Total volatile flux from Mount Etna

Authors


Abstract

[1] The Total Volatile (TV) flux from Mount Etna volcano has been characterised for the first time, by summing the simultaneously-evaluated fluxes of the three main volcanogenic volatiles: H2O, CO2 and SO2. SO2 flux was determined by routine DOAS traverse measurements, while H2O and CO2 were evaluated by scaling MultiGAS-sensed H2O/SO2 and CO2/SO2 plume ratios to the UV-sensed SO2 flux. The time-averaged TV flux from Etna is evaluated at ∼21,000 t·day−1, with a large fraction accounted for by H2O (∼13,000 t·day−1). H2O dominates (≥70%) the volatile budget during syn-eruptive degassing, while CO2 and H2O contribute equally to the TV flux during passive degassing. The CO2 flux was observed to be particularly high prior to the 2006 eruption, suggesting that this parameter is a good candidate for eruption prediction at basaltic volcanoes.

1. Introduction

[2] Measuring the total flux of volatiles emitted from active volcanoes provides constraints on the rates of magma supply and degassing in magmatic plumbing systems, and as such contributes to the monitoring of volcanic activity [Oppenheimer, 2003]. Although volcanoes emit a large variety of chemical species, this Total Volatile flux (TV flux) can be approximated to the flux of the three main volcanic gases, H2O, CO2 and SO2.

[3] SO2 has been focused on for volcano-geochemical investigation due to its low ambient concentration and strong UV absorption bands, which expedite detection by remote sensing. Since the early application in Volcanology of the Correlation Spectrometer in the 1970s [Stoiber et al., 1983], in fact, the SO2 flux has been extensively measured for volcano monitoring purposes, demonstrating its utility as a tracer of magmatic processes [Edmonds et al., 2003; Caltabiano et al., 2004], and, in some circumstances [e.g., Daag et al., 1996], as a precursor to eruptive activity.

[4] In contrast, measurement of H2O and CO2 fluxes, the two major volcanic gas components, has long been a goal of volcanology, yet one frustrated by the dwarfing of the volcanogenic signals by these species' high atmospheric background levels. Whilst Fourier Transform Infrared (FTIR) spectroscopy has been used to record the total gas composition from some volcanoes, e.g., Stromboli [Burton et al., 2007], Yasur [Oppenheimer et al., 2006], Masaya [Burton et al., 2000] and Nyiragongo [Sawyer et al., 2008], these data are not widespread; in particular FTIR H2O and CO2 data have been collected only sporadically on Etna [Allard et al., 2005]. As such, while Etna's SO2 flux is well characterised [Caltabiano et al., 2004], its H2O and CO2 fluxes remain poorly constrained.

[5] Recently, Aiuppa et al. [2007] and Shinohara et al. [2008] reported H2O, CO2 and SO2 concentration time series for the gas plume emitted from Etna's summit vents, using the MultiGAS technique. Here, we combine these data with new MultiGAS measurements and remotely-sensed SO2 fluxes (all obtained between January 2005 and July 2007). In so doing, we make the first assessment of Etna's H2O flux, and extend further the presently limited CO2 emission database [Allard et al., 1991; Aiuppa et al., 2006]. Finally, by summing up H2O, CO2 and SO2 contributions, we derive a first-order quantitative evaluation of the TV flux from Etna, and discuss the wider utility of such data for the geochemical forecasting of basaltic eruptions.

2. Methods and Volcanic Activity

[6] H2O, CO2 and SO2 concentrations in Etna's gas plume were simultaneously measured by an in-situ MultiGAS (Multi-component Gas Analyser System) unit, consisting of a Licor Li-840 NDIR spectrometer (for H2O and CO2) and a Membrapor S-100 electrochemical sensor (SO2), as detailed by Aiuppa et al. [2007]. The MutiGAS data used in this study are from Aiuppa et al. [2007, Figure 3]; they were collected in a series of discrete field surveys on Etna's summit area, mostly at Voragine (Figure 1), the most profusely degassing crater during the study period, and also, to a lesser extent, at the Bocca Nuova and North–East craters. These MultiGAS results were here combined with contemporaneously measured SO2 fluxes, obtained by traversing underneath the plume (∼3–5 km downwind of the summit) by car with a zenith-pointing Ocean Optics USB2000 spectrometer [Galle et al., 2003].

Figure 1.

Map of Etna's central craters, with location of sampling points (modified from Neri et al. [2008]). Solid lines with arrows at each end denote tracks of walking traverses performed during the July 2007 campaign.

[7] These discrete survey data were augmented with routine CO2 and SO2 measurements from a second MultiGAS device (with no H2O monitoring capability) permanently installed at Voragine crater. The derived CO2/SO2 ratios [Aiuppa et al., 2007], are here combined with SO2 flux data from the network of UV scanners continuously operating on Etna [Salerno et al., 2004], to provide a higher time resolution record of CO2 emissions from the volcano.

[8] In the investigated period, Etna was characterized by quiescent (passive) degassing from March 2005, when the 2004–2005 effusive eruption terminated, until mid July 2006. Strombolian and effusive activity again took place at the South–East Crater (SEC) from 14–24 July, followed by a quiescent phase, lasting until August 31. Further violent activity then resumed, at the SEC, until December 2006 (see reports at www.ct.ingv.it for details). Etna was passively degassing in July 2007, when our final measurements were performed.

3. Results

[9] The composition of Voragine crater's gas plume varied considerably and rapidly during the study period (Figures 2a and 2b), as adduced from the discrete measurements, with CO2/SO2 and H2O/SO2 ratios ranging 0.5–17 and 7–60, respectively. Measurements taken in May–June 2006 showed particularly high CO2/SO2 ratios (12 to 17; Figure 2a): these are part of a main phase of increasing CO2/SO2 ratios detected by the permanent Multi-GAS, previously interpreted as reflecting pre-eruptive ascent of CO2-rich magmas, triggering the onset of the July 2006 eruption [Aiuppa et al., 2007]. The peculiar CO2-rich and SO2-poor compositions in May 2006 are also supported by the low H2O/CO2 ratios (∼1), and by the largest H2O/SO2 ratios (∼60) in the Voragine dataset (Figure 2b). A distinct volcanic gas composition was instead observed at Voragine following the resumption in eruptive activity on July 14 2006, when the CO2/SO2 (0.5–4) and H2O/SO2 (7–25) ratios were significantly lower, and the H2O/CO2 ratios higher (range, 2–42), than during May–June 2006.

Figure 2.

Temporal plots (January 2005–December 2006) of (a) CO2/SO2 and (b) H2O/SO2 molar ratios from Etna's summit crater plumes, derived from the discrete campaign MultiGAS measurements (VOR, Voragine; BN, Bocca Nuova; NEC, North–east crater; data from Aiuppa et al. [2007]). (c) SO2 (red, right scale) and CO2 (blue, left scale) fluxes (all fluxes in t·day−1); the former obtained from car traverses, and the latter by multiplying these SO2 emission rates by the Voragine crater CO2/SO2 ratios shown in Figure 2a. (d) H2O and total volatile (TV) fluxes; the former derived by scaling Voragine's H2O/SO2 by the SO2 fluxes, and the latter by summing the CO2, SO2 and H2O flux series. (e) CO2 fluxes obtained by multiplying data from the permanently installed Multi-GAS at Voragine crater and the network of UV scanning spectrometers. The unshaded areas represent periods of passive degassing from the central craters; whilst the grey areas mark periods of semi-continuing effusive and mild-explosive eruptive activity at the South–East Crater.

[10] Whilst the Bocca Nuova and North–East crater emissions were characterised with less detail than those from Voragine (Figures 2a and 2b), our measurements suggest a reasonable consistency between the compositions of Bocca Nuova and Voragine, with the former crater emissions peaking at a higher CO2/SO2 ratio (26) during May 2006. The North–East crater emissions varied a little more randomly (CO2/SO2 ranging 1.7–7; H2O/SO2 ranging 16–80) over the measurement period.

[11] In order to estimate H2O and CO2 fluxes (Table 1), we multiplied the discrete campaign MultiGAS CO2/SO2 and H2O/SO2 ratios by the UV-sensed SO2 fluxes (shown in Figure 2c). Note that whilst SO2 flux is representative of the bulk plume (e.g., it integrates gas contributions from all the actively-degassing vents), MultiGAS measurements are representative of individual vent compositions, which may differ among each other to some extent. As such, we have taken Voragine's ratios as proxies for bulk plume composition in the H2O and CO2 flux calculations, with results shown in Figures 2c and 2d. This assumption was made on the basis that: (i) visual observations confirmed Etna to be dominantly degassing through the Voragine crater during 2005–2006; (ii) where measured in tandem, the spreads in the three vents' ratios were relatively minor, typically within a factor of ∼1.5, ranging at most by a factor of ∼3.5 (Figures 2a and 2b).

Table 1. Fluxes of SO2, CO2, H2O, With Corresponding Errors (±)a
DateSO2 Flux t·day−1(±)CO2 Flux t·day−1(±)H2O Flux t·day−1(±)TV Flux t·day−1(±)
  • a

    CO2 and H2O fluxes were derived by scaling Voragine's CO2/SO2 (±20 %) and H2O/SO2 (±30 %) ratios to SO2 fluxes (±20 %). Total volatile (TV) fluxes were calculated by summing the CO2, SO2 and H2O fluxes.

07/05/20051711(342)11413(3228)4762(1717)17886(3672)
27/08/20051422(284)4985(1410)9381(3382)15788(3676)
28/08/20051236(247)4420(1250)10126(3651)15783(3867)
29/08/20051051(210)4336(1226)9300(3353)5387(3577)
02/09/2005738(148)2812(795)4141(1493)7691(1698)
03/09/2005982(196)3705(1048)6062(2186)10748(2432)
23/05/20061064(8213)9508(2689)17503(6311)28075(6863)
07/06/20061631(326)19062(5392)13257(4780)33950(7213)
19/06/20061657(331)7065(1998)14450(5210)23172(5590)
16/07/20065698(1140)1959(554)33656(12135)41313(12201)
19/07/20064910(982)1688(477)24165(8713)30763(8781)
18/08/20062260(452)1553(4394448(1604)8261(1723)
19/08/20062209(442)1974(558)15344(5532)19527(5578)
05/09/20063283(657)1129(319)18469(6659)22881(6699)
06/09/20062661(532)1280(362)16762(6044)20703(6078)
10/09/20063679(736)1264(358)11380(4103)16323(4184)
04/10/20064463(893)2761(781)10167(3666)17392(3853)
25/10/20063104(621)9603(2716)19643(7082)32350(7611)
01/11/20062173(435)6275(1775)5134(1851)13582(2601)

[12] The calculated CO2 fluxes are plotted in Figure 2c, showing high (20,000 t·day−1) values before the onset of the July 2006 eruption, and substantially reduced (<5,000 t·day−1) emissions after the eruption started. This trend is also manifest in the results derived from the permanent MultiGAS and the network of continuously-operating UV scanners, shown in Figure 2e. The higher temporal resolution of the latter dataset enables identification of several distinct peaks in CO2 flux, not captured in the discrete campaign data, prior to a resumption of eruptive activity in August 2006, and corresponding to major explosive episodes in October–November 2006, respectively.

[13] According to our measurements, the H2O flux from Mount Etna ranged 4,400–33,600 t·day−1 during the study period (Figure 2d), averaging ∼13,000 t·day−1. In contrast to the CO2 fluxes, H2O emissions were larger during the July 2006 eruptive phase than in the pre-eruptive period. The temporal changes in 2006 H2O flux largely mirrored trends in SO2 emissions, which showed a factor ∼6 increase with the onset of eruptive activity in July (Figure 2c).

[14] By summing these H2O, CO2 and SO2 emissions, the TV flux from Mount Etna (Table 1 and Figure 2d) is hereby evaluated at 5,400–41,000 t·day−1 during the study period. Assuming this to be representative of Etna's long term behaviour, the volcano emits on average as much gas during passive degassing as during eruptive phases (∼21,000 t·day−1). Interesting, the TV flux is primarily (∼70%) H2O during eruptive periods, while CO2 and H2O emission rates are nearly equal during passive degassing. SO2 accounts for a relatively-minor fraction of the TV flux both during passive (∼8%) and eruptive (∼17%) degassing.

[15] In order to empirically validate our applied methodology of scaling the 2005–2006 MultiGAS ratios of the dominantly degassing crater (then Voragine) by the bulk SO2 fluxes, we performed an additional field survey, in July 2007, during which we constrained TV fluxes by performing walking traverses to determine the SO2 fluxes from the individual craters [McGonigle et al., 2002]. We then combined these data with simultaneous MultiGAS observations at each target to derive the CO2 and H2O flux from each crater, and summed these, with the SO2 fluxes to derive the total crater and volcanic TV (∼7,000 t·day−1) fluxes shown in Table 2. The latter emission rate falls within the range presented above (5,400–41,000 t·day−1). Furthermore, simple scaling of the total SO2 flux by the CO2/SO2 and H2O/SO2 ratios of the then dominantly degassing (North–East) crater, to compute TV emissions, results in the rather similar value of ∼7,700 t·day−1, providing confidence that first order TV flux assessments can be justifiably made on this basis. Indeed this calculation is rather insensitive to which of the significantly degassing crater ratios are used; even using the Voragine ratios (degassing from this crater was rather modest at that time following several minor collapse events), resulted in TV emissions of ∼7,200 t·day−1.

Table 2. Results of Individual Crater Measurement Performed on July 24–25, 2007a
 H2O/SO2 (molar)CO2/SO2 (molar)SO2 Flux t·day−1CO2 Flux t·day−1H2O Flux t·day−1TV Flux t·day−1
  • a

    (Errors in parentheses). Etna was passively degassing by the time these measurements were obtained. The molar gas ratios (measured by MultiGAS) fall within the range observed in 2005–2006 (see Figure 2); and confirm contrasting compositions for the different Etna's craters [Aiuppa et al., 2006], which we ascribe to ascent and degassing of CO2-rich magmas through the Central Craters (Voragine and Bocca Nuova), followed by intrusion (and degassing) of CO2-depleted magmas below North–east and South–east craters. The SO2 fluxes were obtained during walking traverses along the tracks shown in Figure 1; CO2 and H2O fluxes were calculated by multiplying these data together (as the Bocca Nuova and Voragine plumes were indistinguishable in the traverses, we have multiplied the combined flux from these two, by MultiGAS data taken at site A in Figure 1, fumigated by both plumes). The TV flux was generated by summing together the Total Species fluxes. Note that North–east crater was the main gas source at this time (as confirmed by visual observations [Miraglia, 2007]) and since mid-2007, when several minor collapse events mitigated against degassing from Voragine (which was dominant in 2005–2006).

North–east crater36 (11)2.2 (0.4)352 (70)545 (151)3564 (1283) 
Bocca Nuova crater11 (3)6.5 (1.3)    
Voragine crater25 (7)5.5 (1.1)    
Combined central craters18 (5)5.7 (1.1)233 (47)961 (272)1180 (425) 
South–east crater20 (6)0.8 (0.2)26.5 (5.3)15 (4)149 (53) 
Total species flux from all craters  612 (85)1520 (312)4893 (1352)7025 (1390)

4. Discussion and Implications for Volcano Monitoring

[16] The simultaneous measurement of H2O, CO2 and SO2 in volcanic plumes is relatively sporadic [e.g., Shinohara and Witter, 2005], such that TV fluxes are unconstrained for several open-conduit volcanoes [Oppenheimer, 2003]. Our measurements demonstrate that Etna releases on average ∼21,000 tons of volatiles each day, a large fraction (50–70%) of which are previously unmeasured H2O emissions (average, ∼13,000 t·day−1). We stress that this H2O flux is likely to be entirely magmatic in origin, as our measured H2O/S weight ratios for the bulk plume (∼13) closely match those of primitive silicate melt inclusions (∼11 [Spilliaert et al., 2006]). According to our calculations, Etna emissions would thus correspond to ∼1.6% of global H2O fluxes from arc volcanism (∼800,000 t·day−1 [Wallace, 2005]). In order to sustain such large H2O and TV fluxes, high pre-eruptive gas contents (∼5 wt. % [Spilliaert et al., 2006]) and magma supply rates (3.7 m3·s−1 as long-term average [Allard et al., 2006]) are required.

[17] Our systematic TV flux record also allows some considerations to be made on the potential of gas monitoring in the forecasting of eruptions at Etna, and other basaltic volcanoes. We show that the SO2 flux, whilst having been used in volcano monitoring for several decades at Etna, accounts for only a minor fraction (<17%) of the TV flux. Thus, while the role of SO2 flux measurements in volcanology remains invaluable, and in spite of increasing SO2 flux trends have been reported before eruptions on Etna [Bruno et al., 1999], we stress that extension of geochemical observations to include other species' fluxes appears imperative. CO2, in particular, is characterised by deep exsolution from mafic silicate melts; CO2-dominated fluid inclusions in primitive olivines formed at P = 650–700 MPa, have been retrieved on Etna [Kamenetsky and Clocchiatti, 1996]. As such, measurements of CO2 flux definitely provide the advantage of expanding our view on pre-eruptive degassing of decompressing magmas deeper in the volcano's plumbing system than possible with SO2 alone (the latter being released by Etnean magmas only for P < 100–140 MPa [Spilliaert et al., 2006]. It has recently been proposed, based on melt inclusion investigations [Spilliaert et al., 2006], that the shallow (<10 km bsl) Etnean plumbing system is persistently flushed by deep-rising CO2-rich gas bubbles, and that a CO2-rich gas phase (XCO2 ∼0.65) coexists at depth with the basaltic melt within the volcano's main magma storage zone (at depth of ∼5 km bsl [Spilliaert et al., 2006]). In this context, the ten-fold increases of the CO2 flux 1–2 months before onset of the 2006 eruption (Figures 2c and 2e) suggest that the accumulation at depth and pre-eruptive ascent of such CO2-rich magmas, leading to over-pressurization of the shallow volcano's plumbing system and finally triggering Etna's eruptions [Patanè et al., 2003; Allard et al., 2006; Spilliaert et al., 2006], can be tracked by a pre-eruptive release of CO2-rich volcanic gases at the surface. We thus propose that increasing the number of CO2 flux determinations at volcanoes should be a priority target of volcano gas geochemistry in the years to come. Comparison between Figure 2c and 2e demonstrates, however, that continuous observations, rather than discrete measurement surveys, are required to track the pre-eruptive fast ascent of basaltic magmas; and that our recently-acquired capability to perform such measurements in real-time and with relatively-cheap instrumentation should be exploited further.

[18] Measurements of the H2O flux are also likely to become sources of additional information. However, further data are required to test if pre-eruptive increases of the H2O flux can be resolved at basaltic volcanoes: our preliminary measurements highlight that no clear peak in H2O flux was captured before the 2006 eruptions, rather a syn-eruptive increase in the second half of 2006 (Figure 2d). The observation that H2O fluxes essentially mimic variations in SO2 release (Figures 2c and 2d), reflects the species' relatively-shallow degassing (compared to CO2) from Etnean magmas.

Ancillary