Shallow Submarine CO₂ Emissions in Coastal Volcanic Areas Implication for Global Carbon Budget Estimates: The Case of Vulcano Island (Italy)

Abstract

The Earth’s degassing is an important factor in evaluating global carbon budget estimates and understanding the carbon cycle. As a result, numerous studies have focused on this topic. However, current estimates predominantly focus on subaerial CO₂ emissions and CO₂ deep submarine emissions, particularly along mid-ocean ridges (MORs), whereas very few and only spatially limited estimates of shallow submarine CO₂ emissions have been reported, despite being widespread features of the seafloor. This study reports the results of measuring the dissolved CO₂ concentrations in shallow submarine environments along the coast of Vulcano Island (Aeolian Islands, Italy). For the areas exhibiting the highest concentrations, we calculated the amount of diffuse degassing by computing the sea–air CO₂ flux. The results revealed extremely high dissolved CO₂ concentrations, reaching up to 24 vol.% in areas with visible hydrothermal activity, including one location far from the island’s main crater. Notably, elevated CO₂ levels were also detected in areas with minimal or no apparent hydrothermal discharge, indicating the occurrence of diffuse degassing processes in these areas. In addition, the calculated diffuse degassing flux was comparable in magnitude to the CO₂ flux directly emitted into the atmosphere from the island’s main bubbling pools.

1. Introduction

The Earth’s degassing causes the release of deep fluids stored in the crust and mantle toward the surface, contributing to the global emissions of CO₂ into the atmosphere and the carbon cycle. The Earth’s degassing occurs preferentially in tectonic and volcanic areas where faults allow the release of deep stored fluids [1,2,3,4,5,6,7]. Although global CO₂ emissions from the Earth’s degassing have been widely studied, there is still significant uncertainty in their estimates [8] for both subaerial and submarine environments. Global estimates of subaerial CO₂ fluxes vary widely, ranging between 232 and 540 Mt yr⁻¹ [9,10]. According to these authors, one-third of subaerial emissions, between 83 and 183 Mt yr⁻¹, are emitted into the atmosphere through diffuse emissions [9,10]. Diffuse subaerial CO₂ emissions resulting from the Earth’s degassing, though less well constrained, are recognized as significant contributors to global CO₂ fluxes [7,9,10]. Estimates of submarine CO₂ emissions exhibit considerable variability, ranging from as low as 4.4 to as high as 792 Mt yr⁻¹ [11,12]. This broad range reflects the methodological differences and uncertainties related to parameters such as dissolved carbon concentrations and global ³He fluxes [10], as well as the inherent challenges of conducting direct measurements in underwater environments [7]. More recent assessments have narrowed these estimates to a range of between 60 and 110 Mt yr⁻¹ [7,13,14,15,16]. Although submarine CO₂ emissions are generally considered smaller than their subaerial counterparts, they play a crucial role in the Earth’s carbon cycle, particularly in modulating ocean chemistry. Current estimates predominantly derive from deep-water degassing along mid-ocean ridges (MORs), while estimates of shallow submarine CO₂ emissions remain scarce and spatially limited (e.g., [17]). These shallow submarine emissions primarily occur through shallow water hydrothermal vents (SWHVs), typically found at depths of less than 200 m, close to volcanic and tectonically active areas. These vents release a gas phase predominantly composed of CO₂, accompanied by various trace gases [18,19]. Price and Giovannelli [18] presented a comprehensive overview of the global distribution of SWHVs, noting their frequent occurrences near coastal or island volcanoes and shallow seamounts. Additionally, such emissions are common in tectonically active coastal regions, where fluids ascend through zones of high permeability, including regional fault systems and actively deforming sectors [20,21,22]. Several examples of this latter type of submarine CO₂ emission have been reported in the southern Tyrrhenian Sea (Italy): Tor Caldarara [23,24], Napoli Bay [25], San Giorgio (northeast coast of Sicily) [26], Zannone Island (western Pontine Archipelago) [27], Secca delle Fumose (Pozzuoli Bay) [28]. Examples have been reported all around the world as well: the German North Sea [29], offshore northeast Taiwan [30,31], and Baja California (Mexico) [32]. In the field, SWHVs include single, gas-rich vents that bubble through the overlying seawater column, and emissions occur diffusely from the seafloor through clusters of small vents and “diffuse vents”, i.e., diffuse bubbling through sand and fractured rocks [18,19,33]. Studies of SWHVs have focused mainly on their ecological and biogeochemical significance [19,24] or used SWHVs as field laboratories for investigating the consequences of ocean acidification [17,34,35] and as analogs of seepage from sub-seabed CO₂ geological storage areas [36,37]. However, relatively few studies have quantified their CO₂ fluxes. Notable exceptions include investigations of the Aegean Islands (Greece) [17], especially around the Milos Islands [38,39,40], and Aeolian Islands (Italy) [37,41,42,43]. According to Price and Giovannelli [18], taking into account the abundance, global distribution, and shallow depths of these vents, they could represent major contributors to global CO₂ degassing. In addition, given that diffuse CO₂ subaerial emissions account for one-third of the total subaerial emissions, diffuse submarine emissions via SWHVs should likewise constitute a substantial component of submarine CO₂ emissions. Recently, Bastianoni and coauthors [44], based on limited data and through three different data expansion techniques, estimated that SWHVs can contribute between 20 and 128 Mt CO₂ yr⁻¹, with a conservative value estimate of 64.1 Mt CO₂ yr⁻¹. In this study, we examined the shallow submarine CO₂ emissions around Vulcano Island (Aeolian Islands, Italy), where SWHVs are present at multiple sites. Vulcano Island has exhibited persistent hydrothermal activity since the most recent eruption of the La Fossa cone between 1888 and 1890 [45]. The island hosts several fumarolic fields, including one located on the northern rim of the La Fossa crater and another near the shoreline in the northeastern sector [46]. In addition, Vulcano Island is characterized by CO₂-rich groundwater and notable soil CO₂ emissions [46,47,48]. Submarine emission zones are visible at shallow depths along the beaches in the southern and northeastern sectors. The investigation involved direct measurements of the dissolved CO₂ concentrations in shallow seawater at five selected sites (Figure 1).

The selected sites included locations along the seashore and rocky coastlines. For areas with high CO₂ concentrations, we also provide estimates of the CO₂ diffuse degassing flux from the sea surface to the atmosphere.

2. Materials and Methods

The measurements were taken with a submergible probe EosGP sensor (Eosense, Dartmouth, NS, Canada). The operation of the device is based on the separation process of dissolved gases via a semipermeable membrane. When the device is immersed in water, the dissolved gases permeate inside the system, and an equilibrium between the dissolved gases and the gas in the system headspace is reached [49]. The probe is equipped at its bottom with a semipermeable membrane to allow permeation of dissolved gases inside the headspace of the probe. The dissolved CO₂ concentration is determined via non-dispersive infrared sensing housed in the probe’s headspace, which operates within a calibration range of 0–12.5% and has a full-scale capacity of 24%. The probe was connected to a tailor-made acquisition device that registered the value and position (latitude and longitude) every 5 s. To calculate the partial pressure of CO₂ (pCO₂), the recorded concentration was multiplied by the atmospheric pressure value at sea level, given the shallow water depths (<100 cm) at which all the measurements were performed. The isotopic ratio ³He/⁴He and ⁴He were measured by a split-flight-tube mass spectrometer (GVI Helix SFT, GVinstrument) equipped with a double collector system that enabled the simultaneous detection of ion beams after the purification of the sample from the major gaseous species and separation from the other noble gases.

The isotopic composition of CO₂ carbon was analyzed using a Thermo Delta V Plus isotope ratio mass spectrometer (Thermo Fisher Scientific, Waltham, MA, USA) coupled with the GasBench II and reported in δ‰ versus the V-PDB standard, with a standard deviation of the ¹³C/¹²C ratio of ±0.15 ‰.

Measurements of the dissolved CO₂ concentration were conducted in five areas: two in the northern sector, two along the eastern coast, and a final site in the southern sector (Figure 1). We measured the dissolved CO₂ concentrations both in areas with minimal or no visible hydrothermal activity and in areas with evident hydrothermal activity with gas-rich vent bubbling.

3. Results

The first investigated area (Site A1 in Figure 1) is a clear example of an area exhibiting pronounced hydrothermal activity. This zone, located between the Vulcanello Peninsula and the Faraglione cone, is widely documented in the literature as Baia di Levante—a well-characterized site of substantial hydrothermal fluid discharge. The area is marked by low pH conditions and the presence of trace gases such as H₂S, CH₄, and H₂ [19,35,42,43,50,51]. The hydrothermal manifestations include seafloor vents near the shoreline with vigorous bubbling as well. The emitted gas phase is predominantly composed of CO₂, accompanied by a few percent of H₂S and hundreds of ppm of CH₄ and H₂ [52].

Estimates of the CO₂ flux emitted directly into the atmosphere via bubbling in this area range from 1.3 to 3.6 tons per day [42,43]. During the most recent volcanic crisis on Vulcano Island, which began in September 2021, the hydrothermal discharge activity in the area markedly increased from May 2022 onward [53]. This intensification was evidenced by the appearance of extensive whitish sulfur flocculation and intense bubbling activity [53]. In September 2022, we conducted measurements of the dissolved CO₂ concentrations along the shoreline at a depth of about 50 cm below the sea surface (Figure 2a). The recorded concentrations consistently exceeded the air-saturated seawater (ASSW) value of 420 ppm along the entire surveyed shoreline. The lowest dissolved CO₂ concentration, recorded at the northernmost end of the transect, was 1800 ppm. The concentrations increased substantially in proximity to visible bubbling zones, with the highest values—exceeding the probe’s saturation limit of 24 vol.%—observed near the most intense bubbling points at the southern extremity. Measurements in the other areas were taken in June and July 2022. The second examined site (Site A2 in Figure 1), located between the Porto di Levante dock and the northern slopes of the volcanic cone of La Fossa, was surveyed along the dock, the adjacent beach, and selected offshore locations about 50 m from the coastline using a boat-based approach (Figure 2b). The observations here included sporadic bubble trains rising from the seabed and milky water coloration near the volcanic slopes due to fluid emissions. The dissolved CO₂ concentrations ranged from 650 to 4800 ppm, which were markedly lower than those at Site A1 but consistently surpassed the ASSW concentrations.

Along the Porto di Levante dock, the highest recorded CO₂ concentration (1800 ppm) was observed nearest to the Faraglione cone. Along the adjacent beach, the distribution of CO₂ concentrations was heterogeneous, with the peak values reaching up to 4800 ppm at both extremities. Notably, concentrations substantially exceeding the typical ASSW values extended across an area spanning more than 100 m. In the northwestern sector of the island (Box B in Figure 1), measurements were conducted near and along the Civil Protection pier (Figure 3a). The dissolved CO₂ concentrations in this area ranged from 420 to 680 ppm. The highest values were recorded along the small beach east of the pier, which slightly exceeded the ASSW reference level, suggesting a minor additional CO₂ input (Figure 3a). Along the eastern coastline, measurements were performed in the waters off the northeastern slopes of the La Fossa cone, near Scoglio della Sirena (Box C in Figure 1), and within the Punta Quadra cave (Box D in Figure 1), located in the central section of the eastern coastline. In the latter site, occasional small gas bubbles were observed rising to the surface inside the cave. Measurements around Scoglio della Sirena revealed an average CO₂ concentration of 455 ppm, slightly above the ASSW value (Figure 3b).

Measurements inside the Grotta Quadra (Figure 3c) revealed values consistently above the ASSW, with a maximum value of 1600 ppm, thus indicating a contribution of deep CO₂ (Figure 3c). In the southern sector, the measurements were carried out along the beach of Gelso village (Box E in Figure 1). In this area, trains of bubbles rising from the seabed could be observed at various points. The measured CO₂ concentrations varied widely between 0.01% and 17%, with a mean value of 4% (Figure 4). The highest CO₂ concentrations were recorded in the central sector of the beach, while the lowest values were observed in the eastern sector. At one site, where visible bubble trains were present, a gas sample was collected for the chemical and isotopic analysis of the emitted fluids. The sampling was conducted using an inverted funnel positioned over the bubble stream on the seabed. The isotopic composition of the collected gases—specifically, a δ¹³C CO₂ value of −2.10‰ and a helium isotopic ratio (Rc/Ra) of 4.54 (where Rc is the air-corrected ³He/⁴He ratio of the sample and Ra is the ³He/⁴He ratio in air)—clearly indicates a deep magmatic origin for the emitted fluids [54]. The findings from the latter two study areas are particularly noteworthy, as they demonstrate that deep CO₂ emissions can occur in locations far away from the active crater or well-known hydrothermal systems. In this context, the presence of anomalous degassing zones is likely associated with active tectonic structures that facilitate the ascent of deep-sourced gases, extending their influence southward across the island.

4. Discussion

The results of dissolved the CO₂ concentrations revealed areas with significantly elevated levels, particularly in the northeastern sector of the island (Site A) and the southern sector (Site E). For these sites, it is of particular interest to estimate the CO₂ emitted diffusely from the sea surface into the atmosphere (sea–air CO₂ flux).

Estimates of the CO₂ flux in the SWHVs, in most cases, have focused on the amount of CO₂ emitted into the sea (e.g., [28]) or released into the atmosphere through bubbling [41,42,43], while estimates of the CO₂ emitted diffusely from the sea surface into the atmosphere are limited to only the SWHV on Milos Island, where a floating accumulation chamber was employed to measure the CO₂ flux [55]. Sea–air CO₂ flux estimates are generally performed in oceanography and environmental studies [56,57,58]. The open ocean is a consistent carbon sink because of the physical and biological processes that drive CO₂ uptake, whereas coastal regions, although smaller in size, often act as sources because of inputs such as submarine groundwater discharge (SGD), riverine input, or upwelling [57,59]. The basic concept of the sea–air CO₂ flux exchange is that when surface waters are supersaturated with respect to the CO₂ atmospheric concentration, CO₂ tends to escape from seawater into the atmosphere and vice versa. Typically, the sea–air CO₂ flux (F in mol m⁻² d⁻¹) is calculated using a bulk flux equation expressed in terms of the partial pressure [60]:

$$F=k·Kh{(\mathrm{pCO}}_{2\mathrm{sea}}-{\mathrm{pCO}}_{2\mathrm{atm}})$$

(1)

where

k is the gas transfer velocity (piston velocity) (m d⁻¹)

Kh is the solubility of CO₂ in seawater (mol m⁻³ atm⁻¹)

pCO_2sea is the partial pressure of CO₂ in seawater (atm)

pCO_2atm is the atmospheric CO₂ partial pressure (atm)

Specifically, to compute the solubility coefficient Kh in Formula (1), which is a function of the salinity and temperature, we used the formula and the constants reported in Weiss [61]. To compute the gas transfer velocity k, function of the wind velocity and Schimdt number, we used parameterization reported by Whikellof [60]:

$${k=0.31u^2(Sc/660)}^{-0.5}$$

(2)

where

u is the wind velocity (m s⁻¹)

Sc is the Schmidt number (a dimensionless ratio of the kinematic viscosity of the fluid (seawater) and the diffusivity of the gas within that fluid; this value depends on the temperature and salinity). We computed the Schmidt number by using the formula and the constants reported in [60]. Here, 660 is the Schmidt number of CO₂ in seawater at 20 °C.

For the calculations, we used the mean measured seawater temperature of 28 °C, the seawater salinity of 35‰, and the mean wind speed recorded for the area of 5 m s⁻¹. The sea–air CO₂ flux values were calculated for each measurement point. To estimate the total flux over the study area, we generated a spatial distribution map of the CO₂ flux using Kriging interpolation. The bulk daily CO₂ flux was then calculated by integrating the flux values above zero across the interpolated surface. The results are presented in Figure 5 and Figure 6. For Site A, two separate flux maps were generated: one for the northern sector (A1, Figure 5a) and another for the southern sector (A2, Figure 5b). The northern sector (Site A1), corresponding to the Baia di Levante site, yielded a mean sea–air CO₂ flux of 5.2 mol m⁻² d⁻¹ and a total daily bulk flux of 3.3 t d⁻¹. This diffusely emitted sea–air CO₂ flux was notably high and comparable in magnitude to the estimated flux directly released into the atmosphere from bubbling pools at the same site (3.6 t d⁻¹) [43]. In contrast, the southern sector (Site A2) exhibited a significantly lower mean CO₂ flux of 0.12 mol m⁻² d⁻¹, corresponding to a daily bulk flux of 0.09 t d⁻¹, about two orders of magnitude lower than that recorded in Baia di Levante. Although Site A2 exhibited lower sea–air CO₂ fluxes compared to Baia di Levante, the calculated values remained several orders of magnitude higher than those typically reported for coastal regions [56]. This highlights the substantial contribution of diffuse degassing processes in coastal volcanic settings, even in areas where the visible hydrothermal activity is minimal or absent.

Finally, in the Gelso area, located in the southern sector of the island (Figure 6), we recorded a mean sea–air CO₂ flux of 3 mol m⁻² d⁻¹, comparable to that observed at Baia di Levante. This corresponds to a flux of 132 g m⁻² d⁻¹, which is about twice the average CO₂ flux typically reported for soil CO₂ emissions on the lower flanks of the La Fossa cone and in the inhabited area of the Vulcano Porto area [53,62].

The daily bulk flux of 0.18 t d⁻¹ is an order of magnitude lower than that at Baia di Levante because of the smaller surface area investigated. However, it is reasonable to assume that anomalous emissions also extend to neighboring beaches, which has not yet been investigated.

In

Table A1. Shallow-water hydrothermal vent CO₂ flux data.

Site	Depth Range (m)	Surface (m2)	Flux (t yr−1)	Flux (t yr−1 m−2)	Ref.
Castello Aragonese vents, Ischia Island (Italy)	0–10	5000	1498	0.300	[34]
Milos (Greece)	0–200	70,000,000	2,244,000	0.032	[39]
Champagne Hot Springs (Lesser Antilles)	1–5	100	0.198	0.002	[63]
Panarea hydrothermal field (Italy)	7–200	120,000	9600	0.080	[37]
Kraternaya Bight, Yankich Island (Russia)	0–60	700,000	1128	0.002	[64]
Tutum Bay, Ambitle Island (Papua New Guinea)	5–10	450,000	8000	0.018	[65]
Secca delle Fumose, Campi Flegrei Caldera (Italy)	4–20	140,000	18,906	0.135	[28]
Baia di Levante, Vulcano (Italy)	0–1	11,500	1205	0.105	[This study]
Gelso, Vulcano (Italy)	0–1	1200	66	0.055	[This Study]

, we present estimates of the CO₂ flux reported in the literature for SWHVs. In addition to the total annual CO₂ flux values, we include the annual flux per unit area to enable more accurate comparisons across regions of varying sizes. A comparison with our estimates, shown in Figure 7, reveals that our measurements of the diffuse CO₂ flux are comparable with the values reported for similar volcanic environments concerning the CO₂ emissions via bubbling or direct discharge into the sea.

5. Conclusions

The results of the dissolved CO₂ concentration measurements in the submarine environment along the coast of Vulcano Island reveal extremely high dissolved CO₂ concentrations—reaching up to 24 vol.%. These concentrations were measured in areas with visible hydrothermal activity, including one location far from the island’s main crater (La Fossa). Notably, elevated CO₂ levels were also detected in areas with minimal or no apparent hydrothermal discharge, indicating the presence of diffuse degassing processes in these areas. Moreover, the calculated diffuse fluxes—i.e., sea–air CO₂ fluxes—were significantly higher across all the investigated sites compared to the typical values reported for coastal regions. Notably, the diffuse CO₂ flux measured at Baia di Levante was comparable to the flux directly emitted into the atmosphere from major bubbling pools. This finding indicates that a substantial portion of CO₂ at SWHV sites is released diffusely across the sea–air interface, rather than solely through discrete bubbling. This result is particularly important, as most existing CO₂ flux estimates from SWHVs only account for bubbling emissions. Our data suggest that such estimates should be revised to include an equivalent contribution from diffuse emissions. This adjustment is especially critical considering that the total CO₂ flux attributed to SWHVs is comparable in magnitude to the emissions along MORs. Overall, our findings underscore the potential significance of shallow submarine CO₂ emissions in coastal volcanic areas within the context of the global carbon budget. They also highlight the need for further research focused on the spatial variability and quantification of these emissions.