Noble Gas and Carbon Isotope Systematics at the Seemingly Inactive Ciomadul Volcano (Eastern-Central Europe, Romania): Evidence for Volcanic Degassing

2.1 Ciomadul Volcanic Dome Field

Ciomadul volcano is located at the southeastern edge of the Carpathian-Pannonian Region, at the southern end of the Călimani-Gurghiu-Harghita volcanic chain (Szakács et al., 1993; Szakács & Seghedi, 1995; Pécskay et al., 2006; Figure 1). It is part of a postcollisional volcanic belt, which comprises a series of andesitic to dacitic volcanoes, developed parallel with the Carpathian orogeny. The volcanism occurred well after the continent-continent collision between the Tisza-Dacia microplate and the western margin of the Eurasian plate (Cloetingh et al., 2004; Csontos et al., 1992; Mațenco et al., 2007; Mațenco & Bertotti, 2000; Seghedi et al., 2004, 2005, 2011). Ciomadul is part of a lava dome field and this central volcanic complex involves 8–14 km³ of high-K dacitic lavas (Karátson & Timár, 2005; Molnár et al., 2019; Szakács et al., 2015). The volcano developed on the Early Cretaceous clastic flysch sedimentary unit of the Eastern Carpathians that forms several nappes. It consists of binary alternation of sandstones, calcareous sandstones, limestones, and clays/marls from the Sinaia Formation of the Ceahlau nappe and the Bodoc flysch (Băncilă, 1958; Grasu et al., 1996; Ianovici & Rădulescu, 1968; Nicolaescu, 1973). The flysch unit has a thickness up to 2,500 m.

The Ciomadul volcanic complex is made up by amalgamation of several lava domes truncated by two explosion craters called Mohos and Saint Anna (Szakács et al., 2015). This central volcano is surrounded by further isolated lava domes (Baba Laposa, Haramul Mic, Dealul Mare, Büdös-Puturosul and Bálványos; Molnár et al., 2018, Figure 2). Volcanism at the Ciomadul volcanic dome field started around 1 Ma, while the most voluminous Ciomadul volcanic structure has developed over the last approximately 160 kyr (Molnár et al., 2018, 2019). During the first volcanic stage, the intermittent lava dome extrusions were separated by relatively long dormant periods even exceeding 100 kyr. The second volcanic stage was characterized by initial lava dome effusion and then, after approximately 40 kyr of quiescence, a more explosive volcanic activity occurred (from 57 to 30 ka, Harangi et al., 2010; Harangi, Lukács, et al., 2015; Karátson et al., 2016; Molnár et al., 2018, 2019; Moriya et al., 1995, 1996; Vinkler et al., 2007). This stage involved lava-dome collapse events, vulcanian and subplinian to plinian explosive eruptions (Harangi, Lukács, et al., 2015; Karátson et al., 2016; Vinkler et al., 2007). The eruptive products are relatively homogeneous K-rich dacites (Molnár et al., 2018, 2019; Szakács et al., 1993; Szakács & Seghedi, 1987; Vinkler et al., 2007). Petrogenetic and thermobarometric studies on amphiboles as well as combined U-Th/He and U/Th zircon dating suggest the presence of a long-lasting (up to 350 kyr) crystal mush body in the crust. This appears to be mostly at relatively low-temperature just above the solidus (700–750 °C) and is periodically partly remobilized by injections of fresh basaltic magmas that could rapidly trigger volcanic eruptions (Harangi, Lukács, et al., 2015; Harangi, Novák, et al., 2015; Kiss et al., 2014).

The Ciomadul volcano is located near (~50 km) the Vrancea seismic region (Ismail-Zadeh et al., 2012; Wenzel et al., 1999) located at the arc bend between the Eastern and the Southern Carpathians. Frequently occurring earthquakes have deep hypocentres (70–170 km) delineating a narrow, vertical region. This is consistent with a high-velocity seismic anomaly interpreted as a cold lithosphere slab descending slowly into the asthenospheric mantle (Wortel & Spakman, 2000). Further crustal and subcrustal earthquakes (M < 4) occur occasionally around the Perșani basalt volcanic field and the Ciomadul volcano (Popa et al., 2012). The seismic tomographic model indicates a vertically extended low-velocity anomaly beneath Ciomadul. This can be interpreted as transcrustal magma storage with an upper melt-dominated magma chamber (Popa et al., 2012). The seismic tomographic model is supported by the result of combined petrologic and magnetotelluric studies that demonstrated the existence of a low-resistivity anomaly and the depth of 5–20 km beneath the volcanic centers of Ciomadul, inferred to be a melt-bearing crystal mush (Harangi, Novák, et al., 2015). In addition, a deeper low-resistivity anomaly was also detected at a depth of 30–40 km, possibly related to a deeper magma accumulation zone at the crust-mantle boundary.

Another Pleistocene monogenetic basalt volcanic field is approximately 40 km from the Ciomadul, at the southeastern part of the Carpathian-Pannonian Region (Figure 1), at the boundary between the Perşani Mountains and the Transylvanian basin (Downes et al., 1995; Harangi et al., 2013; Seghedi et al., 2016; Seghedi & Szakács, 1994). Basaltic volcanism occurred here between 1.14 Ma and 683 ka (Panaiotu et al., 2004, 2013) and formed several volcanic centers accompanied by maars, scoriacones, and lava flows. The erupted basaltic magma carried significant amount of ultramafic xenoliths from the lithospheric mantle (peridotites and amphibole pyroxenites) revealing the nature of the uppermost mantle of this region (Falus et al., 2008; Vaselli et al., 1995).

2.2 Gas Emissions and Mineral Water Springs at Ciomadul Volcanic Area

Gas emanations in the form of bubbling pools and low-temperature (T ~ 8–10 °C) dry mofettes are characteristic of the Ciomadul volcano. CO₂-bubbling peat bogs can be also found, mainly at the northeastern (Buffogó peat bog) and southern parts of the Puturosul Mountains (Zsombor-Valley, Jánosi et al., 2011). The minimum total CO₂ flux was estimated to be 8.7 × 10³ t/year (Kis et al., 2017). The aquifers of this area are represented by CO₂-rich sparkling mineral water, with temperature up to 22.5 °C (Berszán et al., 2009; Italiano et al., 2017; Jánosi et al., 2011).

3.1 Chemical and Isotopic Composition of Gases

The chemical composition of the samples from the first campaign was analyzed with a Portable Varian CP4900 Micro Gas Chromatograph (GC). This Micro GC is configured for the analysis of He, Ne, H₂, O₂, and N₂ by means of a molecular sieve 5A (20 m unheated) column and CO₂, CH₄, and H₂S by means of a PoraPlot (PPQ 10 m heated) column. The instrument is equipped with a microthermal conductivity detector responding to the difference in thermal conductivity between the carrier gas (argon) and the sample composition. The detection limit is 1 ppm, operating range is from 1 ppm to 100% level concentrations, and repeatability is <0.5% relative standard deviation in peak area at constant temperature and pressure.

For the analysis of δ¹³C_CO2, carbon dioxide was cryogenically removed from the gas samples by liquid nitrogen and measured by Thermo Finnigan Delta ^PLUS XP isotope ratio mass spectrometer. Isotope ratios are given in the standard δ notation in permil (‰) versus VPDB (Vienna Pee Dee Belemnite). Errors for δ¹³C are 0.5‰.

Noble gas isotopic ratios (³He/⁴He and ⁴He/²⁰Ne) were measured from each gas sample that was inserted into the preparation line of the VG5400 noble gas mass spectrometer. The argon and the other chemically active gases (N₂, CO₂, etc.) were separated in a cryogenic cold system consisting of two cold traps and were adsorbed in an empty trap at 25 K. The Ne and He were adsorbed in a charcoal trap at 10 K. He was desorbed at 42 K and neon at 90 K and measured sequentially. The measurement procedure was calibrated with known air aliquots. The analytical uncertainties are 1% for He concentrations and 5% for Ne concentrations and 2.5% for ³He/⁴He. ³He/⁴He ratio is expressed as R/R_a (being R_a the He isotope ratio of air and equal to 1.384·10⁻⁶. He isotopic composition was corrected for the atmospheric He contamination (R/R_ac) considering the ⁴He/²⁰Ne ratio; R/R_ac = [R/R_a * (X − 1)]/(X − 1), where X is the air-normalized ⁴He/²⁰Ne ratio taken as 0.318 (Sano & Wakita, 1985).

For the samples of the second analysis campaign, the chemical and isotopic composition of He-Ne and ¹³C_CO2 was determined in the laboratories of INGV-Palermo.

The concentrations of CO₂, CH₄, O₂, and N₂ were analyzed using an Agilent 7890B gas chromatograph with Ar as carrier and equipped with a 4-m Carbosieve S II and PoraPlot–U columns. A thermal conductivity detector was used to measure the concentrations of He, O₂, N₂, and CO₂ and a flame ionization detector for CO and CH₄. The analytical errors were 10% for He and 5% for O₂, N₂, CO, CH₄, and CO₂. More details on the analytical procedures used during this analysis are given in Liotta and Martelli (2012).

The carbon isotopic composition of CO₂ (δ¹³C_CO2) was determined using a Thermo Delta XP isotope ratio mass spectrometer equipped with a Thermo Scientific™ TRACE™ Ultra GC, and a 30-m Q-plot column (i.e., of 0.32 mm). The resulting δ¹³C_CO2 values are expressed in permil with respect to the international VPDB standard and analytical uncertainties are ±0.15‰. The method for the δ¹³C determination of total dissolved carbon is based on chemical and physical CO₂ stripping (Capasso et al., 2005). Isotopic ratios were measured using a Finnigan Delta Plus Mass Spectrometer. The results are expressed in permil of the international VPDB standard. The standard deviations of the ¹³C/¹²C ratios are ±0.2‰.

³He, ⁴He, and ²⁰Ne and the ⁴He/²⁰Ne ratios were determined by separately inserting He and Ne into a split flight tube mass spectrometer (GVI-Helix SFT, for He analysis) and into a multicollector mass spectrometer (Thermo-Helix MC plus, for Ne analysis), after standard purification procedures (Rizzo et al., 2015). The analytical reproducibility was <0.1% for ⁴He and ²⁰Ne. However, the estimation of He and Ne concentration agrees within 10% uncertainty respect to GC measurements. In this study, the time from sampling to analysis was lower than 2 weeks and results are fully reliable. The analytical error for He and Ne concentration measurements is generally below 0.3%.

3.2 Noble Gas Isotope Data for the Perşani Clinopyroxene

The chosen xenolith is a fresh spinel lherzolite with about 12% clinopyroxene content. Here we performed new noble gas analyses. The preparation, single-step crushing and analysis of fluid inclusions was the same as described by Correale et al. (2012) and references therein. Helium (³He and ⁴He) isotopes were measured separately by two different split-flight-tube mass spectrometers (Helix SFT-Thermo). The analytical uncertainty of the determination of the TGC (thermal conductivity detector) and the He and Ne abundances was ~10%. Error in the ³He/⁴He ratios is reported at the 1σ level.

4.1 Chemical and Isotopic Composition of Gases

The CO₂ concentration in the collected gases ranges from 6.40% to 98.36%. Besides CO₂, H₂S (2.7 × 10⁻⁴ to 1.72 × 10⁻¹%), He (5.91 × 10⁻⁵ to 1.66 × 10⁻²%), Ne (6.39 × 10⁻⁷ to 5.80x10⁻³%), H₂ (1 × 10⁻⁵ to 2.3 × 10⁻¹%) CO (6 × 10⁻⁵ to 5 × 10⁻⁴%), CH₄ (3.5 × 10⁻² to 1.69%), N₂ (1.5 × 10⁻¹ to 74.5%), and O₂ (2 × 10⁻³ to 18.99) are present in the gas samples. The ternary diagram CO₂/50–N₂–O₂ (Figure 3) shows a progressive enrichment in N₂ and O₂ of the samples, indicating a variable amount of air.

The ³He/⁴He ratios range between 0.77 and 3.10 R_a and the ⁴He/²⁰Ne ratios from 0.36 to 1,700, which show that some of the collected gases are affected by air contamination (Table 3). The ³He/⁴He ratios after corrections for the air contamination (R/R_ac) are up to 3.25. The δ¹³C_CO2 ranges between −1.40‰ and −17.2‰ versus VPDB (Table 3).

4.2 Noble Gas Ratios of Fluid Inclusions From Persani Clinopyroxenes

Helium content in the fluid inclusions in clinopyroxenes ranged between 4.06 × 10^–12 and 3.81 × 10^–12 mol/g and Ne content between 2 × 10^–15 and 2.74 × 10^–15 mol/g, so the He/Ne ratios ranged between 1,390 and 2,030. The He isotopic signature in fluid inclusions was 5.95 Ra ± 0.01 (Table 4).

5.1 Crustal Assimilation Versus Mantle Metasomatism

Helium comes from three different sources (mantle, crust, and air), which can be readily distinguished based on their characteristic isotopic ratios (Sano & Wakita, 1985). Helium isotopes are useful tracers for detecting deep fluids and their possible origin (crust, mantle, or atmosphere; Ozima & Podosek, 2002). It has been demonstrated that in the case of quiescent volcanoes, the active degassing of deep volatiles can occur for a long time after the last volcanic activity (Caracausi et al., 2009, 2015; Carapezza & Tarchini, 2007; Tassi et al., 2013).

The last eruption in Ciomadul occurred 30 ka (Harangi et al., 2010; Harangi, Lukács, et al., 2015; Molnár et al., 2019), yet there is an intense CO₂ degassing with a minimum flux of 8.7 × 10³ t/year (Kis et al., 2017), which is comparable to other dormant volcanic areas such as Panarea (1.72 × 10⁴ t/year) and Roccamonfina (7.48 × 10³ t/year) from Italy or Jefferson (7.92 × 10³ t/year) from the United States.

In addition, previous investigations (Althaus et al., 2000; Vaselli et al., 2002; Túri et al., 2016) highlighted the outgassing of mantle-derived volatiles at Ciomadul volcano. He isotopic ratios in the fluids collected in this study are up to 3.1 R_a similar to those obtained from previous studies (Figure 4 and Table 3). These values are higher than those obtained from the surrounding areas such as in the Carpathian Foredeep and the Transylvanian Basin where He isotopic ratios are between 0.02 and 0.03 R_a (Baciu et al., 2017; Italiano et al., 2017; Vaselli et al., 2002, Figure 4). These latter values are typical of crustal fluids dominated by ⁴He produced by decay of U and Th (e.g., Ozima & Podosek, 2002). The higher R_a values measured at Ciomadul could imply a higher contribution of magmatic He. Nevertheless, the 3.1 R_a value is significantly lower than the MORB and subcontinental lithospheric mantle (SCLM) value (Sano & Marty, 1995) requiring addition of radiogenic ⁴He that decreased the pristine isotopic signature.

The mantle xenoliths of the Perşani volcanic field (approximately 40 km from the Ciomadul area) could provide the He isotopic signature of the lithospheric mantle beneath the region. The He isotopic ratios in fluid inclusions of the Persani clinopyroxenes are 5.95 ± 0.01 (Table 4), and these are lower than those of of previous measurements, from 6.5 to 7.3 R_a, obtained by Althaust et al. (1998) but consistent with the values of the SCLM (R/R_a = 6.1 ± 0.9 R_a, Gautheron & Moreira, 2002). The continental crust (R/R_a = 0.02, Ozima & Podosek, 2002) and atmosphere (R/R_a = 1) have distinct isotopic values, and ⁴He/²⁰Ne can be used to infer how mixing between the three possible end members can support the He isotopic signature of the fluids that outgass in the Ciomadul region (Figure 4). Most Ciomadul samples indicate a possible trend between air and a magmatic source, where the He ratio of the magmatic end member (3.1 R_a) is lower than that of the ECLM and the Perşani clinopyroxene. This is also supported by the trend line in the ³He–CO₂–⁴He ternary diagram (Figure 5), where the Ciomadul samples are along a trend showing variable amounts of CO₂ and R/R_ac values between 2 and 3. This trend reflects the dominance of radiogenic He in the fluids outgassing from the Ciomadul volcano. We have now to assess the possible processes that can add the radiogenic He component to the mantle component.

Such a relatively low He isotope ratio of the magma source is not uncommon in volcanic arc settings (e.g., Hilton et al., 1992; Allard et al., 1997; Martelli et al., 2004) and can be due to several processes involving the addition of radiogenically produced ⁴He, such as magma aging, crustal assimilation, mixing between mantle, and crustal-derived fluids (Kennedy & van Soest, 2006; Torgersen et al., 1995). Unfortunately, there are no undifferentiated mantle-derived mafic rocks in the region of the Ciomadul volcano, so we cannot investigate the He isotope composition of the mantle directly below the volcano. In Ciomadul, only high-K dacitic volcanic products are found (Mason et al., 1996; Molnár et al., 2018, 2019; Vinkler et al., 2007), although occurrence of high-Mg minerals such as olivine and clinopyroxene in the dacites suggest involvement of primitive mafic magmas in the magma evolution of Ciomadul (Kiss et al., 2014; Vinkler et al., 2007).

Magma aging and crustal assimilation are two mechanisms that could account for the addition of the radiogenic He component to the mantle-derived melts. Both these processes have been invoked to explain low He isotopic ratios (<MORB and SCLM) in different volcanic regions, worldwide, such as Aeolian Island, Italy (Mandarano et al., 2016), and Iceland (Condomines et al., 1983). The magma-aging mechanism considers an addition of ⁴He by radiogenic decay in the magma itself. In constrast, crustal assimilation furnishes ⁴He by interaction between magma and the whole rock. First, we investigated the likelihood that the magma aging model can interpret the low He isotopic signature in the fluids that outgas at Ciomadul volcano.

The ³He/⁴He ratio of the fluid inclusions of the Persani clinoproxene (5.95 R_a ± 0.01) can be assumed to represent the mantle end member value beneath of the region. Thus, the primary magmas of Ciomadul could be also characterized by such isotope ratio. The Ciomadul dacites have U and Th concentrations of 3 and 15 ppm, respectively (Molnár et al., 2018, 2019; Vinkler et al., 2007). Using these data, the magma-aging model calculation yield ³He/⁴He ratio around 4.65 R_a after 30 kyr (Figure 6). Thus, this process alone cannot be responsible for the low He (approximately 3.1 R_a) isotopic signature of the Ciomadul fluids. Furthermore, if we assume the U (1.5 ppm) and Th (5.5 ppm) contents of the Persani basalts (Harangi et al., 2013), the magma-aging model is still not a viable process to provide the required ⁴He addition and generate the low ³He/⁴He for Ciomadul gases.

The relatively low He isotopic ratio can also be explained by high-level crustal assimilation (e.g., van Soest et al., 2002), which has to also be evaluated. Assuming the U and Th amount of the typical upper crust, 2.7 and 10.5 ppm, respectively (Rudnick & Gao, 2003), and an age of 5 Ma, 3% of crustal assimilation could be sufficient to achieve the observed low He isotopic ratios. The Sr–Nd–O isotope compositions of the erupted magmas sensitively reflect such a process. Mason et al. (1996) published isotopic data for three samples of the Ciomadul volcanic system. They have distinct isotopic features compared to the calc-alkaline volcanic suite of the Calimani-Gurghiu-Harghita chain. Although the Sr–Nd isotopic data could suggest an assimilation and fractional crystallization process with 10–35% assimilation of flysch sediment, such a high crustal contamination is not feasible, based on the fairly low δ¹⁸O values (6.3–7.1‰) of the phenocrysts from the dacites (Mason et al., 1996). Instead, they suggested that these isotopic characteristics could also be explained by source contamination from subduction-related fluids. In fact, the bulk-rock composition of the Ciomadul dacites has unique characteristics with high Sr, Ba (both showing typically >1,000 ppm), and high K compositions and low concentrations of heavy rare Earth elements (Molnár et al., 2018, 2019; Seghedi et al., 1987; Vinkler et al., 2007). Furthermore, the high-Mg pargasitic amphiboles thought to have derived from the less differentiated magmas have also relatively high Ba content (Kiss et al., 2014). Thus, these peculiar compositional characters can be due to the nature of the magma source rather than magma differentiation processes. The elevated K, Sr, and Ba contents of the assumed mantle source of the Ciomadul primary magmas can be due to metasomatism, and this is in contrast what the peridotite xenoliths from the Persani volcanic field show (Vaselli et al., 1995). In fact, the He signature of the outgassed volatiles at Ciomadul resembles the values in fluids from other subduction-related volcanic systems (i.e., Italy, and Indonesia; Hilton et al., 1992; Martelli et al., 2004), where the mantle source regions seem to be contaminated by crustal material that added radiogenic ⁴He and decreased the pristine He isotopic signature (Hilton et al., 2002).

Such a small-scale spatial heterogenity of the lithospheric mantle beneath this area can be explained by the closer location of Ciomadul to the collision front, where subduction is expected to have occurred during the Miocene up to around 11 Ma (Cloetingh et al., 2004; Mațenco et al., 2007; Royden et al., 1982; Seghedi et al., 2011). Such a scenario is not unique; Martelli et al. (2004) suggested that the relatively low He isotopic ratio in the volcanic rocks of Central Italy can be explained by magma source features (i.e., contribution of radiogenic He from metasomatic, subduction-related fluids and ingrowth of ⁴He in the lithospheric mantle). We note that the ⁸⁷Sr/⁸⁶Sr isotopic ratio of the Ciomadul dacites and the highest ³He/⁴He isotopic values of the emitted gases plot into the same trend (Figure 5 in Martelli et al., 2004) what the Central Italian volcanic areas form.

In summary, considering the petrology of the Ciomadul volcanic products, the relatively low He isotope magmatic end member of the Ciomadul gases can be interpreted as due to magma-source characteristics, where the radiogenic He was added via subduction-related fluids and increased radioactive ingrowth following the metasomatism. However, a mixing between mantle-derived fluids with and SCLM He isotopic signature and ⁴He-rich crustal fluids coming from shallow crustal layers should still be further explored as a possible process responsible of the low He isotopic ratios in the Ciomadul fluids. This likelihood will be discussed in the next section.

5.2 Sources and Origin of Carbon Dioxide

The carbon isotopic composition of CO₂ (δ¹³C_CO2) from the studied fluids range between –1.40‰ and –4.61‰ versus VPDB, consistent with previous measurements in the area (–2.77‰ to –4.70‰; Frunzeti, 2013; Sarbu et al., 2018; Vaselli et al., 2002). In the Pannonian Basin (Central Europe), the carbon isotopic composition of CO₂ gases shows values in a narrow range between –3‰ and –7‰ with an average value of –5‰ VPDB based on hundreds of measurements (Bräuer et al., 2016; Cornides, 1993; Palcsu et al., 2014; Sherwood-Lollar et al., 1997). These values are consistent with a mantle origin. In contrast, crustal-derived CO₂ is characterized by a δ¹³C of about –25‰ in case of biogenic sedimentary source and around 0‰ considering thermometamorphism of limestone (Sano & Marty, 1995 and references therein). The Ciomadul gases overlap the range of mantle composition, even if some samples have more positive values that cannot be explained by the addition of a crustal biogenic component (Table 3 and Figures 7 and 8). To constrain the origin of CO₂ in the fluids emitted by the Ciomadul volcano, we used the relationship between the elemental ratio CO₂/³He and the isotopic signature δ¹³C_CO2 (Sano & Marty, 1995; Figure 7).

The CO₂/³He ratios of the Ciomadul gases are higher than 2 × 10⁹, the expected mantle ratio (Marty & Jambon, 1987) and which suggests an addition of a crustal component. It is interesting that these ratios fall into the same trend as shown by volcanic and fumarolic gases measured at volcanic arcs, worldwide (Mason et al., 2017; Figures 8a and 8b). Almost all the Ciomadul samples fall close the mixing line between a mantle component and a limestone end member suggesting that mixing of the two sources could be the main process that controls the CO₂–³He systematics in these fluids. In contrast, CO₂ fluids in the Transylvanian Basin, (Baciu et al., 2007, 2017) west of the volcano have distinct character and fall closer to the mantle-organic sediment mixing line. Rayleigh-type fractionation due to gas exsolution from water is not a plausible process to produce the carbon isotopic signature and the CO₂/³He of the studied fluids (Figure 7; Holland & Gilfillan, 2013; Roulleau et al., 2015). However, the ¹³C_CO2 values of most of the samples fall in the narrow range of −2‰ and −5‰, which is a typical signature for mantle-derived carbon. We obtain the same trend in the He isotopic ratios (R/R_a) versus ¹³C_CO2 (VPDB) plot (Figures 8a and 8b), where the Ciomadul samples clearly approach the mantle end member and overlap the isotopic values of many other volcanic systems related to subduction areas. Remarkably the Ciomadul samples show similarities in He–C isotopic composition with active and dormant volcanic regions (e.g., Italy and Indonesia).

The involvement of carbonatic component can be explained by mixing with fluids derived from thermometamorphic decomposition of carbonates in the flysch sedimentary pile or by mantle source contamination via subducted carbonatic material. The mantle source of the Ciomadul magmas is considered to be a metasomatic lithospheric mantle based on the compositional features of the dacitic rocks. The relatively low He isotopic ratio can be due to these source characteristics, whereas metasomatism was the result of slab-derived fluids during the Miocene subduction along the Eastern Carpathians followed by ingrowth of radiogenic He by radioactive decay. The Sr–Nd–O isotope data of the volcanic rocks do not support significant upper-level crustal contamination but rather crustal component addition to the source region via slab-derived fluid metasomatism (Mason et al., 1996). The combination of He and C isotopic data suggests that this crustal component consisted of decomposed subducted carbonate material as suggested also for the volcanic rocks in Italy, although addition of fluids from carbonate decomposition at shallow crustal level cannot be unambiguously excluded.

5.3 Relationship With the Deep Magmatic System

Dormant volcanoes pose a particular hazard to society since there is much less awareness about a possible eruption event. However, the scientific community is giving increased attention to these volcanoes and the surrounding areas that are generally characterized by intense gas emissions (Burton et al., 2013, and references therein). Recent investigations highlighted the presence of an active plumbing system even below volcanoes which last erupted >10 kyr (e.g., Colli Albani, Italy; Trasatti et al., 2018; Uturuncu, Bolivia; Sparks et al., 2008; Comeau et al., 2015; Tatun, Taiwan; Konstantinou et al., 2007; Lin & Pu, 2016). Harangi, Novák, et al. (2015) suggested the term PAMS volcano, that is, volcano with potentially active magma storage for these long-dormant volcanoes, which have clear implication for a subvolcanic melt-bearing magma plumbing system. Ciomadul belongs to this category, since there are a number of observations suggesting that a melt-bearing magma body could still exist beneath it (Harangi, Novák, et al., 2015; Popa et al., 2012; Szakács & Seghedi, 2013). The isotopic composition of the emitted gases coupled to the high localized heat flow in the area of the Ciomadul vocano gives additional support to this interpretation.

This involves the similarities in the isotope composition of CO₂ and He of the gases emitted at the Ciomadul with those found in other active and dormant volcanic arc systems worldwide and their proposed high magmatic component. Furthermore, the Ciomadul volcanic system is characterized by relatively high CO₂ gas fluxes (Kis et al., 2017). This is consistent with the presence of a still-degassing magma below the Ciomadul system as inferred by geophysical investigations that recognized a low-resistivity and low-velocity anomaly in the crust, below the volcano (Harangi, Novák, et al., 2015; Popa et al., 2012) as well as petrologic observations suggesting the involvement of a mafic magma in the petrogenesis of the erupted dacite (Kiss et al., 2014). The measurements of U–Th and U–Pb spot ages on zircons suggest a long-standing magma storage that could go back as far as about 350 ka (Harangi, Lukács, et al., 2015; Lukács et al., 2018). Molnár et al. (2018, 2019) presented a detailed eruption chronology for the Ciomadul lava dome field involving the Ciomadul volcanic complex and emphasized that volcanic activity could be renewed even after long (>100 kyr) repose times. Several tens-of-kiloyears quiescence periods between the active phases have also been pointed out also during the evolution of the Ciomadul volcanic complex (Harangi, Lukács, et al., 2015; Molnár et al., 2019). However, the zircon U–Th and U–Pb ages suggest that crystallization was ongoing also during the long quiescence periods, that is, there was an active magma storage beneath the apparently inactive volcano This suggests a long-standing felsic upper-crustal crystal mush system underlain by a mafic hot zone in the lower crust, as has already been suggested by petrologic interpretations (Kiss et al., 2014). The diverse amphibole compositions in the dacites are consistent with a polybaric magma evolution, that is, with transcrustal magma storage (Cashman et al., 2017; Sparks & Cashman, 2017) comprising ephemeral melt-dominated bodies, that is, magma chambers at various depths. In addition, fluid-gas accumulation zones can also have developed within this magma storage (Christopher et al., 2015; Sparks & Cashman, 2017). Thus, a possible source of the CO₂ gases could be these fluid entrapment zones within the crystal mush during quiescent period. However, gas emission is more common around the Ciomadul volcanic complex and significantly lower within the volcano itself (Kis et al., 2017). Allard et al. (1991), and Edmonds (2008) pointed out that stronger degassing around the volcanic edifice is not uncommon in volcanic regions. An alternative source of the CO₂ gases could be mafic magma residing at deeper level, possibly at the lower crust. Indeed, the occurrence of high-Mg minerals, such as olivine, clinopyroxene, and orthopyroxene in the dacites (Kiss et al., 2014; Vinkler et al., 2007) suggests that mafic magma also played an important role in the magma evolution. Harangi, Novák, et al. (2015) detected a lower crustal low resistivity anomaly, which might represent the mafic magma accumulation. Thus, we propose that most of the CO₂ gases could come directly from the presumed mafic-magma accumulation zone at the lower crust through fractures (Kis et al., 2017), whereas only limited amount of gases is derived from the mushy magma storage.

Vaselli et al. (2002) already suggested that the emitted gases in Southern Harghita could have a magmatic component. Based on our new measurements, we support this interpretation, particularly in the area of Ciomadul volcano. Assuming that a deep-seated mafic magma body can be the main source of the CO₂ gases and considering that it is characterized by relatively low ³He/⁴He isotope signature (3.1 R_a) inherited by the mantle source region, we can use this value to calculate the relative magmatic component of the emitted gases (Sano & Wakita, 1985). If no interaction with crustal fluids occurred, the magmatic component in the gases could exceed even the 80%. Remarkably, we obtained such high values for the areas having a larger diffusive CO₂ flux. This high magmatic He content of the gases is not unique and resembles what Trasatti et al. (2018) proposed for Colli Albani volcanic complex, another long-dormant volcanic field, where they assumed more than 80% mantle-derived component in the emitted CO₂ gases. However, the magmatic component can be lower, if interaction between the ascending gases with crustal gases occurred at shallow crustal depth, a possibility what we cannot test at this stage but requires further studies.