Geophysical Research Letters

Correlated increase in CO2 fumarolic content and diffuse emission from La Fossa crater (Vulcano, Italy): Evidence of volcanic unrest or increasing gas release from a stationary deep magma body?



[1] The peculiarity of the quiescent La Fossa volcano is the occurrence of “crises” characterized by strong increases of fumarole T and output and by chemical changes indicative of an increasing input of magmatic fluids. Several surveys carried out during a new “crisis” began in November 2004 indicate that the total diffuse CO2 emission for the crater area increases by one order of magnitude during crises (up to 1600 ton·d−1 in December 2005). Concern exists on the possibility that these crises be related to an unrest process leading to eruption. The repetition along decades of the same gas compositional variations during crises, their temporal coincidence with increases of the local shallow seismicity, and the lack of any significant ground motion, rather suggest that they correspond to moments of increasing volatile release from a stationary magma system.

1. Introduction

[2] In quiescent explosive volcanoes it is of crucial importance, for the safety of the exposed population, to recognize the early symptoms of the process leading to eruption. Together with the monitoring of seismicity and ground deformations, fluid geochemistry can play a relevant role to this respect, as magmatic gases are increasingly released by the rising magma, in proportions that depend on their solubility in the silicate melt, and reach the surface, along fumarole feeding fractures, well before their parent magma.

[3] La Fossa cone, formed in the last 6 Ka in the island of Vulcano (Aeolian Islands, Italy), is the volcano whose historic activity gave name to the “vulcanian” type of explosive eruptions. Since it last erupted in 1888–1890, it has been affected by two main episodes of increasing fumarolic activity (that have been frequently called “volcanic unrest”) none of which was followed by eruption. The first one occurred in 1913–1923 and the maximum temperature of the crater fumaroles (T) increased from 200°C to 615°C [Sicardi, 1941]. The second episode began in 1977 and T progressively increased from 210°C to 690°C in May 1993 [Chiodini et al., 1995]. Then T decreased, slowly and with fluctuations, to values that in 2004 were around 340–360°C.

[4] On the basis of the chemical and isotopic signatures, the fumarolic fluids have been interpreted as a mixture of hydrothermal and magmatic components which varied during time their relative proportions. The magmatic component is richer in CO2, N2 and He with respect to H2O [Chiodini et al., 1995; Italiano et al., 1998; Nuccio et al., 1999; Capasso et al., 1999]. The high values of the 3He/4He R/Ra ratio (up to 6 times the air ratio Ra during the crises) [Tedesco et al., 1995; Italiano and Nuccio, 1996] fall in the variation range of the magmatic helium of island arc volcanoes [Farley and Neroda, 1998]. It was recognised that during crises the magmatic component dominated the fumarolic composition. In particular inputs of magmatic fluids were detected in 1979–1981, 1985, 1988 and 1996. The associated seismic activity was characterized by swarms of low-magnitude shocks (M < 2.5) with generally shallow foci (h < 4 Km). Waveform feature and frequency of events indicate that they were linked to rising gases in the fumarolic feeding system [Chiodini et al., 1992; Bonaccorso, 2002].

[5] In November 2004 and in November 2005 La Fossa crater was affected by new phases of local anomalous seismicity (INGV-Catania, internal reports) with characteristics similar to the crises registered in 1985, 1988 and 1996. The occurrence of these crises was the occasion to perform a series of surveys on the soil CO2 flux and fumarolic gas composition in the crater area. The main objective was to investigate the correlation between CO2 soil degassing and the input of magmatic fluids highlighted by changes in the fumarole composition.

2. Fumarolic Composition During the 2004–2005 Period

[6] The systematic sampling and analysis of the high-T crater fumaroles carried out since the late '70s allowed the development of a conceptual model capable of explaining the composition of the fumarolic gases and the variations observed during strong degassing episodes. According to a so called “dry” model, La Fossa fumaroles are essentially fed by two components: i) a deep one released from a magma body and ii) a shallower one, formed by the total evaporation of hydrothermal fluids of marine origin entering the zone near the conduits where deep hot fluids rise up [Chiodini et al., 1995]. The best tracers of the magmatic component are CO2, He, N2 and the stable isotopes of H2O [Chiodini et al., 1993; Nuccio et al., 1999]. It has been shown that since 1979 the fumarolic gases fall, in the H2O-CO2-He triangle, on linear trends describing the mixing of the two components. During crises, the magmatic component prevails and the gas composition shifts toward the CO2-He side. The H2O-CO2-He diagram (Figure 1a) clearly shows the persistence with time of the same linear trend and that the composition of the gas collected in 2004 and 2005 is very similar to that of the gases emitted in the previous crises.

Figure 1.

(a) Gas composition of the crater high temperature F1 (1979–mid 1988) and F11 (mid 1988–2005) fumaroles in the H2O-CO2-He plot. During degassing crises (full circles) CO2 content exceeds 12 vol.% approaching the composition of the inferred magmatic component (grey circle). (b) Time variation of the CO2 concentration in the fumarolic gas compared, for the 1995–2006 period, with the CO2 flux measured in the crater target area (CTA). Dashed lines refer to periods without data. Chemical data of the fumaroles for 1979–1991 period are from Chiodini et al. [1995], the following ones are unpublished data of INGV-Osservatorio Vesuviano. F. Tassi (University of Florence) provided some of the 2002–2004 analyses.

[7] The periodical increase of magmatic gases in La Fossa fumaroles is shown in the CO2 chronogram of Figure 1b. In 1979–1981, 1985, 1988 and 1996, the CO2 concentration, as well as the He concentration [Nuccio and Paonita, 2001], shows similarly shaped peaks. In each case CO2 concentrations reach maximum values of 15–20 vol.% which approach the composition of the inferred magmatic gas. The peaks, which generally persist for some months, are characterised by a relatively sharp increase followed by a gentler decline. The compositions measured in December 2004 and December 2005 suggest the occurrence of two new inputs of magmatic gas similar to those detected in the previous crises.

3. CO2 Diffuse Degassing at La Fossa Crater

[8] Eleven surveys of CO2 soil diffuse degassing have been performed in the crater area in the period 1995–2006, using the accumulation chamber method [Chiodini et al., 1996]. Results are shown in Table 1. The data of the 1995 and 2002 campaigns have been published by Chiodini et al. [1996, 2005]; all raw data are included as auxiliary material. The sequential Gaussian simulations (sGs) algorithm was used for mapping the CO2 fluxes and to quantify the total output and the relative error following the method described in Cardellini et al. [2003].

Table 1. CO2 Soil Flux Statistics
DateInvestig. Area, Km2Meas. Points, nMean (Min-Max), gm−2d−1Mean (Min–Max)CTA, td−1
Apr.19950.8291435 (0.50–8830)165 (136–193)
Jul. 19950.47257289 (0.50–7228)109 (94–124)
Jul. 19980.75429288 (0.60–9633)107 (98–116)
Mar. 20020.1954302 (0.70–2360)68 (58–79)
Jun. 20030.7576129 (0.00–1238)41 (37–44)
Dec. 20040.613081370 (1.52–29386)331 (307–355)
Apr. 20050.83403826 (1.06–31942)223 (207–239)
May 20050.40170780 (0.56–18300)152 (138–167)
Jul. 20050.80381333 (1.06–6386)82 (76–87)
Dec. 20050.804892750 (7.74–32510)724 (693–754)
Jan. 20060.251781992 (1.4–37346)504 (451–556)

[9] In order to identify the degassing structures active in the area, the data of the most extended campaigns (July 1998, December 2004, April, July and December 2005) were used. The entire data set consists of 1558 points distributed over a ∼0.8 Km2 area covering the upper part of La Fossa cone (Figure 2). Because the mean flux values and the total output of CO2 varied considerably in the different campaigns (see Table 1) each data set was transformed into a normal distribution (normal scores transforming) before the mapping, by substituting the original values with the corresponding quantiles. The transformed data were used by the sGs algorithm to simulate the normal score value at each location of the computational domain. In total 100 equiprobable simulations were carried out, whose average values at each location are mapped in Figure 2b. The main degassing structure is represented by the La Fossa crater with the exception of its bottom, characterized by impervious soils. Part of this anomalous area coincides with the fumarolic field. Other anomalies, located NE, NW and E of La Fossa, seem to be linked to the volcano-tectonic structures that controlled the formation of the old craters (Figure 2b).

Figure 2.

(a) Location of Vulcano island and limits of the broad area and of the crater target area (CTA) investigated for diffuse CO2 degassing. (b) Map of the CO2 soil degassing of La Fossa obtained using the data of the five most extended surveys of 1998–2005. The rims of the various craters that were active in different times are indicated (a. Present crater; b. Pietre Cotte; c. Palizzi and Commenda; d. Unidentified; e. Punte Nere [Frazzetta and La Volpe, 1991]).

[10] In order to compare the 1995–2006 variation of the CO2 flux a subset of data was chosen, selecting the measurement points falling in an area that had been conveniently covered in all the surveys (crater target area, CTA, 500x500 m, Figure 2). In Figure 1b the CO2 output from CTA is compared with the CO2 fumarole concentration. It is convenient to stress out that CTA fluxes do not depend on the few measurements carried out in the proximity of the fumaroles. Data show that the CO2 fluxes remained relatively low in the campaigns before December 2004, with a gentle decreasing trend from 110–160 ton·d−1 in 1995 to 40 ton·d−1 in 2003. These pre-2004 campaigns were performed during periods of no crisis, that is, when the fumarole composition was dominated by the H2O rich hydrothermal component. The campaign of December 2004 and the following ones were instead carried out specifically to follow the evolution of the soil diffuse degassing during the last two crises occurred at the end of 2004 and at the end of 2005, when the fumarole composition was dominated by the CO2 rich magmatic component. In December 2004 the CO2 flux from CTA significantly increased, with a value (330 ton·d−1) 3 to 4 times higher than in the preceding years and then progressively decreased in April, May and July 2005 concurrently with the decreasing of the CO2 content of the fumaroles. In December 2005, when a new increase in the CO2 content of the fumaroles and a new seismic swarm were recorded at La Fossa cone, CO2 CTA flux increased again at 700 ton·d−1 and decreased at 500 ton·d−1 in January 2006 (Figure 1b).

[11] The time variation of CO2 degassing is shown in the maps of Figure 3. In July 1995 and July 1998 the main CO2 releasing zone was the inner northern slope of the crater, around the fumarolic field and there were several zones with low CO2 flux, indicating the local existence of a low permeability soil, either primary, due to the presence of a wet pyroclastic surge cover, or secondary, related to self-sealing processes by fumarolic sublimate precipitation as in the crater bottom [Chiodini et al., 1996]. In December 2004 the entire crater area was strongly degassing including most of the previous low-permeability zones that hence had been affected by new fractures. From April to July 2005, the surface of the anomalous degassing area strongly decreased and low flux zones reappeared in the crater bottom and on its southern rim. Finally in November–December 2005 the CO2 soil release furtherly reincreased and again the entire crater area was strongly degassing, with average values and total flux that were more than the double of those measured one year earlier (Figure 3 and Table 1).

Figure 3.

Variation of the CO2 soil flux of La Fossa crater area recorded in April 1995–January 2006.

[12] In summary the CO2 flux increase is well correlated with the increase of the magmatic component in the fumaroles and clearly indicates that the 2004 and 2005 crises are the expressions of episodes of intense degassing of the magmatic source.

4. Discussion and Conclusions

[13] Considering that the CO2 output from entire crater area is, on average, 2.2 times the CO2 output from CTA, our data show that in relatively quiet periods La Fossa crater area releases diffusively about 150–200 ton·d−1 of CO2 from a surface of ∼0.5–0.6 km2. During “crises”, the CO2 diffuse output may increase of one order of magnitude, as in December 2005 when a total soil flux of 1600 ton·d−1 was measured. The CO2 total emission integrated over the period October 2004–January 2006 was ∼185,000 tons. This value is impressive considering that an equivalent, if not higher, amount of CO2 is contemporaneously convectively emitted from the fumaroles [Chiodini et al., 1996].

[14] These data indicate, to one side, that the diffuse soil degassing has to be carefully estimated in order to assess the total degassing budget of a volcano and, to the other side that a significant volume of degassing magma must exist at depth. The chemical changes recorded in 1988–1998 at the fumaroles have been attributed by Nuccio and Paonita [2001] to an upraise of rhyolitic magma that in 1998 would have reached a depth of only 2.5 km. There are several difficulties in accepting their conclusion. Should the new 2004–2005 strong degassing episodes be linked to further magma upraise, the conditions should have likely brought to an eruption that instead has not occurred. In addition the inferred rhyolitic composition of the magma is in contrast with several geochemical data. In fact sulphur content of melt inclusions in the phenocrysts of the magmas erupted at La Fossa, compared with COSPEC estimates of the SO2 output from the fumaroles, led Clocchiatti et al. [1994] to infer that the present gas source is likely a primitive basaltic magma, richer in volatile than the evolved latite-rhyolitic magmas of the historical eruptions. The latter have been generated by assimilation-fractional crystallisation processes in very shallow open magma chambers, where no volatile accumulation in the residual melts occurred. Also the 3He/4He ratios of the fumarolic gas (R/Ra ∼ 6) point to a basic degassing magma having a similar R/Ra ratio, in contrast with the 1888–1890 evolved magmas where crustal contamination lowered the R/Ra to a value of 3.8 [Magro, 1997]. An equivalent of the basic degassing magma was identified by Clocchiatti et al. [1994] in the shoshonitic basalt of La Sommata, erupted 50 Ka ago to the South of La Fossa. However, besides the relevant age difference between La Sommata eruption and La Fossa cone formation, the lack of CO2 in La Sommata phenocryst inclusions (detection limit of 50 ppm) indicates that the presently degassing magma must be located at a depth greater than the 3–3.5 km estimated for La Sommata magma [Gioncada et al., 1998] where crystallization and degassing in an open system brought to the almost total release of the least soluble CO2.

[15] In conclusion, data are coherent in indicating that the episodes of anomalous, high degassing of La Fossa crater correspond to an increasing release of high-T, CO2-He rich magmatic fluids. The last of these episodes occurred in 2004 and 2005. Unpublished seismological data of the INGV-Catania Section (A. Bonaccorso, personal communication, 2006) indicate that each CO2 flux peak temporarily coincided with an anomalous increase of shallow low-energy seismicity at La Fossa, as it had been observed during the 1988 degassing episode [Chiodini et al., 1992]. The concurrence of the compositional variation at the fumaroles, caused by the displacement of the hydrothermal fluids by the magmatic component, and the strong increase of the CO2 degassing rate from the entire La Fossa crater area point to a generalized increase of the pore pressure within the volcanic system that could be the cause of the local shallow seismicity. On the other hand, this local seismicity could be at least partly due to a variation of the stress regime of the crater area, caused by tectonic earthquakes localized in the proximity of the island, that actually frequently precede the anomalous degassing and the seismic swarms of La Fossa [Bonaccorso, 2002]. In both cases the increasing gas output would reflect the permeability increase by seismic fracturing of the impervious rock pile above a continuously degassing deep magmatic source.

[16] This interpretation does not require any magma upraise, for which, in addition, there is no geophysical evidence, as no significant uplift of La Fossa cone has been detected [Barberi et al., 1991; Bonaccorso, 2002]. Therefore, strictly speaking, the anomalous degassing episodes would not reflect a volcanic unrest process caused by magma movement, but rather a pulsating degassing process from a deep pressurized, possibly stationary, magma body.


[17] This work was partially funded by the Italian Dipartimento della Protezione Civile in the frame of the 2004–2006 Agreement with Istituto Nazionale di Geofisica e Vulcanologia - INGV. The authors are grateful to Franco Tassi for providing the analytical fumarolic data of the 2002–2004 period.