Fig. 1. The study area: (right) map of Mount Etna, eastern Sicily, showing the location of sampling sites A–D. The grey line is the division between volcanic deposits and sedimentary rocks. Measurements at site D, located about 20 km from the summit vents, are considered as representative of background ambient air, while a very dilute and aged plume was sampled at sites A–C on the Etna's upper flank; (left) detail of the volcano summit, showing the location of sampling sites on the near-vent plumes of Voragine (VOR) and North-East (NEC) summit craters.

Fig. 2. Scatter diagram of total gaseous mercury (Hg_(g)) versus SO_2(g) concentrations in the near-vent volcanic plumes of NEC (grey circles) and VOR (white circles) summit craters of Etna. The composition of ambient air on the volcano's upper flank is denoted by the grey area, based on measurements at sites A–D.

Fig. 3. Particulate (Hg_(p)) versus total gaseous (Hg_(g)) mercury in Mt. Etna's plume as measured adjacent to the North East and Voragine craters (circles, this study), contrasted against compositions of other quiescent volcanic plumes (data from Bagnato, 2007 and Witt et al., 2007): SQUARES—Miyakejima volcano (Japan); triangles—Masaya volcano (Nicaragua); stars—La Soufrière volcano (Guadeloupe, Lesser Antilles). The composition of ambient air is marked by the grey area. The arrows indicate the mean particulate to gaseous proportions in Etna (0.01) and Masaya (0.05) plumes, respectively (see Witt et al. (2007) for further details about the Masaya measurements).

Fig. 4. Results of model calculations (carried out with HSC software) simulating mercury speciation at thermodynamic equilibrium within Etna's high-T magmatic emissions. The logarithmic molar concentrations calculated by the model of the most-relevant Hg-bearing gaseous or solid species are plotted against the corresponding equilibrium temperature (T °C). All model runs were carried out for 0.1-MPa pressure, and the initialising gas compositions are shown in Table 2. Model run results shown: (a) when initialised with 100% magmatic gas composition and no atmospheric dilution, to simulate the pristine plume; and (b) when initialised with a mixture of “60% magmatic gases and 40% ambient air”. is the prevalent equilibrium form of mercury for all simulated conditions. In both diagrams, the grey shaded areas indicate the conditions most likely prevailing during magmatic degassing at Mount Etna.

Fig. 5. A schematic showing the major pathways governing Hg chemistry in the atmosphere (after Ryaboshapko et al., 2002). They can be summarised as: (i) oxidation of Hg⁰ to Hg^II in the gas phase, (ii) gas/droplet partitioning of Hg⁰ and Hg^II in the presence of clouds and fog, (iii) scavenging of Hg_(p) by droplets and dissolution in water, (iv) oxidation of Hg⁰ to Hg^II in droplets, (v) reduction of Hg^II to Hg⁰ in droplets, and (vi) adsorption/desorption of Hg^II onto insoluble solid particles within the droplet.

Our estimated Hg emission rates from Mt Etna compared with previous evaluations of the Hg flux from Etna

Concentrations of the different Hg species in the near-vent volcanic plumes of Voragine (VOR) and North-East (NEC)

Bulk plume Hg/SO₂ ratios were estimated from total (Hg_(g)+Hg_(p)) Hg contents, except where indicated by the asterisk, where Hg_(g) was used as a proxy for Hg. Hg emission rates (Φ Hg) from the volcano were evaluated by scaling these Hg/SO₂ plume ratios by remotely sensed observations of total SO₂ emission from Etna, taken on the same day by performing traverses below the plume on Etna's upper flank.

Compositions used to initialise the HSC software for equilibrium speciation calculations (in mol%)

The “pure magmatic” composition is representative of high-temperature undiluted emissions from Etna, modified from Aiuppa et al. (2005) and Bobrowski et al. (2007) to include Hg. The mixture of 60% magmatic gas and 40% air is the one for which Bobrowski et al. (2007) found the best agreement between modelled and experimentally determined BrO plume column amounts.

List of mercury species and compounds integrated into the HSC volcanic-gas model