydrocarbons are commonly investigated on the basis of the relative abundance of the C 1 -C 3 alkanes in fumarolic gases [Oremland et al. 1987, Whiticar and Suess 1990, Kiyosu et al. 1992, Darling 1998].Gases produced by decay of organic matter at temperatures >150 ˚C are characterized by CH 4 /(C 2 H 6 + C 3 H 8 ) ratios <100, whereas higher ratios are expected when processes related to bacteria activity at low temperatures are responsible for hydrocarbon production.Based on these assumption, the light hydrocarbons present at Irruputuncu, Putana, Olca and Alitar would be characterized as originating from a bacteriarelated source, because the CH 4 /(C 2 H 6 + C 3 H 8 ) ratios range between 158 and 603.However, we suggest that a bacteriarelated origin is not likely.We do not anticipate that hydrocarbon production can be attributed to bacterial activity in fluids so strongly related to magmatic degassing, and specifically those with discharge temperatures higher than the temperatures at which bacteria are anticipated to survive (~120 ˚C).Instead, we prefer an alternative process in which the high temperatures and highly oxidizing conditions created by the presence of water-rich, high-temperature, magmatic-related fluids results in a thermal cracking process that preferentially destroys long-chain hydrocarbons, altering the CH 4 /(C 2 H 6 + C 3 H 8 ) ratio.This hypothesis can explain: (i) the lack of C 3+ species, with the possible exception of benzene, furane and thiophene, which are stable at relatively high temperatures [Montegrossi et al. 2003, Capaccioni et al. 2005, Tassi et al. 2010b]; (ii) the CH 4 "excess" with respect to the C 2 H 6 and C 3 H 8 concentrations; and (iii) the relatively low CH 4 / 3 He ratios (1.55 × 10 5 to 1.85 × 10 7 ) that are lower than those observed for gases produced by thermogenic processes in an organic-rich hydrothermal environment (up to 1 × 10 12 ) [Poreda et al. 1988].Relatively high CH 4 concentrations in hydrothermal fluids from submarine ultramafic-hosted systems [Charlou et al. 2002, Proskurowski et al. 2008, Konn et al. 2009] and in springs discharging from ophiolites [Abrajano et al. 1990] were ascribed to CH 4 production through catalytic hydrogenation of carbon oxygenated species, the so-called Fischer-Tropsch-type reaction.Fiebig et al. [2007,2009] suggested that the occurrence of a Fischer-Tropsch-type reaction is responsible of excess methane in continental hydrothermal emissions.Isotopic data (d 13 C and dD) of CH 4 and light alkanes, which were not available for the present study, could provide useful information about the origin of these species [Sherwood Lollar et al. 2006, Fu et al. 2007, Taran et al. 2007, McCollom et al. 2010, Taran et al. 2010].However, on the basis of the reported chemical evidence, we can argue that the concentrations of CH 4 are independent of light alkanes, with the former being regulated by the CO 2 -CH 4 chemical equilibrium instead of the decay of organic matter.

Geothermometry
We can estimate the thermodynamic conditions of a deep reservoir source for fumarolic fluids by assuming equilibrium conditions for reactions involving specific gas compounds.The H 2 /H 2 O log-ratio (R H ) is considered to be the most suitable parameter for describing the redox state of most natural fluids [Giggenbach 1987].In hydrothermal fluids, redox conditions are regulated by the FeO-FeO 1.5 pair in rocks, which produces a R H of −2.8 almost independent of temperature.Alternatively H 2 concentrations and the redox potential in gases related to magmatic degassing are controlled by the SO 2 -H 2 S gas buffer [Giggenbach 1997, and references therein].Equilibrium constraints in the H 2 O-CO 2 -CO-H 2 system are provided by the following pressure-independent reaction: (4) Reaction ( 4) is strongly controlled by a temperaturedependent equilibrium constant, according to [Giggenbach 1996]: (5) In the log(H 2 /H 2 O) versus log(CO/CO 2 ) diagram (Figure 10), the Irruputuncu and Putana gases correspond to equilibrium temperatures ranging between 420 ˚C and 580 ˚C, whereas the Olca and Alitar fumaroles appear to equilibrate at 340 ˚C to 410 ˚C.It is worth noting that in these systems, the SO 2 -H 2 S pair is the dominant redox buffer, which is consistent with the presence of significant concentrations of these two gas compounds in the fumarolic gases (Table 1).The H 2 O-CO-CO 2 -H 2 geothermometer indicates ureliable low temperatures (<100 2 C) for the Alitar bubbling gases that are significanlty affected by H 2 O condensation and CO dissolution within the bubbling pools.The gas species in the H 2 O-CO 2 -CH 4 -H 2 system are controlled by the Sabatier reaction, as follows: (6) Assuming that log f H 2 O = 4.9 − 1820/T [Giggenbach 1987], the temperature dependence of the equilibrium constant of reaction (6) can be expressed by the following equation: (7) According to Equations ( 6) and ( 7), the thermodynamic conditions of the equilibria among the carbon gas species (i.e.CO 2 , CH 4 and CO) can be shown by the log(CH 4 /CO 2 ) versus log(CO/CO 2 ) diagram [Giggenbach 1987], where compositions that have equilibria controlled by the FeO-FeO 1.5 and SO 2 -H 2 S redox pairs are also reported (Figure 11).The CO 2 -CH 4 -CO equilibrium temperatures of the Irruputuncu and Putana gases are 320 ˚C to 390 ˚C, whereas those of the Olca and Alitar fumaroles are <180 ˚C.The decoupling between the fluid temperatures, calculated by using the H 2 O-CO 2 -CO-H 2 and CO 2 -CH 4 -CO geothermometers, is probably caused by the relatively low kinetics of reaction (6) with respect to that of reaction (4).Indeed, the gas samples in Figure 11 plot far from both the FeO-FeO 1.5 and SO 2 -H 2 S equilibrium lines.The CO 2 -CH 4 pair is quenched at relatively high temperatures (between 370 ˚C and 650 ˚C, considering the FeO-FeO 1.5 redox buffer), whereas H 2 and CO tend to reequilibrate at lower temperatures as fumarolic fluids rise toward the surface.
Because of non-equilibrium (quenching) conditions, we additionally investigate the thermodynamic conditions that characterize deep fluid source regions on the basis of the FLUID GEOCHEMISTRY OF CHILEAN VOLCANOES  (8) The temperature dependence of the equilibrium constant for reaction (8) is [Capaccioni et al. 2004]: (9) De-hydrogenation of C 3 H 8 to produce C 3 H 6 is as follows: (10) The temperature dependence of the equilibrium constant for reaction (10) is [Capaccioni et al. 2004]: (11) We calculated equilibrium temperatures from Equations ( 9) and (11) (T C2-C2 and T C3-C3 , respectively) using the H 2 concentrations measured.These calculations were not performed for the Alitar bubbling gases, as the H 2 concentrations of these samples are strongly affected by liquid-gas interactions at low temperatures (Figure 10).As shown in Table 2, both of the sets of calculated temperatures indicate that the Irruputuncu and Putana gases equilibrate at 478 ˚C to 536 ˚C, whereas those of the Olca and Alitar gases range from 280 ˚C to 400 ˚C, in agreement with the temperature estimations in the H 2 O-CO 2 -CO-H 2 system.

Concluding remarks
The chemical and isotopic (R/R a , d 18 O-H 2 O, dD-H 2 O and d 13 C-CO 2 ) compositions of gases emitted from the Irruputuncu, Putana, Olca and Alitar volcanoes suggest a mixture between magmatic, hydrothermal and atmosphericrelated fluids, as typically occurs in volcanic systems along convergent/subduction plate boundaries.We suggest that magmatic degassing is the dominating fluid source at Irruputuncu and Putana.The Olca and Alitar gases show high-temperature gas compounds (i.e.presence of SO 2 ), and significant contributions from hydrothermal components.The R/R a and the CO 2 / 3 He ratios of the four investigated volcanic systems are consistent with those for the CAVZ and subduction arc systems worldwide.All four systems emit gases with magmatic carbon components below 10%, while the lower proportion of magmatic carbon in Alitar fluids supports our assertions that fluids discharging at this volcano are fed by a well-developed hydrothermal system.Gas geothermometry in the H 2 O-CO 2 -CO-H 2 , CO 2 -CH 4 and C 2 -C 3 alkenes-alkanes systems, consistently indicates that the Irruputuncu and Putana fluids are quenched at relatively high temperatures (>500 ˚C), which suggests hot magmaticrelated gases rapidly interact with cooler hydrothermal fluids as they ascend to the surface.However, only a very small amount of interaction between magmatic fluids and shallow aquifers is feasible based on our data.Conversely, significantly lower equilibrium temperatures (<400 ˚C) of the fluids from the Olca and Alitar volcanoes are likely to be attained within hydrothermal systems overlying the magmatic source.
On the whole, the Irruputuncu, Putana, Olca and Alitar volcanoes are variably affected by mixing between deepseated (magmatic) and shallower (hydrothermal) sources, both which suggest that each of these edifices remain active.Consequently, although these volcanoes are located in remote areas of northern Chile, we suggest that periodic (e.g. on a yearly basis) sampling of the fumarolic gas discharges is required to monitor the evolution of the most relevant geochemical and isotopic parameters.Periodic monitoring will allow researchers to develop an accurate geological hazard assessment for these volcanoes and the gain an understanding of the volcanic hazard potential that can result from their activity.

Figure 1 .
Figure 1.Schematic map of the central volcanic zone of the Andes in northern Chile, with the locations of the Irruputuncu, Olca, Putana and Alitar volcanoes.

Figure 2 .
Figure 2. Aerial photograph (a) and schematic map (b) of the summit crater of Irruputuncu volcano, showing the locations of the fumarolic fields and sampling sites.

Figure 3 .
Figure 3. Aerial photograph (a) and schematic map (b) of the summit crater of Olca volcano, showing the locations of the fumarolic fields and sampling sites.

Figure 4 .
Figure 4. Photograph (a) and schematic map (b) of the summit crater of Putana volcano, showing the locations of the fumarolic fields and sampling sites.

Figure 5 .
Figure 5. Aerial photograph (a) and schematic map (b) of the summit crater of Alitar volcano, showing the locations of the fumarolic fields and sampling sites.

Figure 10 .
Figure 10.Log(H 2 /H 2 O) versus log(CO/CO 2 ) diagram.Solid lines refer to equilibria controlled by the FeO-FeO 1.5 and the SO 2 -H 2 S redox pairs.Dashed lines are the calculated isotherms for the simultaneous equilibrium of the H 2 -H 2 O and the CO-CO 2 geothermometers.Symbols as in legend to Figure 6.

Figure 11 .
Figure 11.Log(CH 4 /CO 2 ) versus log(CO/CO 2 ) diagram.Solid lines refer to equilibria controlled by the FeO-FeO 1.5 and the SO 2 -H 2 S redox pairs.Dashed lines are the calculated isotherms for the simultaneous equilibrium of the H 2 -H 2 O and the CO-CO 2 geothermometers.Symbols as in legend to Figure 6.