Kolumbo submarine volcano (Greece): An active window into the Aegean subduction system

Abstract

Submarine volcanism represents ~80% of the volcanic activity on Earth and is an important source of mantle-derived gases. These gases are of basic importance for the comprehension of mantle characteristics in areas where subaerial volcanism is missing or strongly modified by the presence of crustal/atmospheric components. Though, the study of submarine volcanism remains a challenge due to their hazardousness and sea-depth. Here, we report ³ He/⁴ He measurements in CO₂ -dominated gases discharged at 500 m below sea level from the high-temperature (~220 °C) hydrothermal system of the Kolumbo submarine volcano (Greece), located 7 km northeast off Santorini Island in the central part of the Hellenic Volcanic Arc (HVA). We highlight that the mantle below Kolumbo and Santorini has a ³ He/⁴ He signature of at least 7.0 Ra (being Ra the ³ He/⁴ He ratio of atmospheric He equal to 1.39×10^-6 ), 3 Ra units higher than actually known for gases-rocks from Santorini. This ratio is also the highest measured across the HVA and is indicative of the direct degassing of a Mid-Ocean-Ridge-Basalts (MORB)-like mantle through lithospheric faults. We finally highlight that the degassing of high-temperature fluids with a MORB-like ³ He/⁴ He ratio corroborates a vigorous outgassing of mantle-derived volatiles with potential hazard at the Kolumbo submarine volcano.

Kolumbo submarine volcano (Greece): An active window into the Aegean subduction system

R

P

Andrea Luca Rizzo, Antonio Caracausi, Valrie Chavagnac, Paraskevi Nomikou, Paraskevi N. Polymenakou, Manolis Mandalakis, Georgios Kotoulas, Antonios Magoulas, Alain Castillo & Danai Lampridou

dominated

has a

derived volatiles with potential hazard at the Kolumbo submarine volcano.

Most of the volcanic activity worldwide occurs in the oceans1, with newly formed volcanoes at mid-ocean ridges, hot spots and volcanic island arcs. Any submarine volcanic eruption leads to a major regional disruption of the environment by modifying the chemical composition of adjacent seawaters and releasing a large amount of gases with consequences on the associated deep-sea ecosystems and atmospheric chemistry24. This degassing can be even more catastrophic when high volumes of volcanic gases reach the atmosphere and become lethal for humans, as occurred in 1650 A.D. at Kolumbo volcano (South Aegean Sea; Fig.1a,b). During this submarine eruption ~70 people were killed at the nearby island of Santorini5. Thus, the study of submarine volcanic degas-sing is of great importance to better constrain the characteristics and evolution of geodynamic settings showing elevated volcanic hazard with considerable impact on the daily life of local populations.

The geochemical studies carried out on submarine uids have been mainly focused on mid-ocean ridge hydrothermal systems68, whereas only a few of them investigated subduction zone environments912. The HVA in the South Aegean Sea (East Mediterranean) results from the subduction of the African plate below the European one13,14 (Fig.1a). In the last decades, several studies were dedicated to better constrain the mantle characteristics below the HVA and understand the dynamics of this complex geodynamic setting but a general understanding has not been reached thus far1522. The recent discovery of the high-temperature hydrothermal eld at the Kolumbo submarine volcano23 (Fig.2ac), located 7km northeast o Santorini Island (Fig.1b), has opened up a new opportunity to deepen our knowledge of Aegean mantle features. Indeed, Kolumbo is considered to be the most active volcanic system of the region at present time15,2326. Its magmatic activity is evidenced by numerous hydrothermal sites venting vigorously and continuously up to ~220C warm uids27 and CO2-rich gases at the crater seaoor (Fig.2ac)4. This ongoing intense degassing contrasts radically with the low-temperature phenomena typically observed in the Santorini caldera, next to the Kameni islands4,23,28 (Fig.1b).

Istituto Nazionale di Geofisica e Vulcanologia, Sezione di Palermo, Italy. CNRS, Gosciences Environnement Department of Geology and Geoenvironment, National and

SCIENTIFIC REPORTS

1

www.nature.com/scientificreports/

Figure 1. (a) Simplied map of the present day geodynamic structure of the HVA, showing the modern volcanic arc developed behind the Hellenic trench, the PeloponneseCrete island arc and the Cretan back-arc basin. The study area is located in the center of the HVA (modied from ref. 28). (b) Swath bathymetric map of Christianna-Santorini-Kolumbo (CSK) volcanic elds (modied from ref. 26). The location of the NE-SW prole of Fig.3 is also reported (purple dotted line), as well as CSK tectonic alignment.

The Santorini-Amorgos Basin surrounding Kolumbo system, divides the HVA into a seismic-active eastern zone and a relatively quiet western area29. The main seismic hypocentres beneath Kolumbo are located at depths of 69km24, indicating a high level of magmatic activity below this volcano. Accordingly, this range of depth corresponds to the location of the main magma chamber24,25 (Fig.3).

The Kolumbo hydrothermal system is poorly studied4,30, and little is known about the origin of uids escaping from its seaoor (Fig.2c) and the potential link with those emitted from Santorini subaerial vents, as well as other active volcanoes across the HVA. Here, we report on the 3He/4He ratios of gases discharging from seven dierent hydrothermal chimneys of the Kolumbo submarine volcano in order to constrain the He isotopic characteristics of the mantle beneath this key sector of the HVA. We focus on the 3He/4He ratio because this is the most powerful geochemical tracers to dene the origin of volatiles released from the solid earth and the magmatic/mantle features31,32. Indeed, 3He has a primordial origin and is preferentially degassed from the Earths interior. Because He is highly mobile, chemically inert, physically stable and nonbiogenic, the interaction of this noble gas on its movement toward the surface is minimized, and its isotopic composition is not aected by subsequent chemical reactions8. It also gives unambiguous signals of magma rising up in volcanic plumbing systems3336.

Geodynamic setting

The HVA is a 500km-long curving chain of volcanoes, Pliocene to modern in age, extending from the Methana peninsular zone on the Greek mainland, through the islands of Milos-Antimilos, Santorini, Kos-Nisyros, towards the western peninsular zone in Turkey. The HVA results from the extended and elongated subduction of the African plate beneath the Eurasian continent (Fig.1a). The plate horizontal velocities measured by GPS in the Eastern Mediterranean highlight the presence of a clockwise rotation of a broad region relative to Eurasia at rates in the range of 2030 mm/yr19,22. This rotation is combined with the continuing southward rollback of the Hellenic subduction around a pole located in Albania19,22. The complex regime of movements causes a NWSE extension in the Western AnatoliaAegean system, which in turn leads to the thinning of the Aegean crust from roughly 50 to 25km19. Based on seismic tomography models, the Aegean subduction is characterized by a single slab, more than 1,500km-long, which extends down to the lower mantle22. These models also suggest that the slab does not extend far eastward and that possible ruptures are present below Nisyros volcanic system and western Turkey, where modern alkaline magmatism is present15,22. In addition, P-wave velocities along the HVA indicate the presence of a low velocity zone just above the downgoing slab, which can be interpreted as the occurrence of asthenospheric mantle18,37. As a result, the interface between the slab and mantle wedge is inuenced by primary magmas generated in the underlying asthenosphere18.

Within the HVA, Santorini volcano is one of the most active and it is very famous worldwide because of its large explosive eruption that occurred in Bronze Age, which caused signicant impacts to human populations in the eastern Mediterranean38. In the Santorini area, volcanism extends northeastward into the submarine environment evidenced by a series of small craters and cones aligned along Christianna-Santorini-Kolumbo (CSK) tectonic line28, which likely controls the pathways of hydrothermal circulation within the region (Fig.1b). This is a vertical northeast trending zone, up to 3040km large and 45km deep29. The CSK line created an area of structural weakness, characterized by increased seismic activity25, enabling the upward migration of uids at Kolumbo. The focus of this study is dedicated to a better understanding of the composition of these uids in order to further constrain the mantle features within the HVA.

SCIENTIFIC REPORTS

2

www.nature.com/scientificreports/

Figure 2. (a) 3D Bathymetric map of Kolumbo submarine volcano, showing the shape of the crater in whose bottom are located the hydrothermal chimneys; (b) High resolution swath bathymetric map of Kolumbo crater bottom (Fig.1c) with the location of the sampled hydrothermal chimneys as labeled in Table1. (c) Active hydrothermal vent discharging both gases (>99% CO2) and high-temperature uids collected in this work.

Results

The chemical composition of gases collected from each of the seven chimneys (Fig.2b,c) consists of almost pure CO2 (Table1), in line with the ndings of previous work at the same area4. This feature is typical of submarine volcanic emissions and it has also been observed either in hot spot systems as Loihi39 (Hawaii), or in subduction-related environments like Panarea (Aeolian islands, Italy9), oshore NE Taiwan40, along the Mariana and Tonga-Kermadec arcs11. The O2 (<2.1%) and N2 (<8.5%) concentration measured in Kolumbo gases suggests a contamination by atmospheric air, which, nevertheless, has a minor inuence on helium isotopic composition, as evidenced by the 4He/20Ne ratio (9270; Table1) being 30 to 250 times higher than that of ambient air (4He/20Ne = 0.318). At the bottom of Kolumbo crater, seawater exhibits acidic conditions (pH = 5), as a consequence of CO2 release and dissolution3,4,30. This dissolution, the extent of which varies depending on the actual ux of gas coming out from each vent, is responsible for the variable helium concentrations (1541 part per million) measured in our samples (Table1).

The 3He/4He ratios of seven gas samples are virtually identical to one another, irrespective of the vent being sampled and the extent of CO2 dissolution, and vary in the narrow range of 7.07.1 Ra (Fig.4a and Table1). These values are slightly higher than the single previous measurement of 6.8 Ra reported for Kolumbo4. Because the

3He/4He ratio does not suer any signicant fractionation during gas-water interaction, it can be used to assess the origin of emitted uids.

Discussion

The 3He/4He ratios measured at Kolumbo (7.07.1 Ra) are within the typical range of values found in arc volcanoes worldwide and correspond to a MORB-like 3He/4He signature

SCIENTIFIC REPORTS

3

www.nature.com/scientificreports/

Figure 3. Cross section along ~40km NESW prole (reported in Fig.1b) moving from the Kolumbo area towards Santorini Island showing a sketch of the distinct plumbing systems beneath the two volcanoes. The depth of the magmatic chambers (dimensions not to scale) and the main volcanological, seismic and petrologic features characterizing Santorini and Kolumbo volcanic systems are also reported15,2325,28,29,55,57.

Sample Depth (m) Latitude Longitude 4He (ppm) 20Ne (ppm) O2 (%) N2 (%) CO2 (%) R/Ra 4He/20Ne Rc/Ra Error (+/)

A2 497 3631.5700N 2529.2110E 23 0.09 0.17 1.1 98.0 7.01 270.3 7.02 0.13 V2 498 3631.5700N 2529.2050E 15 1.69 2.09 8.5 88.3 6.84 9.1 7.05 0.017 V3 498 3631.5843N 2529.2046E 26 0.10 0.10 0.9 98.4 7.04 248.3 7.05 0.011 V4 498 3631.5846N 2529.2378E 22 0.84 1.08 5.0 91.5 6.95 29.1 7.05 0.012 V5 500 3631.5790N 2529.2060E 25 0.51 0.67 3.1 95.4 7.05 48.0 7.10 0.016 V6 498 3631.5824N 2529.2012E 19 0.08 0.00 0.6 97.4 7.00 234.4 7.01 0.018 V7 498 3631.5580N 2529.2160E 41 0.29 0.17 1.3 96.7 6.98 139.4 7.00 0.015

NA007-081* 502 3631.5735N 2529.2034E 24 0.03 0.01 0.3 99.4 6.84 990.0 6.84

Table 1. Analyses of major gaseous components, helium and neon in gases from Kolumbo hydrothermal vents. *Sample from ref. 4.

of 81 Ra8,41 (Fig.4a). This nding indicates that magma degassing beneath Kolumbo and feeding its hydrothermal system has a 3He/4He representative of the primary source. This signature is surprisingly higher (more than 3 Ra units) than that measured in gases and rocks (uid inclusions) of the adjacent (<7 km) Santorini island34

(Fig.4a). Indeed, gases collected between 1988 and 2001 from both Nea and Palea Kameni islets were in the range of 3.23.8 Ra17,4244, while during the 20112012 seismic, geodetic and geochemical unrest these ratios increased from 3.6 to 4.0 Ra (Ra=0.4)34. Also, the 3He/4He ratios measured in some mac enclaves, hosted by dacitic lavas recently erupted at Nea Kameni, yielded a range of 3.13.6 Ra34 (Fig.4a). All these data imply that the magma residing at shallow depths into the Santorini plumbing system has a distinct and fairly consistent He-isotope signature ranging between 3 and 4 Ra.

Substantial spatial variations of the 3He/4He ratios over a 10-km distance have been recently observed at the mid oceanic ridge (7.510.2 Ra45). However, to the best of our knowledge, it is rare that two distinct and active volcanic systems located only a few km apart from one another display at the same time such a wide variation in

3He/4He ratio. This nding emphasizes the importance of investigating uids being discharged by submarine volcanic clusters with sufficient spatial coverage and resolution, if precious insights about the mantle are to be gained around the planet. In light of this, while the high 3He/4He ratios at Kolumbo can be considered representative of the mantle source beneath the submarine volcano, the lower 3He/4He values measured at Santorini lead to two fundamental questions:

(1) Do the 3He/4He ratios measured at Santorini reect a local mantle heterogeneity, despite the very short distance among the two volcanic systems?

(2) Do these ratios reect an addition by radiogenic 4He within the Santorini plumbing system of magma and/or uids originated from a homogeneous mantle?

To decipher between the two hypotheses, we compare the lowest 87Sr/86Sr ratios measured in whole rocks from Santorini (0.703515,46,47) with 3He/4He in gases and uid inclusions of the two volcanic systems (Fig.4b). For completeness, in Fig.4b we report the range of 87Sr/86Sr ratios measured in bulk rocks erupted in each volcanic

SCIENTIFIC REPORTS

4

www.nature.com/scientificreports/

Figure 4. Along-arc variations of (a) 3He/4He4,17,34,4244 and (b) 87Sr/86Sr ratios in the HVA15,16,46,47. Solid black

rectangle in plot 4a reproduces the range of He isotope ratios of Kolumbo gases. In the insert diagram of plot 4b we report a binary mixing between a MORB-like mantle (3He/4He=8.0 Ra; 87Sr/86Sr= 0.7020) and a hypothetical Aegean continental crust (3He/4He=0.03 Ra; 87Sr/86Sr=0.7150)50,51. See text for further details. Based on this mixing, the secondary y axis of plot 4b reports the expected 3He/4He ratios (red coloured) corresponding to the 87Sr/86Sr ratios reported in the le y axis.

system. In magmatic environments, these geochemical tracers are inversely correlated, in the sense that the highest He isotopic ratios correspond to the lowest Sr isotopic ratios, which are measured in the most mac and primitive products, and vice versa. This behaviour has been observed in other magmatic systems worldwide48,49,

and is simplied in the insert diagram of Fig.4b in which we plot a binary mixing line between a MORB-like mantle (3He/4He =8.0 Ra;87Sr/86Sr = 0.7020) and a hypothetical Aegean continental crust (3He/4He =0.03 Ra;

87Sr/86Sr= 0.7150)50,51. We obtain a straight line because for simplicity we assumed k = 1 (being k= (He/Sr)crust/ (He/Sr)mantle), however slight variations of this ratio do not modify our inferences. Based on this mixing, the right y axis of plot 4b reports the expected 3He/4He ratios corresponding to the 87Sr/86Sr ratios reported in the le y axis.

Petrological studies on the HVA indicated that volcanic rocks from Santorini display compositional features typical of subduction-related tectonic settings, with a high variability of trace elements and Sr-Nd-Pb isotopes mainly ascribed to crustal assimilation processes15,46,47. Considering that the most primitive products erupted at Santorini showed 87Sr/86Sr values as low as 0.703515,46,47 (Fig.4b), the genesis of parental magmas can be ascribed to the partial melting of a depleted, MORB-like, mantle wedge15,18,47,52. As shown in the insert plot of Fig.4b, these

87Sr/86Sr values are highly consistent with the 3He/4He ratios measured in gases at Kolumbo (7.07.1 Ra; see also secondary y axis in Fig.4b). Conversely, the 3He/4He ratios measured in gases and rocks of Santorini (3.14.0 Ra) cannot account neither for the very low 87Sr/86Sr measured in basalts, nor for the ratios measured in mac enclaves hosted in dacitic lavas of Nea Kameni (87Sr/86Sr~0.704834). This implies that the wide variability of Sr isotope composition observed among the erupted volcanic at Santorini (87Sr/86Sr from 0.7035 to 0.7062; Fig.4b) is related to magmatic dierentiation which predominantly takes place in a shallow magma chamber plus assimilation of crustal rocks during magma ascent plays an important role in the evolution of Santorini magmas15,46,47

SCIENTIFIC REPORTS

5

www.nature.com/scientificreports/

(Fig.3). Accordingly, several models of assimilation plus fractional crystallization involving Santorini basalts as mac end-members and crustal basement rocks as contaminants have been proposed18,46,47.

The composition of rocks erupted from Kolumbo is much less investigated than in Santorini. Available data indicated that rocks erupted from Kolumbo volcano range from andesites to rhyolites and belong to calcalkaline series, similarly to most of the Santorini rocks15. Pumices are the main products of the last Kolumbo eruption at 1650 A.D. and lie along a magmatic evolution line similar to the suite of silicic pyroclastics from the nearby Santorinivolcano53. Unfortunately, no Sr isotopic data have been published yet for Kolumbo rocks.

Recent volcanological and geochemical investigations suggested the presence of two distinct plumbing systems beneath Santorini and Kolumbo volcanic systems15,24 (Fig.3). These two reservoirs are also characterized by distinct geochemical and mineralogical characteristics, suggesting that primary magmas beneath Santorini are modied by successive processes of crustal assimilation plus fractional crystallization and magma mixing15,18,24.

As already highlighted34, the occurrence of magma contamination by crustal rocks in Santorini plumbing system would eventually lead to the reduction of 3He/4He ratio with regard to parental magmas. On the other hand, magmas feeding Kolumbo would derive directly from the mantle24.

Santorini and Kolumbo volcanic systems lie along the same tectonic alignment that starts from Christianna north-eastwards (CSK28, Fig.1b) and share a common region of lithospheric weakness. This unequivocally indicates that these volcanic systems originated from the same geological structure, but distinct magmatic reservoirs are active in the crust beneath each one of them15,24 (Fig.3). Furthermore, our data argue that the mantle below Kolumbo and Santorini can be considered homogeneous in terms of He isotope ratios (3He/4He 7.0 Ra). The observed He signature is MORB-like, as in most arc volcanoes worldwide in which the subduction of oceanic crust does not aect He isotopic composition of the mantle wedge8,41.

If petrological features argue for primary magma undergoing dierentiation plus crustal assimilation process in the magmatic plumbing systems below Santorini and Kolumbo volcanoes, then why do we observe 3He/4He ratios as high as 7.1 Ra only at Kolumbo? Two possible hypotheses can be proposed:

(1) Evolution and crustal assimilation by magma beneath Kolumbo are very low and do not imply a contribution of 4He-rich rocks and/or uids;

(2) Predominant degassing at Kolumbo occurs directly from the mantle and uxes the plumbing system, overprinting the He isotopic signature of evolved/contaminated magmatic products.

The geological features following a cross section through Santorini and Kolumbo volcanoes indicate a similar crustal thickness (~25 km54; Fig.3) and common basement rocks24,30, discarding the rst hypothesis. Furthermore, local seismological studies carried out in the area revealed an active tectonic regime beneath Kolumbo volcano characterized by the presence of faults24,55, whereas there is no evidence of continuous seismic activity beneath the caldera of Santorini volcano with the exception of unrest periods24,25,29,56. These geological

and geophysical observations are in very good agreement with recent volcanological and marine studies that showed an intense high-temperature hydrothermal activity at Kolumbo, in comparison to the low-level activity of the Santorini caldera (submarine and subaerial uid temperatures <100)23,28,34. As a result, we conclude that magma residing in the shallow plumbing system beneath Santorini could be volumetrically small, strongly degassed in terms of helium, and prone to the crustal-derived 4He contamination by surrounding rocks. On the other hand, the 3He/4He ratios at Kolumbo could be the result of mantle degassing (uxing) in a regional extensive regime characterized by the presence of active lithospheric faults beneath the central part of the HVA (Fig.1b), which would overprint the signature of 3He/4He of evolved/contaminated magma residing at shallow depths, thus allowing to measure mantle ratios (Figs.3 and 4a).

Under this prism, the Kolumbo submarine volcano is conceived to represent a sort of window into the mantle of the central HVA.

The 3He/4He ratios of Kolumbo (up to 7.1 Ra) are also the highest ever measured in gases and rocks belonging to the HVA (Fig.4a). Until now, the highest values were measured in gases emitted at Nisyros (6.2 Ra), located at the eastern side of the arc17. In this case, the presence of crustal assimilation at shallow depths below the island or the addition of radiogenic 4He to a MORB-like mantle wedge were considered unlikely, while the 3He/4He ratio of Nisyros was deemed to be representative of South Aegean mantle and consistent with the European Sub-Continental Lithospheric Mantle (SCLM; 6.1 0.9 Ra). The same study identied a regional trend toward a westward decrease of 3He/4He ratios from Nisyros to Sousaki volcanoes (Fig.4a) and it was attributed to a variable degree of crustal assimilation and/or to dierent magmatic activity in each volcanic system17.

Our current 3He/4He measurements from Kolumbo provide new constrains on these interpretations, because the mantle below the HVA (at least in its central sector, i.e., Kolumbo-Santorini) has an He isotopic signature compatible with a MORB-like source, and the high 3He/4He ratios at Kolumbo rule out the presence of a regional trend. In order to evaluate if the mantle signature beneath Kolumbo may extend up to Nisyros, we used the same approach described in the previous section based on He and Sr isotope signatures in gases and rocks. Petrological studies of the HVA indicate that the most primitive products erupted at Nisyros are characterized by 87Sr/86Sr values as low as 0.703415,46 (Fig.4b). These Sr isotope ratios are perfectly compatible with the high 3He/4He ratios measured at Kolumbo (7.07.1 Ra; Fig.4b), indicating that this He isotopic signature can be considered also representative of the mantle below Nisyros. On the other hand, 3He/4He ratios of 6.2 Ra would be compatible with

87Sr/86Sr of ~0.7049 (Fig.4b). The 87Sr/86Sr variability of Nisyros (0.70340.706415,46), which is in line with measured values at Santorini, shows that magmatic dierentiation plus assimilation of crustal rocks during magma ascent could also occur in the evolution of Nisyros magmas. The occurrence of these processes in the plumbing

SCIENTIFIC REPORTS

6

www.nature.com/scientificreports/

system of the volcano is reasonably responsible of the slight lowering of 3He/4He ratios at Nisyros (6.2 Ra) compared to the pristine values obtained at Kolumbo volcano (7.07.1 Ra; Fig.4a).

These new considerations can be extended westward up to Milos volcanic system, whose lowest 87Sr/86Sr (i.e., 0.703715) are comparable to those measured at Santorini and Nisyros (Fig.4b). On this ground, the 3He/4He ratios <4 Ra17,44 (Fig.4a) in our opinion indicate strong crustal contamination process that would mask the pristine isotope signature. On the other hand, the westernmost part of the HVA (from Methana to Sousaki) showed the lowest 3He/4He (0.172.6 Ra17,44) and 87Sr/86Sr values >0.7055 even in the most mac volcanics (basaltic-andesite to andesite of Poros and Methana15,46) (Fig.4a,b). A recent investigation based on Sr and Nd isotopes highlighted an extensive crustal assimilation at Methana, higher than in Santorini and Nisyros46. These features strongly support the recognized variability of the He isotopic signature in these volcanic systems, even if magma mixing and mingling could simultaneously have occurred to generate the recognized variability in the petrological features of erupted volcanics46.

Based on these arguments, we propose that the mantle below the central and eastern part of the HVA (from Milos through Santorini-Kolumbo up to Nisyros-Kos) is rather homogeneous in terms of 3He/4He and preserves its MORB-like magmatic signature8,41.

Volcanic hazard at Kolumbo. Among the subaerial volcanoes belonging to the HVA, Nisyros and Santorini volcanoes are considered as the most active ones. In 19951998, the former was characterized by a seismic activity, aer which the geochemical monitoring of gases collected from the fumaroles highlighted an increase of the

3He/4He (Ra = 0.7)17. This variation was related to the upward movement of subsurface magma, which triggered an enhanced contribution of mantle helium17. In 20112012, a seismic, geodetic and geochemical unrest took place in Santorini56, during which an increase of 3He/4He ratios (Ra= 0.4) has been recorded in the uids emitted at Nea and Palea Kameni34. This increase was interpreted as the result of the intrusion of a more-primitive

3He-rich magma into the shallow plumbing system. These unrests observed at Nisyros and Santorini led to a general fear that an eruption could be imminent.

Based on our present results, we infer that Kolumbo submarine volcano is very active at present time and the associated volcanic hazard may be potentially high. With this in mind, the development of a regular geochemical monitoring program for this potentially dangerous submarine volcano is strongly suggested.

In summary, hydrothermal uids from Kolumbo submarine volcano exhibited the highest 3He/4He ratios (up to 7.1 Ra) across the HVA, which are surprisingly higher (>3 Ra units) than those previously reported for the adjacent island of Santorini. These uids reect a direct mantle degassing that can only be produced in an extensive tectonic regime characterized by the presence of active lithospheric faults. The range of 3He/4He ratio measured at Kolumbo is the highest of all the HVA and is representative of the mantle below the central and eastern part of the HVA. The degassing of high-temperature uids from the bottom of Kolumbo crater with a mantle-like 3He/4He ratio suggests that this submarine volcano is characterized by a vigorous activity with potential volcanic hazard.

Methods

Sampling. The Kolumbo submarine volcano is characterized by a well-defined 1500-meter-wide crater (Fig.2a), with a rim as shallow as 17m and a oor ~500m below the sea level26,28 (Fig.2a,b). On the active hydro-thermal vent eld at the crater oor of Kolumbo (Fig.2b), rst discovered in 200623, relatively high temperatures (up to ~220 C4,27) have been measured. During the 4-SeaBiotech cruise on board of the R/V Aegaeo (Hellenic Centre for Marine Research) in May 2014, seaoor exploration of hydrothermal activity has been carried out with the Greek ROV Max Rover. Due to the pressure-temperature (P-T) conditions of the vents (P~50 bar, T>200C), hydrothermal uid/gas discharges are in form of clear waters together with continuous gas bubbling at dierent chimneys (Fig.2c and supplementary movies).

High temperature hydrothermal gases were collected in titanium gas-tight bottles of 200 ml capacity each equipped with funnels. These bottles have been built specially to avoid out-gassing and gas leakage during recovery and were initially designed by IFREMER for the manned submersible Nautile. Additional samplers has been projected and developed by Istituto Nazionale di Geosica e Vulcanologia (INGV), Sezione di Palermo (Italy), and experimentally used at Kolumbo. Both types of bottles have been pre-evacuated on board of the oceano-graphic vessel, just before each submersion (internal pressure <103 bar). We point out that we did not observe any dierence in using the two types of gas-tight samplers. For sake of clarity, sample A2 was collected with INGV-type bottle (Table1). The gas-tight bottles were held by the ROV Max Rover of HCMR arm over the bubble streams discharged from the chimneys and then triggered to collect a sample (see supplementary movies).

We collected gas samples from seven dierent chimneys located at the bottom of Kolumbo crater (Fig.2b,c). Aer each sampling, once the gas-tight syringes were onboard of the research vessel, gases at 50 bar were immediately extracted in stainless steel or titanium bottles for safe transportation (at lower pressures) towards the laboratory.

The analysis of chemical composition and He-Ne isotopes of the collected gases has been performed in the laboratories of INGV-Palermo. The concentrations of CO2, O2 and N2 were analyzed by a Perkin Elmer Clarus 500 gas chromatograph equipped with a 3.5-m Carboxen 1000 column and double detector (hot-wire detector and ame ionization detector), with analytical errors of <3%. 3He, 4He and 20Ne and the 4He/20Ne ratios were determined by separately admitting He and Ne into a split ight tube mass spectrometer (GVI-Helix SFT, for He analysis) and into a multicollector mass spectrometer (Thermo-Helix MC plus, for Ne analysis), aer standard purication procedures32,34. The 3He/4He ratio is expressed as R/Ra (being Ra the He isotope ratio of air and equal to 1.39106)58, and the analytical error is generally below 0.3%. The 3He/4He values were corrected for the atmospheric contamination based on the measured 4He/20Ne ratio34 as follows:

SCIENTIFIC REPORTS

7

www.nature.com/scientificreports/

=

R/Ra ( ) ( )

M M A M A

where subscripts M and A refer to measured and atmosphere theoretical values, respectively [(He/Ne)A=0.318]58. The corrected 3He/4He ratios reported in the text and in Table1 are expressed as Rc/Ra values. The correction is small or negligible for most of the samples, with the maximum bias of ~0.2 Ra appearing in the sample showing the lowest 4He/20Ne.

1. Crisp, J. A. Rates of magma emplacement and volcanic output. J. Volcanol. Geotherm. Res. 20, 177211 (1984).2. Chavagnac, V., German, C. R. & Taylor, R. N. Global environmental eects of large volcanic eruptions on ocean chemistry: evidence from hydrothermal sediments (ODP Leg 185, Site 1149). J. Geophys. Res. 113, B06201 doi: 10.1029/2007JB005333 (2008).3. Brewer, P. G. A dierent ocean acidication hazardThe Kolumbo submarine volcano example. Geology 41, 10391040 (2013).4. Carey, S. et al. CO2 degassing from hydrothermal vents at Kolumbo submarine volcano, Greece, and the accumulation of acidic crater water. Geology 41, 10351038 (2013).

5. Fouque, F. Santorini and its Eruptions (translated and annotated by A.R. McBirney, 1998): Baltimore, Maryland. John Hopkins University Press, 1321 (1879).

6. Buttereld, D. A. et al. Geochemistry of hydrothermal uids from Axial Seamount Hydrothermal Emissions Study Vent Field, Juan de Fuca Ridge: Subseaoor boiling and subsequent uid-rock interaction. J. Geophys. Res. 95, 12,89512,921 (1990).

7. Von Damm, K. L. et al. Evolution of East Pacic Rise hydrothermal vent uids following a volcanic eruption, Nature 375, 4750 (1995).

8. Graham, D. W. Noble gas isotope geochemistry of Mid-Ocean ridge and Ocean island basalts: characterization of mantle source reservoirs. In NobleGases in Geochemistry and Cosmochemistry, Rev. Mineral. Geochem. 47, edited by Porcelli, D. P., Ballentine, C. J. & Wieler, R. Mineralogical Society of America, 247317 (2002).

9. Caracausi, A. et al. Changes in uid geochemistry and physico-chemical conditions of geothermal systems caused by magmatic input: The recent abrupt outgassing o the island of Panarea (Aeolian Islands, Italy). Geochim. Cosmochim. Acta 69, 30453059 (2005).

10. Lupton, J. et al. Submarine venting of liquid carbon dioxide on a Mariana Arc volcano, Geochem. Geophys. Geosyst. 7, Q08007, doi: 10.1029/2005GC001152 (2006).11. Lupton, J. et al. Venting of a separate CO2-rich gas phase from submarine arc volcanoes: Examples from the Mariana and Tonga-Kermadec arcs. J. Geophys. Res. 113, B08S12, doi: 10.1029/2007JB005467 (2008).

12. Lan, T. F. et al. Evaluating Earth degassing in subduction zones by measuring helium uxes from the ocean oor, Earth Planet. Sci. Lett. 289, 317322 (2010).

13. Le Pichon, X. & Angelier, J. The Hellenic arc and trench system: a key to the neotectonic evolution of the eastern Mediterranean area. Tectonophysics 60, 142 (1979).

14. McKenzie, D. P. Active tectonics of the Mediterranean region Geophys. J. Int. 30, 109185 (1972).15. Francalanci, L., Vougioukalakis, G. E., Perini, G. & Manetti, P. A West-East traverse along the magmatism of the South Aegean volcanic arc in the light of volcanological, chemical and isotope data. Developments in Volcanology (Amsterdam, Olanda). In The South Aegean Active Volcanic Arc, Present Knowledge and Future Perspectives, 7, edited by Fitykas, M. & Vougioukalakis, G. E. 65111 (2005).

16. Agostini, S., Doglioni, C., Innocenti, F., Manetti, P. & Tonarini, S. Neogene volcanism and extension in the Western Anatolian-Aegean area: A new geodynamic model. IOP Conf. Ser.: Earth Environ. Sci. 2, doi: 10.1088/1755-1307/2/1/012008 (2008).

17. Shimizu, A., Sumino, H., Nagao, K., Notsu, K. & Mitropoulos, P. Variation in noble gas isotopic composition of gas samples from the Aegean arc, Greece. J. Volcanol. Geotherm. Res. 140, 321339 (2005).

18. Bailey, J. C., Jensen, E. S., Hansen, A., Kann, A. D. J. & Kann, K. Formation of heterogeneous magmatic series beneath North Santorini, South Aegean island arc. Lithos 110, 2036 (2009).

19. Agostini, S., Doglioni, C., Innocenti, F., Manetti, P. & Tonarini, S. On the geodynamics of the Aegean ri, Tectonophysics 488, 721 (2010).

20. Suckale, et al. High-resolution seismic imaging of the western Hellenic subduction zone using teleseismic scattered waves. Geophys. J. Int. 178, 775791 (2009).21. Royden, L. H. & Papanikolaou, D. J. Slab segmentation and late Cenozoic disruption of the Hellenic arc, Geochem. Geophys. Geosyst. 12, Q03010, doi: 10.1029/2010GC003280 (2011).

22. Jolivet et al. Aegean tectonics: Strain localisation, slab tearing and trench retreat. Tectonophysics 597598, 133 (2013).23. Sigurdsson, H. et al. Marine Investigations of Greeces Santorini Volcanic Field. EOS, 87(34), 337 (2006).24. Dimitriadis, I. et al. Seismicity and active tectonics at Coloumbo Reef (Aegean Sea, Greece): monitoring an active volcano at Santorini Volcanic Center using a temporary seismic network. Tectonophysics 465, 136149 (2009).

25. Dimitriadis, I. et al. P and S velocity structures of the SantoriniKolumbo volcanic system (Aegean Sea, Greece) obtained by nonlinear inversion of travel times and its tectonic implications. J. Volcanol. Geotherm. Res. 195, 1330 (2010).

26. Nomikou, P. et al. Submarine volcanoes of the Kolumbo volcanic zone NE of Santorini Caldera, Greece. Global Planet. Change 9091, 135151. doi: 10.1016/j.gloplacha.2012.01.001 (2012).

27. Carey, S. et al. Exploration of the Kolumbo volcanic rift zone. Oceanography 24 (1 suppl), 2425 doi: 10.5670/oceanog.24.1. supplement (2011).

28. Nomikou, P., Papanikolaou, D., Alexandri, M., Sakellariou, D. & Rousakis, G. Submarine volcanoes along the Aegean volcanic arc. Tectonophysics 597598, 123146 (2013).

29. Bohnho, M. et al. Microseismic activity in the Hellenic Volcanic Arc, Greece, with emphasis on the seismotectonic setting of the Santorini Amorgos Zone. Tectonophysics 423(14), 1733 (2006).

30. Kilias, S. P. et al. New insights into hydrothermal vent processes in the unique shallow-submarine arc-volcano, Kolumbo (Santorini), Greece. Sci. Rep. 3, 2421 (2013).

31. Martelli, M. et al. Noble-gas signature of magmas from a heterogeneous mantle wedge: the case of Stromboli volcano (Aeolian Islands, Italy). Chem. Geol. 368, 3953, doi: 10.1016/j.chemgeo.2014.01.003 (2014).

32. Di Piazza, A. et al. Geochemistry of the mantle source and magma feeding system beneath Turrialba volcano, Costa Rica. Lithos 232, 319335 (2015).

33. Rizzo, A. et al. New insights into magma dynamics during last two eruptions of Mount Etna as inferred by geochemical monitoring from 2002 to 2005. Geochem. Geophys. Geosyst. 7, Q06008, doi: 10.1029/2005GC001175 (2006).

34. Rizzo, A. L. et al. New mac magma relling a quiescent volcano: Evidence from He-Ne-Ar isotopes during the 20112012 unrest at Santorini, Greece. Geochem. Geophys. Geosyst. 16, 798814 (2015).

35. Rizzo, A. L. et al. The 2014 eusive eruption at Stromboli volcano (Italy): Inferences from soil CO2 ux and 3He/4He ratio in thermal waters. Geophys. Res. Lett. 42, 22352243 (2015b).

36. Sano, Y. et al. Ten-yearheliumanomalyprior to the 2014 Mt Ontakeeruption. Sci. Rep. 5, 13069, doi: 10.1038/srep13069 (2015).

(R /Ra)(He/Ne) (He/Ne) / (He/Ne) (He/Ne) (1)

SCIENTIFIC REPORTS

8

www.nature.com/scientificreports/

37. Papazachos, B. C., Dimitriadis, S. T., Panagiotopoulos, D. G., Papazachos, C. B. & Papadimitriou, E. E. Deep structure and active tectonics of the southern Aegean volcanic arc. The south Aegean active volcanic arc: present knowledge and future perspectives. Develop. Volcanol. 7, 4764 (2005).

38. Doumas, C. G. Thera: Pompeii of the Ancient Aegean. (Thames and Hudson, 1983).39. Hilton, D. R., McMurtry, G. M. & Go, F. Large variations in vent uid CO2/3He ratios signal rapid changes in magma chemistry at Loihi seamount, Hawaii. Nature. 396, 359362 (1998).

40. Yang, T. F. et al. Gas compositions and helium isotopic ratios of uid samples around Kueishantao, NE oshore Taiwan and its tectonic implications. Geochem. J. 39, 469480 (2005).

41. Hilton, D. R., Fischer, T. P. & Marty, B. Noble gases and volatile recycling at subduction zones. In Noble Gases in Geochemistry and Cosmochemistry, Rev. Mineral. Geochem. 47, edited by Porcelli, D. P., Ballentine, C. J. & Wieler, R., Mineralogical Society of America. 319370 (2002).

42. Nagao, K., Kita, I., Matsuda, J. I. & Mitropoulos, P. Noble gas isotope geochemistry of volcanic gases from the Aegean island arc, Bull. Geol. Soc. Greece, 25, 3341 (1991).

43. Chiodini, G. et al. Geochemistry of Santorini uids. In edited by R. Casale et al. The European laboratory volcanoes. Proceedings of the 2nd Workshop, Santorini, Greece, 24 May 1996, EUR 18161 EN, pp. 193232, European Commission, Luxembourg (1998).44. DAlessandro, W., Brusca, L., Martelli, M., Rizzo, A. & Kyriakopoulos, K. Geochemical characterization of natural gas manifestations In Greece, edited by Koukis, G. et al. in Bulletin of the Geological Society of Greece: Proceedings of the 12th International Congress, Patras, Greece (2010).

45. Graham, D. W., Hanan, B. B., Hmond, C., Blichert-To, J. & Albarede, F. Helium isotopic textures in Earths upper mantle. Geochem. Geophys. Geosyst, doi: 101002/2014GC005264 (2014).

46. Elburg, M. A., Smet, I. & De Pelsmaeker, E. Inuence of source materials and fractionating assemblage on magmatism along the Aegean Arc, and implications for crustal growth. From: Gomez-Tuena, A., Straub, S. M. & Zellmer, G. F. (eds). OrogenicAndesites and Crustal Growth. Geological Society, London, Special Publications 385 (2013).

47. Zellmer, G., Turner, S. & Hawkesworth, C. Timescales of destructive plate margin magmatism: New insights from Santorini, Aegean volcanic arc, Earth Planet. Sci. Lett. 174, 265281 (2000).

48. Kurz, M. D., Kenna, T. C., Lassiter, J. C. & De Paolo, D. J. Helium isotopic evolution of Mauna Kea Volcano: First results from the 1 km drill core. J. Geophys. Res. 101, 11, 78111, 791 (1996).

49. Martelli, M. et al. Heliumstrontium isotope constraints on mantle evolution beneath the Roman Comagmatic Province, Italy. Earth Planet. Sci. Lett. 224, 295308 (2004).

50. Ozima, M. & Podosek, F. A. Noble Gas Geochemistry (Cambridge Univ. Press, NY, 1983).51. Druitt, T. H. et al. Santorini volcano. Geol. Soc. Mem. 19, 165 (1999).52. Vaggelli, G., Pellegrini, M., Vougioukalakis, G. E. & Francalanci, L. Highly Sr-radiogenic tholeiitic magmas in the latest interplinian activity of Santorini volcano, Greece, J. Geophys. Res. 114, B06201, doi: 10.1029/2008JB005936 (2009).

53. Cantner, K., Carey, S. & Nomikou, P. Integrated volcanologic and petrologic analysis of the 1650AD eruption of Kolumbo submarine volcano, Greece. J. Volcanol. Geotherm. Res. 269, 2843 (2014).

54. Makris, J. Geophysical investigations of the Hellenides. Hamburger geophysischerEinzelschrien 34, 1124 (1977).55. Hubscher, C., Ruhnau, M. & Nomikou, P. Volcano-tectonic evolution of the polygenetic Kolumbo submarine volcano/Santorini (Aegean Sea). J. Volcanol. Geotherm. Res. 291, 101111 (2015).

56. Parks, M. M. et al. Evolution of Santorini volcano dominated by episodic and rapid uxes of melt from depth. Nat. Geosci. 5, 749754 (2012).

57. Gertisser, R., Preece, K. & Keller, J. The Plinian Lower Pumice 2 eruption, Santorini, Greece: Magma evolution and volatile behaviour. J. Volcanol. Geotherm. Res. 186, 387406 (2009).

Acknowledgements

We thank Giuseppe Riccobono and Paolo Cosenza for the technical support in the projecting and construction of the INGV-Palermo gas tight sampler used to collect part of the submarine uids. The former also contributed to the gas collection and extraction, and in the activities on board the R/V Aegaeo of the Hellenic Centre for Marine Research. The officers and the crew of the R/V Aegaeo are especially acknowledged for their valuable help during sampling. Sampling campaign 4-Seabiotech was supported by Seabiotech project (spider.science.strath.ac.uk/ seabiotech/) funded by the European Commission within its FP7 Programme with Grant Number 311932. We are grateful to Mariano Tantillo for the isotopic analyses of noble gases and Mauro Martelli and Francesco Salerno for the analyses in gas chromatography. Comments made by three anonymous Reviewers strongly improved the manuscript.

Author Contributions

A.L.R. and P.N. conceived the study. A.L.R., V.C., P.N., P.N.P., M.M. G.K., A.M. and A.C. participated to the collection of the gas samples. A.L.R. and A.C. analyzed the gases and elaborated the data. All authors contributed to the preparation of the manuscript.

Supplementary information accompanies this paper at http://www.nature.com/srep

Competing nancial interests: The authors declare no competing nancial interests.

How to cite this article: Rizzo, A. L. et al. Kolumbo submarine volcano (Greece): An active window into the Aegean subduction system. Sci. Rep. 6, 28013; doi: 10.1038/srep28013 (2016).

This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the articles Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/

SCIENTIFIC REPORTS

9