Review of the evolution of geochemical monitoring, networks and methodologies applied to the volcanoes of the Aeolian Arc (Italy)

Accepted Manuscript Review of the evolution of geochemical monitoring, networks and methodologies applied to the volcanoes of the Aeolian Arc (Italy)

Salvatore Inguaggiato, Iole Serena Diliberto, Cinzia Federico, Antonio Paonita, Fabio Vita PII: DOI: Reference:

S0012-8252(17)30130-7 doi:10.1016/j.earscirev.2017.09.006 EARTH 2486

To appear in:

Earth-Science Reviews

Received date: Revised date: Accepted date:

9 March 2017 7 August 2017 7 September 2017

Please cite this article as: Salvatore Inguaggiato, Iole Serena Diliberto, Cinzia Federico, Antonio Paonita, Fabio Vita , Review of the evolution of geochemical monitoring, networks and methodologies applied to the volcanoes of the Aeolian Arc (Italy). The address for the corresponding author was captured as affiliation for all authors. Please check if appropriate. Earth(2017), doi:10.1016/j.earscirev.2017.09.006

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT Review of the evolution of geochemical monitoring, networks and methodologies applied to the volcanoes of the Aeolian Arc (Italy) Salvatore Inguaggiato, Iole Serena Diliberto, Cinzia Federico, Antonio Paonita, Fabio Vita Istituto Nazionale di Geofisica e Vulcanologia Via Ugo la Malfa 153, Palermo, Italia

AC

CE

PT E

D

MA

NU

SC

RI

PT

Abstract Fluids discharged from volcanic systems are the direct surface manifestation of magma degassing at depth and provide primary insights for evaluating the state of volcanic activity. We review the geochemical best practice in volcanic surveillance based to a huge amount of monitoring data collected at different active volcanoes using both continuous and discontinuous approaches. The targeted volcanoes belong to the Aeolian Arc located in the Tyrrhenian Sea (Italy), and they have exhibited different activity states during the monitoring activities reported here. La Fossa cone on Vulcano Island has been in an uninterrupted quiescent stage characterized by variable solfataric activity. In contrast, Stromboli Island has shown a persistent mild explosive activity, episodically interrupted by effusive eruptions (in 1985, 2002, 2007, and 2014). Panarea Island, which is the summit of a seamount rising from the seafloor of the southern Tyrrhenian Sea, showed only undersea fluid release. The only observable clues of active volcanism at Panarea Island have been impulsive changes in the undersea fluid release, with the last submarine gas burst event being observed in November 2002. The geochemical monitoring and observations at each of these volcanoes has directly involved the volcanic plume and/or the fumarole vents, thermal waters, and diffuse soil degassing, depending on the type of manifestations and the level of activity encountered. Through direct access to the magmatic samples (when possible) and the collection of as much observable data related to the fluid release as possible, the aim has been (i) to verify the thermodynamic equilibrium condition, (ii) to discern among the possible hydrothermal, magmatic, marine, and meteoric sources in the fluid mixtures, (iii) to develop models of the fluid circulation supported by data, (iv) to follow the evolution of these natural systems by long-term monitoring, and (v) to support surveillance actions related to defining the volcanic risk and the evaluation and possible mitigation of related hazards. The examples provided in this review article show the close relationships among data analysis, interpretation, and modeling. We particularly focus on describing the fieldwork procedures, since any theoretical approach must always be verified and supported by field data, rather than just by experiments controlled in laboratory. Indeed the natural systems involve many variables producing effects that cannot be neglected. The monitored volcanic systems have been regarded as natural laboratories, and all of the activities have focused on both volcanological research and surveillance purposes in order to ensure that these two goals have overlapped. An appendix is also included that explains the scientific approach to the systematic activities, regarding geochemical monitoring of volcanic activity.

ACCEPTED MANUSCRIPT

1. Introduction People are both deeply fascinated by and often fear volcanoes. An erupting volcano transmits the power and beauty of nature, but at the same time may cause catastrophic events and risks that remain unpredictable and inescapable. Wealthy civilizations have grown nearby active volcanoes, and sometimes have also been destroyed. Some of the active volcanoes producing the most

PT

spectacular and/or damaging eruptions worldwide in recent centuries are Mount Vesuvius in Naples, Italy (in 1631); Krakatoa volcano in Indonesia (in 1883); Mount Peleé in Martinique (in

RI

1902); Bezymianny in Kamchatka, Russia (in 1956); Mount St. Helens in Cascades Range, USA (in 1980); Nevado del Ruiz in Colombia (in 1985); Pinatubo in the Philippines (in 1991), Ontake in

SC

Japan (2014) and Calbuco in Chile (in 2015). Since ancient times humans have realized the importance of understanding natural events and so have studied volcanoes and the effects of their

NU

activity. This has led to many volcanic observatories worldwide that study and forecast volcanic eruptions. Italy has the highest density of active volcanoes in Europe (including Vesuvius, Campi

MA

Flegrei, Ischia, Aeolian Arc, Mt Etna, Pantelleria, and other seamounts), and many of these are very close to urban areas. The Osservatorio Vesuviano in Naples, founded in 1841 on the flanks of

D

Mount Vesuvius, is the oldest scientific institution in the world that studies volcanic activity. The state-of-art volcanic monitoring is presently performed mainly by the Istituto Nazionale di

PT E

Geofisica e Vulcanologia (INGV), which was formed in 2001 by combining several Italian research institutes.

CE

Volcanic monitoring has evolved progressively, turning from essentially visual monitoring to observational methodologies based on geophysical and geochemical approaches. This review article presents the type of results that can be obtained by applying a geochemical approach to

AC

volcano monitoring and surveillance. We focus on applied theoretical concepts (primary based on the classic thermodynamic equilibrium theories), the techniques for measuring observable variables, and their evolution over time. We illustrate each of these aspects by presenting case studies from three active volcanoes belonging to the Aeolian Arc.

2. The Geochemical Approach to Volcanic Systems 2.1 A Typical Volcanic System

ACCEPTED MANUSCRIPT Shallow magmas located beneath active volcanoes release volatiles both during eruptive activity and inter-eruptive periods, which are referred to as passive degassing. Fluids discharged from volcanic systems represent the only surface manifestation of a magma degassing at depth. The composition of fluids ascending directly from magma is characterized mainly by H 2O, CO2, SO2, H2S, HF, and HCl (condensable gases) plus a few noncondensable gases (e.g. H 2, N2, CO, CH4 , Ar and He ). The released fluids can be measured in order to investigate the physicochemical

PT

conditions of the magma and related processes. Intensive parameters such as the chemical composition of the discharged gases and their

RI

relative ratios give useful information about pressure, temperature, and O fugacity conditions (Gerlach & Nordlie, 1975; Gerlach, 1993; Giggenbach, 1980, 1987, 1996; Giammanco et al., 1998;

SC

Chiodini and Marini, 1998; Taran et al., 1998; Nuccio & Paonita, 2001; Paonita et al., 2002), while the isotopic compositions of stable elements (e.g., He, C, Ar, N, H, and O) indicate the origin of the

NU

gases and the physicochemical processes that took place during degassing and during their migration to the surface (Allard et al., 1983; Taran et al., 1986; Capasso et al., 1997; Nuccio &

MA

Paonita, 2001; Inguaggiato et al., 2004a). Dissolved volatiles in magma reflect the thermodynamic conditions at equilibrium (pressure, temperature, and O fugacity); any variation in their initial

D

composition at depth will depend on new thermodynamic conditions, which can be evaluated at

PT E

the Earth’s surface using geochemical modeling based on the composition of the emitted fluids. A decrease in total pressure, caused by fracturation of hosting rocks and/or ascent of a magma batch, normally triggers the exsolution of dissolved volatiles that will easily reach the surface

CE

thanks to their high mobility, thus providing information about the physicochemical conditions of the magmatic plumbing system system. Because the kinetics of several chemical reactions are

AC

much slower than the velocity of the ascending gases, the molecular compositions often undergo quenching phenomena, so that the mixture of gases reaching the surface carries information about the temperatures and pressures at depth, which are higher than the measured outlet values. This convert chemical markers of gases into useful geothermometers and geobarometers. The concentration of magmatic species or their molecular ratios can be determined via direct sampling of gases (fumaroles, bubbling gas vents, or soil emissions) or by telemetric observation methods. Extensive parameters (mass and energy outputs) such as the fluxes of CO 2, H2O, SO2, and heat released in volcanic areas are fundamentally important in providing information about the aerial extent of the ongoing volcanic phenomena and the arguably amount of magma involved in

ACCEPTED MANUSCRIPT volcanic activity. Moreover, coupling the intensive and extensive parameters yields basic information that is useful when formulating models of volcanic fluid degassing (Italiano et al., 1997; Brusca et al., 2004; Inguaggiato et al., 2011a).

2.2 Fluids Geochemical Approach

PT

The first step in the framework of the geochemical investigation of a volcanic system is characterizing the chemical and isotopic compositions of fluids, which allows the formulation of a

RI

geochemical model (Inguaggiato et al., 2011). Such a model can be used to interpret observed changes in any single investigated parameter. The geochemical approach aimed at volcanic

SC

surveillance consists of the following steps:

magmatic and hydrothermal ones.

NU

1. Identifying the main fluid endmembers involved in the studied system, focusing on the

2. Characterizing the chemical and isotopic compositions of these endmembers and their

MA

degree of mixing.

3. Identifying the types and degrees of interaction processes among gas, water, and rock.

PT E

D

4. Formulating a comprehensive geochemical model.

2.3 Geochemical Surveillance

CE

The geochemical monitoring of an active volcano aims at recognizing signals related to changes in volcanic activity (Inguaggiato et al., 2011b, 2011c). As magma ascends into the plumbing system

AC

and/or there is refilling of new batches, the volatiles dissolved therein are progressively released according to their relative solubilities. As they approach the surface these fluids—which are discharged during magma degassing—may interact with shallow aquifers and/or be released along the main volcano-tectonic structures. The emissions of volatiles are typically monitored at the following strategic sites: 1. Fumaroles (Gerlach & Nordlie, 1975; Giggenbach, 1980, 1996; Taran et al., 1986; Gerlach, 1993; Capasso et al., 1997; Giammanco et al., 1998; Nuccio & Paonita, 2001; Paonita et al., 2002; Liotta et al., 2010).

ACCEPTED MANUSCRIPT 2. Diffuse soil degassing (Chiodini et al., 1996, 1998, 2005; Favara et al., 2001; Diliberto et al., 2002; Cardellini et al., 2003; Brusca et al., 2004; Pecoraino et al., 2005; Werner and Cardellini, 2006 Cardellini, 2006; Inguaggiato et al., 2011a, 2012; Mazot et al., 2011). 3. Degassing associated with geothermal waters (Capasso et al., 1998; Chiodini & Marini, 1998; Taran et al., 1998, 2008; Inguaggiato et al., 2000, 2004b, 2005, 2010, 2016; Inguaggiato C. et al., 2016; Federico et al., 2004, 2013; Inguaggiato & Rizzo, 2004b;

PT

Liotta et al., 2006). 4. Plume (Oppenheimer et al., 1998; Bruno et al., 1999; Galle et al., 2003; Aiuppa et al.,

RI

2004, 2005a, 2009; McGonigle et al., 2005; Burton et al., 2007a, 2007b).

SC

3. Case Study: Aeolian Islands

NU

The Aeolian Islands are located in Italy between the southern Tyrrhenian Sea back-arc basin (northeast of Sicily) and the Calabrian Arc (Fig. 1a). The volcanic arc of the Aeolian Islands lies on

MA

an inclined seismic zone that extends down to a depth of 450 km beneath the Tyrrhenian Sea. The Aeolian Islands rise in the southern Tyrrhenian Sea after the subduction of the Ionian plate beneath the European plate (Fig. 1b), and they consist of seven islands of volcanic origin, three of

D

which are still active: Vulcano, Panarea, and Stromboli (Fig. 1a). The magmas of the Aeolian Islands

PT E

are typical of subduction zones, but over time they have evolved into more-basic compositions, with less SiO2 and more K: from andesites and basaltic-andesites, with associated dacites and

CE

rhyolites, to the shoshonites of Vulcano and Stromboli (De Astis et al., 2003, 2013). 4. Vulcano Island

AC

Vulcano is the southernmost island of the Aeolian Arc in the southern Tyrrhenian Sea. It is part of a volcanic complex (that includes Lipari Island) that developed inside a graben-like structure controlled by the north-northwest–south-southeast strike-slip Tindari–Letojanni fault (TLF) system (Gioncada et al., 2003; Ventura, 2013) (Fig. 1b). Vulcano Island is the exposed summit of a volcanic edifice rising from the local sea floor (ca. 1000 m b.s.l.) up to the maximum height of 499 m a.s.l. at Monte Aria. It is composed of volcanic products ranging from shoshonitic to high-K calcalkaline and potassic series (De Astis et al., 2013). Since the last eruption that occurred in AD 1888–1890, the volcano has exhibited fumarolic activity, with the main fluid releases occurring in the northern sector of the island, at La Fossa cone, and in the coastal area of Baia di Levante. The thermal

ACCEPTED MANUSCRIPT aquifers and widespread CO2 soil degassing at Vulcano Porto village and La Fossa cone have also provided evidence of volcanic activity (Capasso et al., 1997; Inguaggiato et al., 2012).

4.1 Monitoring of the Vulcano Porto thermal aquifer The Vulcano Porto area at the base of the La Fossa volcanic cone hosts a thermal aquifer that has received considerable interest from geochemists, for several decades (Fig. 2). Some drilling

PT

campaigns and several studies of shallow circulating fluids have attempted to characterize the physicochemical properties of the hydrothermal systems and their influence on the composition

RI

of the shallow aquifer (Carapezza et al., 1983; Panichi and Noto, 1992; Capasso et al., 1992, 1997,

SC

2000, 2001; Bolognesi and D’Amore, 1993; Aiuppa et al., 2000; Federico et al., 2010). Since the previous studies, the fluids separating from the Vulcano hydrothermal system appeared to be a

NU

promising tool for revealing deeper phenomena. Since 1977 the systematic monitoring of some thermal wells in the Vulcano Porto area, although only performed twice a year, aimed at the

MA

analysis of water chemistry and water stable isotopes (δD and δ18O) (Martini et al., 1979; Carapezza et al., 1983). The dataset, increasingly wealthy on Vulcano Porto cold and thermal wells, allowed the first characterizations of the shallow aquifer and the identification of the main

D

fumarolic steam contributions (Carapezza et al. 1983; Dongarrà et al., 1988; Capasso et al., 1992).

PT E

According to isotopic data, the Vulcano Porto aquifer was thought to be primarily fed by meteoric water, and locally entered by fluids coming from a brackish (Capasso et al., 1992; Chiodini et al., 1995) or a basicly meteoric geothermal aquifer, modified by magmatic fluids (Bolognesi and

CE

D’Amore, 1993). The shallow aquifer (in the first 100-200 m of depth) appeared as a very immature meteoric system, permanently fluxed by H2S vapour coming from underlying boiling

AC

aquifers, lying at 185-236 m depth (Bolognesi and D'Amore, 1993; Cortecci et al., 2001), as intercepted by exploratory wells (Sommaruga, 1984). Additional data on trace elements, dissolved gases, and carbon isotopes of dissolved CO2 have helped to discriminate between steam-heated groundwaters in the Levante-beach area and waters more directly fed by Cl-SO4-rich deep fluids that either condense from the fumarolic area along the flanks of the La Fossa cone or come directly from the central conduits (Bolognesi and D'Amore, 1993; Chiodini et al., 1996; Fulignati et al., 1996; Capasso and Inguaggiato, 1998; Aiuppa et al., 2000; Capasso et al., 1997, 2000, 2001). In Vulcano Porto, the shallow aquifer is made acidic by the continuous input of CO 2, and the dissolved salt content is either related to water-rock exchange at shallow levels or to the input of seawater, while the base of La Fossa cone would represent the area where near-neutral Na-Cl-rich and hot

ACCEPTED MANUSCRIPT fluids massively contaminate the shallow meteoric aquifer, also conveying some mobile trace metals (Capasso and Inguaggiato, 1998; Aiuppa et al., 2000). In this area, a single hot well (temperature about 60°C, called Camping Sicilia, Capasso et al., 1992; W2, Bolognesi and D'Amore, 1993; Sicilia well, Chiodini et al., 1996; VH2, Federico et al., 2010, Fig. 2) revealed variations in both water chemistry and stable isotopes, synchronous with those of crater fumaroles, several times, since 1979 (Fig. 3). The first time that the composition of crater fumarolic fluids (Martini et

PT

al., 1979; Bolognesi and D'Amore, 1993) and the fluid dissolved in this hot well (Bolognesi, 1997; 2000) changed concurrently, was at the time of the occurrence of a seismic sequence 10 km

RI

south-east of Vulcano, occurred on 15 April 1978. The main shock with M 5.5 (Del Pezzo and Martini, 1981), was thought to be responsible for the observed variations. In some pioneering

SC

models, water-stable isotopes and major-ion chemistry in the shallow aquifer were assumed to vary according to the changeable contribution of magmatic fluids to the hydrothermal aquifer,

NU

modulated by the stress-related fracturing (Martini, 1983; Bolognesi, 1997). Both, the chemical and isotopic compositions of Camping Sicilia hot well and the crater fumaroles showed related

MA

changes during recurrent unrest phases; the strongest one occurred in the period 1988-1993 (Capasso et al., 1992; Capasso et al., 1999). Federico et al. (2010) proposed a comprehensive

D

model, describing steam condensation and boiling phenomena. This model is based on differences

PT E

in partition coefficients between liquid water and vapor characterizing oxygen and hydrogen isotopes, as well as volcanic gases (CO2, S species, and HCl). According to the model by Federico et al. (2010), the pristine composition of the sampled thermal waters is modified by both the input of

CE

chemicals and enthalpy from the ascending fumarolic vapor and the consequent boiling and steam separation, differently than what is proposed by the simple mixing models (Fig. 4). In this light, the

AC

variations observed in Camping Sicilia well during the period 1988-90 can be explained by a different composition of the vapor entering the aquifer paralleled by a higher mass rate relative to the shallow meteoric endmember, which conveyed higher amounts of CO 2, HCl and S-bearing gases. Far from central conduits, the flux of deep vapour is minor, therefore a preferential partition of gases (e.g., CO2), with respect to both HCl and S-bearing gases occurs (Fig.5). Despite the initial working hypothesis, the relationship between these phases, chiefly characterized by the increased contribution of magmatic gases and peculiar seismicity, or regional tectonics, has not been unambiguously demonstrated (Madonia et al., 2013). The regular monitoring of the water table head, instead, has put in evidence that a pressure increase in the hydrothermal system, lasting three years, would have caused crack opening and the phase of

ACCEPTED MANUSCRIPT higher vapor output in the fumarole field late in 2004 (Capasso et al., 2014). The following pressure drop would have promoted further vapor separation from the hydrothermal system and finally the drainage of S-rich fluids into the shallow thermal aquifer. According to these authors, the chemistry of the shallow thermal aquifer would be greatly modified by the input of deep fluids only when hydraulically conductive fractures became more permeable. Actually, Madonia et al. (2015) have modelled the pattern of fluid circulation in Vulcano Island, by reconciling the data on

PT

water table head, temperature and electrical conductivity, recorded in groundwater, within the hypothesis of fluid circulation following both vertical and horizontal physical discontinuities. In La

RI

Fossa area, meteoric water would intercept upwelling hydrothermal fluids along vertical volcanotectonics faults and fractures, as inferred by the orientation of the geochemical anomalies, while

SC

the condensed vapour would flow roughly horizontally along volcano-stratigraphic discontinuities (as observed during the drilling of a borehole in 2004), towards the Vulcano Porto aquifer. The

NU

meteoric supply of the deep hydrothermal aquifer would occur along sub-vertical volcano-tectonic

MA

discontinuities, such as faults and caldera borders, downward from the Vulcano Piano plateau. 4.2 THE FUMAROLIC SYSTEM OF LA FOSSA

4.2.1 Geochemical monitoring and observations: Currently, Vulcano presents a wide fumarole

D

field in the northern part of La Fossa crater, with temperature ranging from 100°C to about 450°C.

PT E

This solfataric activity has displayed remarkable evolution in time. In 1978, an increase in the fumarolic activity was observed as an effect of the shock generated by the April 15th earthquake (M6.1), whose epicentre was located on a NW-SE trending fault (crossing Vulcano island) in the

CE

nearby Patti Gulf, about 20 km west of Vulcano island (Chiodini et al., 1992). It was the first occasion in recent times for scientists in Italy to closely observe an unrest of volcanic activity on

AC

the island. Since 1984, a comprehensive geochemical and geophysical surveillance network, now managed by the Italian Istituto Nazionale di Geofisica e Vulcanologia, has been implemented to monitor the evolving volcanic activity. For the fumarolic field of La Fossa, the geochemical monitoring program was based upon a network of stations for real-time monitoring of fumarole outlet temperatures (Barberi et al., 1991). Besides, periodic field surveys started to be performed, addressed to the collection of fumarolic gases from the vents. Thereafter, the periodic field surveys have been carried out from weekly to two-monthly, according to the ongoing level of activity (e.g. Badalamenti et al., 1991). Two vents, located within the crater (FA fumarole) and on the rim (F11 fumarole) were monitored during each campaign, while several other fumaroles were less frequently sampled. Since 2000, fourfumaroles were sampled during each campaign (F0 and

ACCEPTED MANUSCRIPT F5AT, in addition to FA and F11). Concentrations of major (H 2O, CO2), minor (S, HCl, N2) and trace (HF, H2, CO, CH4, He, Ne) species, as well as isotopic compositions of H, O, C and He, were routinely determined in all the samples. The geochemical monitoring has provided evidence that the fumarole system experiences episodes of remarkable change, each lasting no more than a few months. Typical signatures of these short-term anomalies are large increments in CO2, N2 (not shown), and He concentrations, 13

C/12C isotopic ratios (Fig. 6). A main episode took place in 1988-1990,

PT

coupled to increased

marked by quick and intense modifications of both the geochemical features and the output rate

RI

of fumarolic fluids. It was characterized by the increase of outlet temperatures (from 400 to 700°C

SC

in the period 1988-1993, Fig. 7) and steam output from fumarolic vents (Fig. 8)(Barberi et al., 1991; Badalamenti et al., 1991; Chiodini et al., 1993, 1995; Tedesco, 1995; Bukumirovic et al.,

NU

1997; Capasso et al., 1997; Italiano et al., 1998), enhanced concentration of acidic gas species (Figs. 7,8); Chiodini et al., 1995; Capasso et al., 1997), increase of exhaling surface area and steam

MA

output from fumarolic vents (Bukumirovic et al., 1997; Italiano et al., 1998). Marked variations were also observed in temperature, water level, and the chemical and isotopic compositions of thermal groundwaters (Capasso et al., 1992; Bolognesi and D’Amore, 1993), together with

D

increases in diffuse CO2 emissions from the soil of the Vulcano Porto area (Baubron et al., 1990;

PT E

Carapezza and Diliberto, 1993; Di Liberto et al., 2002). In particular, low pH values were due to increased CO2 flux at the base of “La Fossa” crater, while higher Cl and SO4 contents were driven

CE

by the hydrolysis of deep volcanic fluids (Capasso et al., 1999). A subsequent episode occurred in 1996 (Capasso et al., 1999; Nuccio et al., 1999) and was still characterized by geochemical anomalies (Fig.6), which returned to background values after 5

AC

months. The temperature of F5AT fumarole rapidly increased byalmost 150°C, from 350° to 500°C, and went back to background values after 5 months (Fig.7). The contents of acidic species in fumaroles (HCl and Stot; Fig.7) did not vary significantly compared to the 1988 event (Capasso et al., 1997b). Other minor and short-lasting changes were recorded in late 1998 (Chiodini et al., 2000; Paonita et al., 2002). After 2002, the CO2 content in the crater fumaroles has always been very low and the values of δ13CCO2 were generally lower than those achieved during the phases of unrest in 1988 and 1996 (>0‰; Capasso et al., 1999) (Fig. 6). Accordingly, R/Ra values were below 5.0 (Figs. 6,110) and

ACCEPTED MANUSCRIPT were on average lower than those recorded from 1988 to 1997 (5.5–6.2; Italiano and Nuccio, 1997). The temperature monitoring, together with periodic measurements in fumaroles, allowed the early recognition of a new episode of increasing activity in the late 2004 (Diliberto, 2011; Paonita et al., 2013). In particular, a new step of increasing temperatures led to values as high as 468°C in F11 fumarole and 440°C in F5AT fumarole (Fig.7). Conversely, fumaroles on the inner crater slopes

PT

did not show any clear trend of increasing temperature (FA fumarole). All fumaroles showed synchronous CO2 and δ13CCO2 increases (Fig.6), paralleled by He and N2. The CO2 output from the

RI

crater area increased dramatically too (Granieri et al., 2006). The Stot content doubled relative to

SC

the period just before the crisis onset (Fig.7). The helium isotope ratio did not show remarkable variations during the crisis (excluding a single datum for F11), but started a slow increase from 5.1

NU

to 5.7 R/Ra that lasted 2 years (Figs. 6,11).

The concentrations of all incondensable species came back to their background values in April

MA

2005, prior to the new crisis (Figs.6,7). In October 2005 a further anomaly was observed, as indicated by enhanced emission rates, temperature increases, and significant geochemical

D

variations in gas composition. In particular, a temperature as high as 468°C was recorded at F11 (Fig.7). The CO2, He, and N2 contents were again high (Fig.6) and, as generally observed at Vulcano,

PT E

the high CO2 contents corresponded to δ13CCO2 remaining near 0‰ (Fig.6). These parameters reverted to their pre-event values within a few months. After the 2004 and 2005 episodes, further

CE

changes in CO2, He, and N2 contents occurred with progressively lower intensities, as generally observed after the main anomalies. Whereas the contents of Stot roughly paralleled those of CO 2, the HCl variations were largely uncorrelated (Fig.7). Finally, the event in November 2009 had

AC

features similar to the previous ones (Fig.6). 4.2.2 Interpretative models of genesis and circulation of the fumarolic fluids: A main feature emerging from the wide fumarole dataset of Vulcano is the unambiguous correlation between CO 2 concentration and other geochemical parameters, such as He, N 2 (Fig. 9), δ13CCO2, partly HCl and S (Fig. 7), which has been interpreted as a result of mixing between magmatic and hydrothermal fluids by a large number of researchers (Chiodini et al., 1993, 1995, 2000; Tedesco, 1995; Capasso et al., 1997; Nuccio et al., 1999; Di Liberto et al., 2002; Leeman et al., 2005; Taran, 2011). Based on these correlations, it has been concluded that the magmatic fluid is richer in CO 2, He, N2,

13

C

(namely, high δ13CCO2) and Ar, and poorer in H2O, HCl, S and 2H (namely, low δDH2O) relative to the

ACCEPTED MANUSCRIPT hydrothermal vapors (Bolognesi and D’Amore, 1993; Tedesco, 1995; Tedesco and Scarsi, 1999; Capasso et al., 1997, 2001; Chiodini et al., 1993, 1995, 2000; Nuccio et al., 1999; Di Liberto et al., 2002; Paonita et al., 2002). It is noteworthy that the local magmatic gases would be marked by δ13CCO2 and δDH2O values that are more positive and negative, respectively, than those typical for the upper mantle, which would be linked to recycling of the subducted African slab (Capasso et al., 1997, 2001; Tedesco and Scarsi, 1999; Chiodini et al., 2000; Paonita et al., 2002). Coupled to the

PT

magmatic and hydrothermal components, minor contributions of volatiles from meteoric, air, and air-saturated fluids having shallower genesis, have been recognized to occasionally contaminate

RI

the fumarolic gases (Capasso et al., 1997; Tedesco and Scarsi, 1999, Paonita et al., 2002; Taran,

SC

2011).

In the framework of the two-endmember mixing, the estimation of the compositions of the fluids

NU

outgassed from the magmatic and the hydrothermal system, as well as their evolution in time, has greatly improved our knowledge of how these two systems work. Even if the scientific debate is

MA

still open about some questions, a number of processes have been modeled and clarified. 4.2.3 The hydrothermal system: Regarding the hydrothermal systems that would generate the

D

hydrothermal fluids circulating in the island of Vulcano, two main points of view can be found in the literature. According to the “dry model”, the hydrothermal endmember would derive from

PT E

marine water that is totally vaporized when infiltrating below the La Fossa edifice, due to contact with hot igneous rocks (Cioni and D’Amore, 1984; Chiodini et al., 1993, 1995, 2000). Zones of

CE

vaporization at different temperatures, that produce H2O-dominated fluids having different contents of HCl, HF, H2S, and SO2, can be recognized when comparing the concentrations of these species in the fumarolic fluids with the expected compositions of fluids in equilibrium with various

AC

hydrothermal mineral assemblages (Chiodini et al., 1993). Conversely, the “wet model” by Carapezza et al. (1981) consists in a two-phase hydrothermal system at a depth of 1-2 km. Nuccio et al. (1999) reconciled the two models by comparing the compositions of 1970s to the extrapolated hydrothermal fluid composition observed from 1988, the last one showing a decrease in CO2 and an increase in NaCl. They concluded that the “wet model” would have worked up to the end of the 1970s, with the hydrothermal system boiling at about 330°C and 15 MPa. An increase in the magmatic input due to the increasing volcanic activity during the second half of the 1980s would have caused the total vaporization of the central part of the brine, resulting in a onephase central column surrounded by a two-phase system at higher temperature and pressure with

ACCEPTED MANUSCRIPT respect to the 1970s conditions (390°C and 20 MPa). Since then, the vapors coming from the surrounding two-phase system have migrated toward the central column so as to mix with the ascending magmatic gases. The origin of the hydrothermal fluid is less debated, given that the δD H2O value of this endmember is similar to that of the local seawater (Chiodini et al., 1995 and 2000; Paonita et al., 2002). Seawater would however undergo a number of processes, while infiltrating through the hot rocks

PT

(e.g., water-rock interactions and boiling), that modify its O- and B-isotope compositions (Chiodini et al., 1995 and 2000; Paonita et al., 2002; Leeman et al., 2005) and partly the H-isotope

RI

composition (Paonita et al., 2002). Na-Ca exchanges between the water and the local rocks would

SC

also control the pH conditions of this fluid (Di Liberto et al., 2002). While Na is removed from the evolving seawater, Ca enters the solution, undergoes hydrolysis and produces HCl (Bischoff et al.,

NU

1996), lowering the pH of the water down to 2.5 (Fig.8) (Di Liberto et al., 2002). The increasing water-rock ratio within the hydrothermal system, marked by the growing output at the crater

MA

from 1980s to 1995, lowers the Ca availability, so the aqueous solution becomes less acidic (Di Liberto et al., 2002). Seawater flowing towards the boiling hydrothermal brine also dissolves a large quantity of pyrite along its path. In the boiling hydrothermal system, dissolved sulphur

D

precipitates as pyrite and anhydrite, in agreement with the paragenesis of hydrothermal alteration

PT E

minerals recovered in drilled wells at Vulcano (Di Liberto et al., 2002). Water interactions with rocks would also enrich the thermalized seawater in heavy He and Ar isotopes produced by radioactive decay (Tedesco and Nagao, 1996; Tedesco and Scarsi, 1999), although these isotopes

CE

could also come from an old magmatic reservoir (Tedesco, 1995). According to Taran (2011), the hydrothermal fluids would carry a generalized crustal component, which can be associated to

AC

contributions from both the subduction-related lithosphere and the crust below the volcano.

4.2.4 The magmatic system: The estimation of the composition of the magmatic endmember, feeding the fumarolic vents, by using the linear correlations between CO2 and minor species (see above), has followed two different approaches. Chiodini et al. (1993, 1995, 2000) used CO2-He-N2 diagrams, where all fumarolic data were plotted without any temporal distinction. They implicitly meant a fixedcomposition endmember, having 25 mol% CO2. On the other hand, in the chemical-enthalpic “mixing model”, developed by Nuccio et al. (1999), fumaroles were grouped based on their

ACCEPTED MANUSCRIPT sampling date. Nuccio et al., (1999) showed that a thermal balance between a constant magmatic gas (characterized as 950°C and 25 mol% CO2) and the local hydrothermal fluid (characterized as 380°C and 3 mol% CO2) provided temperatures below those of fumarolic emissions in several sampling campaigns. Based on their model, the same authors (Nuccio et al.,1999; Paonita et al. 2002) proposed a magmatic endmember that changes over time. Nevertheless, one main requirement in the “mixing model” was that fumarolic samples—collected at the same time—had

PT

very different compositions and outlet temperatures. This feature commonly occurred in the period 1988 to 1996, so as to guarantee suitable model results. Unfortunately, it has occurred

RI

much more rarely since 1996, which makes the model difficult to apply. More recently, Paonita et al. (2013) developed a different approach to constrain chemical (He/CO2 and N2/He) and isotopic

SC

(13C/12C, 2H/1H, and 3He/4He) ratios of the magmatic endmember. By taking advantage ofthe asymptotic shape of the mixing lines as achieved when plotting a chemical (or an isotope) ratio

NU

versus CO2, Paonita et al. (2013) showed conclusively that the composition of the magmatic fluid changed significantly in time, even if changes in mixing proportions also affected the fumarole

MA

geochemistry. The magmatic He/CO2 (Fig. 10), N2/He,

13

C/12C (Fig. 11b), and 3He/4He (Fig. 11a)

values throughout 1988–1996 differed from those feeding the anomaly at the end of 2004. Early

D

clues of the new magmatic fluid had appeared in 1998-1999, far from any short-term anomaly,

PT E

whereas new and old magmatic fluids have coexisted after 2004 (Fig. 10). Researchers also tried to give answers to two main questions involving the magmatic system, which also have serious implications for volcanic surveillance: (1) type and depth of the magma

CE

presently feeding the fumarolic activity, and (2) meaning of the periodic and short-duration peaks in CO2 (these latter ones being either evidence of volcanic unrest due to magma dynamics or

AC

simply changes in the permeability of the gas pathways). As concerns point (1), Clocchiatti et al. (1994a,b) coupled the SO2 output to the dissolved content of sulfur in the melts to conclude that a magma volume of 0.17 km 3, at the transition between basalt to latite, would feed the fumarolic activity (similar to 0.3 km 3 estimated by Frazzetta et al., 1991). Based on similar considerations, Granieri et al. (2006) assert that a basaltic magma is presently degassing at Vulcano. Magro (1997) compared the He-isotope ratios between fumaroles and volcanic products with different degrees of evolution to support the hypothesis of degassing of a basaltic magma. The S isotope compositions also addressed the involvement of a basaltic magma in feeding fumaroles after 1988 (Cortecci et al., 1996). In contrast, Nuccio and Paonita

ACCEPTED MANUSCRIPT (2001) and Paonita et al. (2002) proposed an acidic-intermediate melt by relating the magmatic fluid variations to those computed by their degassing model. Paonita et al. (2013) implemented the degassing model to predict C isotope fractionation and applied the H 2O-CO2-N2-He- δ13CCO2 degassing model to the estimated chemical and isotopic ratios of the magmatic endmember. These last authors concluded that a magma having a latitic composition is the main source of fluids for the present exhalative activity of La Fossa (Fig. 11b). The differences in He isotope ratio

PT

of the magmatic fluids of 1988-1996 with respect to 2004 also suggested the presence of two magma ponding levels at slightly different pressures (30 to 35 MPa), where bubble-melt

RI

decoupling can occur (Paonita et al., 2013). Such pressure converts to 3.0-3.5 km b.s.l., depths consistent with geophysical and structural evidence on the island (Peccerillo et al., 2006). The

SC

melts remained physically separated and compositionally distinct inside these two levels, thanks to a low hydraulic connectivity in this structurally complex magmatic reservoir. With the aim to

NU

test this result, Mandarano et al. (2016) performed an extensive sampling of rocks with various degrees of evolution, belonging to high-K calcalkaline–shoshonitic and shoshonitic–potassic series

MA

of Vulcano, and they measured the elemental and isotopic compositions of noble gases (He, Ne, and Ar) in olivine- and clinopyroxene-hosted fluid inclusions. A comparison of the He isotope ratios

D

between fluid inclusions and fumarolic gases showed that only the basalts of La Sommata and the

PT E

latites of Vulcanello have comparable values to the fumaroles (Fig. 12). Taking into account that the latites of Vulcanello relate to one of the most-recent eruptions at Vulcano (in the 17th century), the authors confirmed that that the most probable magma actually feeding the

CE

fumaroles is a latitic body. The He isotope ratio also showed a negative correlation with Sr isotopes for all the rock samples, except for the Vulcanello latite, which gave anomalously high values. The authors hypothesized the occurrence of flushing process by fluids coming from the

AC

deepest reservoirs, since an input of deep magmatic volatiles, with high 3He/4He values, would increase the He-isotope ratio, without changing

87

Sr/86Sr. The periodic supply of more-primitive

fluids (having a high 3He/4He) would ascend from a deeper and more-voluminous magma reservoir, located at a depth of about 5 km (Peccerillo et al., 2006). The He-isotope ratio of La Sommata-type basalt makes it the most probable source for these fluids (Mandarano et al., 2016). More rarely, ascent of melts could also occur from this deep reservoir, as testified by the presence of mingling textures in products of the last 6 ka of activity of La Fossa (Clocchiatti et al., 1994b; Piochi et al., 2009).

ACCEPTED MANUSCRIPT With regard to point (2), Chiodini et al. (1992) and Granieri et al. (2006) suggested that the temporal coincidence of the episodic peaks in CO2 (so-called “geochemical crises”) with increases of the local shallow seismicity account for an increase in pore pressure at a shallow depth. The crises would therefore correspond to moments of increasing volatile release from a stationary magma system, probably due to permeability increases by seismic fracturing. Increases in shallow high-frequency microseismicity were observed indeed during the 1988 crisis (Montalto, 1994), in

PT

1998, and in all of the CO2 peaks from 2004 to 2009 (Alparone et al., 2010; Milluzzo et al., 2010; Cannata et al., 2012; Harris et al., 2012), while it did not occur in 1996. In contrast, Nuccio and

RI

Paonita (2001) proposed that each event in the decade from 1988 to 1998 was due to an episode of magma vesiculation and outgassing, due to melt ascent and decompression. This would make

SC

the difference between a degassing event and some quieter periods of volatiles emanations from

NU

a stationary magma.

Each degassing crisis at Vulcano has also been accompanied by i) an increase of H 2O and CO2

MA

output from crater fumaroles, as revealed by the measurements of Italiano et al. (1998), Granieri et al. (2006) and Inguaggiato et al. (2012), and ii) a thermal increase at the gas vents, as real-time measurements of fumarolic outlet temperatures clearly show (Alparone et al., 2010; Diliberto,

D

2011; Harris et al., 2012). Moreover, the lack of both ground deformation (Gambino and

PT E

Guglielmino, 2008; Cannata et al., 2012) and deep seismicity (Chiodini et al., 1992; Cannata et al., 2012), correlated to the CO2 peaks after 2004, would exclude the occurrence of magma movements during these crisis. Taking into account the intensive (gas concentrations) and

CE

extensive (gas output) measurements, as well as the above geophysical information, Paonita et al. (2013) proposed an integrated explanation for these crisis of the volcanic activity. The key fact is

AC

that the magmatic fluids definitely modify their composition because they are coming from different levels of a shallow latitic reservoir, and the same different levels can simultaneously feed the fumaroles. As a corollary, the depth (and pressure) of magma vesiculation and/or gas-melt decoupling can change without the need for magma migration and ascent. The changes after 1996 would indicate that a new shallow level of the latitic reservoir, started to release new magmatic fluids, together with those already degassing from an higher depth. The 2004 event points to the dominant contribution from the new level. Paonita et al. (2013) figured out a foam-like structure growing at the top of the new level, where the magmatic gases can probably accumulate until they are massively released to produce a crisis. This would initiate the chain of the observed geochemical and geophysical signals: (i) displacement of the hydrothermal system far from the

ACCEPTED MANUSCRIPT mixing zone and a relative increase in the magmatic component, (ii) occurrence of high-frequency seismicity due to rock failures, (iii) a consequent increase in permeability, and (iv) increases in the gas output of fumaroles. The crises would then be due to gas accumulation at the top of magma bodies followed by massive escape, or activation of new degassing levels in the reservoir, for which the stress field plays a key role. Such a scenario explains the observed increases in both fumarole output and shallow high-frequency seismicity (due to increased pore pressure) during

PT

the anomalies, while being consistent with the concomitant absence of any deep seismicity or

4.2.5 Continuous Temperature Measurements

SC

RI

ground deformation.

NU

Direct temperature measurements began in 1984 at the main fumarole fields on Vulcano Island. A thermistor was inserted inside the vents at a depth of about 0.5 m below ground level, and

MA

measurements were made with direct contact between the sensor and the flowing fluid, which is considered the best approach for the long-term monitoring of high-temperature fumaroles (HTFs)

D

(Fig. 13a,b). Some of the technical specifications and data annotation procedures for continuous monitoring are provided in the appendix (Section 8.9).

PT E

The first map of the HTF field at the top of La Fossa cone, produced using a geographic information system, was reported by Bukumirovic et al. (1997). Since then the number of

CE

monitored fumarole vents has increased, and the spatial relationships with structural and volcanic features have been analyzed (Harris et al., 2012). The original monitoring sites have not been

AC

moved so as to avoid interrupting long-term data acquisition. The heat release at the surface is currently monitored by two temperature-monitoring stations in the main fumarole area and by two heat-flux monitoring stations in steam-heated soil (SHS) zones. All of these stations are located at the top of La Fossa cone (Fig. 2). The technical evolution and the availability of new types of instrumentation, such as infrared (IR) thermoscanning of the ground surface (Fig. 14; Schopa et al., 2011), have been utilized for extensive temperature mapping, and these new applications augment the information acquired by the long-term monitoring system (Diliberto, 2011). A long-term record of the chemical and isotopic compositions of released fluids can be included with the temperature record since the same fumaroles have been systematically and continuously

ACCEPTED MANUSCRIPT sampled. The main results from the geochemical monitoring are presented in Section 4.2.1 along with numerous references. The site locations of the continuous monitoring network installed at the top of La Fossa cone are shown in Figure 15. From a statistical point of view, we can consider the temperature data supplied by this continuous monitoring system to be robust, of high accuracy, and still representative of the deep processes. However, external perturbations are unavoidable due to the shallow positions of the sensors. The actual technical setup evolved as

PT

solutions were found to various empirical problems encountered during the monitoring period, including the analysis of data quality. The frequency of temperature measurements during the

RI

continuous monitoring has also changed over time. Data are now transmitted via radio to the Sezione di Palermo, Istituto Nazionale di Geofisica e Vulcanologia (INGV-Palermo) acquisition

center and they usually acquire one datum per hour.

SC

center once daily, and the monitoring stations can be programmed remotely by the acquisition

NU

Figure 13 shows all of the validated data that have been recorded in near real time by the monitoring network within the HTF field. No data interpolation has been applied, and each data

MA

point is the daily average of validated hourly values. Some gaps in the time series of temperature data are evident, which were due to temporary system failures. Most of the gaps in the data are

D

shorter than 10 days, with longer gaps—corresponding to the missing symbols in Figure 13—

PT E

occurring only on four occasions (starting in October 2002, August 2003, January 2009, and November 2013).

After many years from the first acquisition, a rescue process for the data (IDRA report; Diliberto

CE

et al., 2014) allowed the continuously recorded temperatures to be compared with the chemical and isotope data obtained by fumarole sampling performed over a 30-year period. These

AC

comparisons reveal that the thermal crises recorded at a few sites were associated with widespread geochemical variations of the fumarole fluids. Moreover, comparison with the results of other thermal surveys (i.e., Madonia et al., 2011; Schöpa et al., 2011; Harris et al., 2012) confirmed that the sites selected for continuous monitoring were still representative of the longterm thermal output. The local disturbances can be recognized more easily and filtered from the continuous monitoring data compared to the measurements made during a discontinuous survey (Diliberto, 2013). Therefore, the continuous monitoring allowed us to acquire high-frequency data of the fumarole temperatures over 25 years, also with a favorable cost-to-benefit ratio, because the amount of time needed to stay in this dangerous area of the active cone was considerably reduced compared to performing periodic surveys.

ACCEPTED MANUSCRIPT The yearly averages of fumarole temperatures shown in Figure 13b reflect the main changes in heat release that occurred in the high-temperature vents since 1991. The error bars in Figure 13b constitute interesting information about the range of values recorded during each year of continuous monitoring. For example, the largest error bars are indicative of increased variance caused by certain mid-term events of heat release (minor cycles in 1996, 1997, 2004, 2005, 2007, and 2011). These mid-term events have been filtered out in the yearly averages, and we were able

PT

to parameterize them (by duration and intensity) only due to the constant time window ensured by the continuous monitoring.

RI

The set of data supplied by continuous monitoring could be compared to other time-series data sampled at a lower frequency. The temperature variation recorded during 1996 (Fig. 13a) at the

SC

HTF field is the most intense and evident variation in the released heat, and it occurred while an increased hydrothermal circulation affected also the base area of the cone. The enhanced thermal

NU

release identified during 1996 in the summit area of La Fossa cone has been temporally associated with a sequence of anomalous peaks in the peripheral diffuse degassing (Fig. 16), that is clearly

MA

distinguishable from the local background values of the CO2 flux (about 50 g m–2 day–1). However, the pulsing behavior pattern of diffuse gas emissions and the discontinuous measurement window

D

hinder the correlation with the continuous-monitoring data of fumaroles. The minor thermal crises

PT E

identified by continuous monitoring (e.g., 1998, 1999, 2004, and 2005) do deserve some consideration too. Their cyclic character and the increased heat release involved an area of the active cone that was larger than the ones identifiable around the main fumarole emissions. The

CE

thermal areas at the margins of the main fumaroles or located elsewhere are SHS zones, and they were identified and described for the first time at the top of La Fossa cone by a survey of

AC

temperature measurements at a depth of 0.3 m (Aubert & Alparone, 2000). Aubert et al. (2008) subsequently reported the results from a new system monitoring the heat flux variations reaching the ground surface during a first testing period (the site location is indicated as “SHST” in Fig. 15a). The outlines for the SHS monitoring method are given in the appendix (Section 8.9), while the practical method and the theoretical background of this new measurement tool were introduced by Aubert (1999) after a test carried out at Mt Etna volcano. The success of this practical monitoring approach is crucially dependent on the site location (Diliberto et al., 2017). The heat flux is best monitored where the condensation zone is just beneath the monitored profile, and so these particular areas are called SHS zones; a schematic section of the ideal relationships between temperature in the ground profile and depth is shown in Figure 17. The SHS zones are not easily

ACCEPTED MANUSCRIPT identified from discontinuous temperature measurements. These zones generally are the heat dispersion areas located where the ascending steam fraction condenses just below the ground level (i.e., above Z1 in Fig. 17); here the temperature of the soil surface is slightly higher than the normal ambient temperature. At the Z1 condensation level the soil temperature is enhanced more by the massive steam condensation than by temperature exchanges between the ascending fluids and the solid structure. Within the uppermost soil layer (i.e., above the Z1 level) the pores are

PT

filled by a residual monophasic mixture of gases (components from atmospheric, hydrothermal, and magmatic endmembers), because the steam fraction in the emitted volatiles is already

RI

completely condensed below this level. Just below the temperature profile considered for the heat flow evaluations, the liquid phase fills the pores and then percolates downward, following

SC

the permeability gradient. This means that diffusive heat transfer dominates the upper layers, with negligible contribution from convective heat transport, and explains why the temperature

NU

variations recorded in the monitored profile reflect any increase in ascending steam condensation that would occur just below it.

MA

In conclusion, when some necessary conditions are fulfilled at a monitoring site, a simplified heat-flow equation can be applied up to the ground level that takes into account the diffuse heat

D

transfer component, as proposed by Aubert (1999) and reported in the appendix 8.10.3.

PT E

Some periods of increased thermal releases that were strictly related to the new episodes of increased solfataric activity (Granieri et al., 2006; Paonita et al., 2013) were identified by the temperature monitoring network and could be followed in near real time, both in the historical

CE

HTF sites and in the new SHS zone, by the new monitoring system (Diliberto, 2011, 2013; Cannata et al., 2012). The first episode is discussed in Section 4.2.1, and the increase in thermal output in the SHS monitored site is shown in Figure 18a,b. The heat flux evaluated at the SHS site (Fig. 20a)

AC

showed a mean value of 47 W m–2 up to November 12, 2004, and 1 day later it increased to the maximum value. When the condensation level (Z1 in Fig. 17b) moved upward, the conductive layer disappeared (CdHT, Fig. 17b) and the heat transfer became essentially advective. During the test period the SHS station underestimated the hydrothermal heat flux, but the strong correlation with the temperature of HTFs suggests that the SHS monitoring data can be used for interpolating other parameters that have been collected discontinuously. The clearest difference between the HTF and SHS temperature records is that after each steep increase, the ground temperatures of SHS suddenly returned to their range of background values. Thus, the SHS ground located out of

ACCEPTED MANUSCRIPT the northern flank of La Fossa cone actually appears to be the best location for dating the end of each pulse of the hydrothermal steam output. More recently, the heat flux monitored in the summit area showed a stationary trend (Fig. 19), with changes related only to external perturbations such as rain, wind, and air temperature. 4.3.1 Monitoring the Soil CO2 Fluxes CO2 is the main constituent of dry magmatic volatiles, and it is dissolved in the magma and is

PT

discharged from the volcanic edifices through summit areas of volcanoes (via plumes or crater fumaroles) as well as in peripheral areas (from soils and aquifers) (Chiodini et al., 2005;

RI

Inguaggiato et al., 2012a). Evaluating the soil CO2 fluxes gives information about the volatile

SC

output linked to a degassing magma batch (Chiodini et al., 1996; Favara et al., 2001; Cardellini et al., 2003; Chiodini et al., 2005; Inguaggiato et al., 2005; Pecoraino et al., 2005; Mazot et al., 2011).

NU

The study of the CO2 output is relevant both for geochemical monitoring to follow the evolution of volcanic activity (Brusca et al., 2004; Carapezza et al., 2004; Werner & Cardellini, 2006; Inguaggiato

MA

et al., 2011a,b,c; 2012b; Pering et al., 2014) and to estimate the volatile output to evaluate the magmatic-CO2 contribution to the global C cycle (Brantley & Koepenick, 1995; Arthur, 2000;

D

Mörner & Etiope, 2002).

Vulcano Island is one of the most-studied volcanoes worldwide. The first estimations of the

PT E

diffuse soil degassing in the peripheral areas of the active cone began in 1984 (Badalamenti et al., 1988, 1991; Baubron et al., 1990). These first output estimate, made for the period of 1984 to

CE

1988, showed a wide range of values from 70 to 1,000 t day –1 calculated for an area of 2.2 km2, and were based on investigations of 55 locations where the soil CO2 flux was measured. Furthermore, CO2 degassing measurements carried out in 1990 and 1996 by Baubron et al. (1991)

AC

and Chiodini et al. (1996) estimated the CO2 emitted from the summit area of the active cone at 150 and 200 t day–1, respectively. An output of 75 t day–1 in 1993 was estimated by Chiodini et al. (1996, 1998) in the same summit area using a finer measurement grid of the soil CO 2 flux (420 locations). Moreover, Diliberto et al. (2002) presented 10 years of temporal variations of diffuse degassing measured at the base of the volcanic edifice (Palizzi area), which also highlighted the temporal relationships between changes in CO2 diffuse flux and volcanic activity. Finally, the total CO2 output from the whole of Vulcano Island was 482 t day–1 (Fig. 20) in 2007. The total balance was derived by an integrated approach on the different sources of data: (i) soil

ACCEPTED MANUSCRIPT degassing, (ii) plume (better defined as volcanic cloud) of the solfataric area, (iii) dissolved CO2 in the aquifer, (iv) and bubbling gases at the Levante Bay and Gelso beaches (Inguaggiato et al., 2012a). That study demonstrated that the degassing of the peripheral area (30 t day –1) contributes less than 10% of the total outgassing. The CO2-degassing maps of the whole island included in Inguaggiato et al. (2012a) highlighted the anomalous degassing areas, and a new suitable site was identified for implementing geochemical systems for monitoring the volcanic activity during this

PT

solfataric phase. Consequently, an automated CO2 soil monitoring station was installed in September 2007 at the active summit crater of La Fossa cone outside the fumarole areas

RI

(Inguaggiato et al., 2012b). The new data on CO2 soil fluxes have been acquired on an hourly basis by utilizing an automated accumulation chamber system (manufactured by West Systems;

SC

Chiodini et al., 1998). These data are transmitted by a Wi-Fi connection to the Carapezza volcanological center and then, via the Internet, to the INGV-Palermo geochemical monitoring

NU

center. The high-frequency acquisition of data supports the geochemical monitoring of volcanic activity for the timely follow-up of volcanic emissions. The daily average of CO2 fluxes evaluated

MA

from 2007 to 2010 ranged from 800 to 16000 g m−2 day−1 (Fig. 21), suggesting a reference background range registered at the permanent station during the current state of normal

D

solfataric activity (1,600±250 g m−2 day−1, mean ± maximum deviation; Inguaggiato et al., 2012b).

PT E

The data obtained by monitoring the diffuse CO2 flux showed a continuous degassing process. Plotting the recorded data on a cumulated probability graph (Sinclair, 1974) resulted in a unimodal logarithmic distribution, with the CO2 fluxes centered around 1,600 ± 250 g m−2 day−1 (Inguaggiato

CE

et al., 2012b). In mid-September 2009 we observed a rapid increase in CO2 flux up to 16,000 g m−2 day−1, with hourly peaks of 20,000 g m−2 day−1 recorded during November and December 2009. At the beginning of January 2010 the fluxes returned to the range of the normal solfataric activity

AC

values, with an average flux of about 1,600 g m−2 day−1. Variations in the summit CO2 soil degassing recorded at the VCS station occurred simultaneously with those in other geochemical parameters, such as the gas/steam ratio of fumaroles and plume SO2 fluxes, therefore confirming that the continuous monitoring of diffuse CO2 flux in the summit area supplies the geochemical monitoring network with a further robust data set that is useful for closely following even minor changes in the solfataric activity. Moreover, the unimodal distribution of CO2 degassing with a single degassing family suggests that the solfataric activity in the summit area of Vulcano during the monitoring period is related to volatile exsolution from a large biphasic hydrothermal system fed from the magma below La Fossa

ACCEPTED MANUSCRIPT cone (Inguaggiato et al., 2012b). This hydrothermal system acts to subdue short-term variations in soil CO2 flux because the volatiles exsolved from the magma undergo a buffering process. In fact, at VCS station we observe that the large variation in the summit CO2 flux occurred over a long time interval (about 2–3 months). In conclusion, continuous monitoring of CO2 fluxes at the summit area of an active volcano by means of the accumulation chamber method represents a useful tool for monitoring deeper

PT

volcanic processes and for highlighting changes in the levels of volcanic activity in real time. The continuous updating of these data has improved investigations of the volcanic system without

RI

requiring difficult fieldwork. Correctly selecting sites for the continuous monitoring of the gas hazard and heat flux could support the evaluation of a changing hazard based on data displayed in

4.3.2 Monitoring Plume SO2 Fluxes

NU

SC

near real time, which would improve the early warning procedures.

MA

SO2 is the main S species released in the HTF gases (Aiuppa et al., 2005a). In the framework of volcanic surveillance activities, the SO2 emission flux has been used to calculate the volume of

D

magma batches in volcanic systems (Bruno et al., 1999; Edmonds et al., 2003). Over the past few

PT E

years there have been many investigations of measurement methods, which have resulted in an increase in SO2 flux data evaluated from steam clouds and volcanic plumes (Bruno et al., 1999; Edmonds et al., 2001). Recent technological advancements in SO2 flux measurements have

CE

simplified plume monitoring (Galle et al., 2003). It is relatively simple to identify SO2 plumes from the normal ambient background, because SO2 is absent in the clean atmosphere. In comparison,

AC

telemetric measurements of CO2 are only possible over short distances, with a plume that occupies a large percentage of the analyzed path. Indeed, over a longer path, the contrast of CO2 measurements between the plume and the atmosphere is not well defined, and the column amount of CO2 in the background atmosphere can far outweigh that in the volcanic plume. Actually, Vulcano only shows a summit degassing fumarole field at around 400°C, with a volcanic cloud, unlike an open-conduit degassing volcano (showing a real plume). New remote-sensing techniques and different prototypes have been tested in recent years in such a natural laboratory that is easy to reach by foot. In particular, the chemical composition of the volcanic cloud was defined and compared with chemical data from direct sampling. At the same time the correlations between different chemical species can be detected, and related to the degassing activity. Mori et

ACCEPTED MANUSCRIPT al. (1995) applied a Fourier-transform infrared (FTIR) spectral radiometer to Vulcano’s plume, Aiuppa et al. (2004) compared the application of different methodologies (FTIR, filter packs, and direct sampling) to the plume of Vulcano Island, O’Dwyer et al. (2003) tested a ultraviolet (UV) spectroscopy prototype for real-time measurements of volcanic H2S and SO2 concentrations in air, and Aiuppa et al. (2005a) compared H2S fluxes at Vulcano, Mt Etna, and Stromboli volcanoes with the aim of estimating the total S budget of volcanic origin (H2S and SO2). Furthermore, Aiuppa et

PT

al. (2004, 2005a, 2005b, 2006) and McGonigle et al. (2008) determined H 2S/SO2, CO2/SO2, and the relative fluxes of SO2 using differential optical absorption spectroscopy (DOAS), IR, and

RI

electrochemical sensor devices. The most common method for measuring the emission rate of SO 2 from a volcano has been passive remote data acquisition. The correlation spectrometer (COSPEC)

SC

was originally developed in the 1970s to monitor SO2 and other gases from factory smokestacks, and it was the first “remote sensing” device used in active volcanic areas. DOAS (Platt, 1994; Platt

NU

& Stutz, 2008) is an alternative remote-sensing method used for quantifying the concentrations of different trace gases. It is based on the principles of absorption spectroscopy (Bouguer–Beer-

MA

Lambert law) by collecting UV spectra using scattered sunlight from the sky (Bobrowski et al., 2003; Edmonds et al., 2003; Tamburello et al., 2011; Vita et al., 2012). Moreover, simultaneous

D

quantitative determination of two-dimensional bromine monoxide (BrO) and sulphur dioxide (SO2)

PT E

distributions in volcanic gas plumes on the island Vulcano (autumn 2004) was carried out with an Imaging DOAS instrument. The SO2 fluxes of several fumaroles was estimated from twodimensional distributions of SO2. Additionally, the first two-dimensional distributions of BrO within

CE

a volcanic plume were successfully retrieved (Louban et al., 2009). Finally, the UV camera is an important new tool in the geochemical monitoring for the high time

AC

resolution of SO2 flux measurements. In fact, in terms of the ability this UV camera approach provided to be able to measure gas fluxes from individual fumaroles at la Fossa crater of Vulcano island

(Tamburello

et

al.,

2011).

4.3.3 SO2 Plume: The NOVAC Network A worldwide network of permanent scanning DOAS instruments was established in 2005 to expand the remote-sensing capabilities of volcanological observatories. The Network for Observation of Volcanic and Atmospheric Change (NOVAC; Galle et al., 2010) has been installed at 19 volcanoes around the world to measure volcanic SO2 emission fluxes, and this technology is

ACCEPTED MANUSCRIPT now used for real-time monitoring and risk assessment. In the framework of the NOVAC international project, two networks for SO2 plume monitoring were installed at the islands of Vulcano and Stromboli, which are characterized by solfataric and Strombolian activity, respectively. Vulcano Island is the only closed-conduit volcano in the worldwide NOVAC network. The active La Fossa cone exhibits solfataric activity, with a wide fumarole field in the summit area that

PT

produces a wide steam cloud, similar to a real volcanic plume. The telemetric measurements of SO2 fluxes in the plume at Vulcano were not performed systematically in the past, but the

RI

persistent volcanic cloud rising from the top of La Fossa cone appeared to be suitable for testing the resolution of telemetry and for comparing the results obtained from long-term geochemical

SC

monitoring based on direct sampling techniques. In March 2008 we installed the first UV scanning DOAS instrument in the Palizzi area (S.D. Palizzi in Fig. 22) within the framework of the NOVAC

NU

project, together with a weather station located on Lentia Hill (Meteo Station in Fig. 22) at the same elevation as the La Fossa cone (350 m a.s.l.). In order to further improve our understanding

MA

of the SO2 fluxes at Vulcano Island, in February 2015 we installed a second UV scanning DOAS instrument in the Porto di Levante area, on the northeast side of the island (N. S.D. Porto di

D

Levante in Fig. 22). Although Vulcano Island is characterized by a very low SO2 flux (10–20 t day–1),

PT E

since it is a closed-conduit system, the UV scanning DOAS monitoring stations allowed us to measure the SO2 fluxes continuously from the steam cloud produced by the main fumaroles in the summit area.

CE

Figure 23 shows temporal variations of the SO2 flux. We observed a relatively constant SO2 flux of around 12 t day–1 from August 2008 to August 2009. This was followed by a large increase in SO2 emissions in mid-September 2009, with fluxes reaching 100 t day–1 in November 2009. In

AC

December 2009 the values of SO2 flux returned to around 12 t day–1, which is proposed as the local background SO2 emission rate (Fig. 23). The temporal variations registered by the UV scanning DOAS monitoring station were strongly correlated with other long-term time series obtained within the framework of the geochemical surveillance activities. The available geochemical parameters that we considered and compared included both the relative S content and the outlet temperatures of the fumarolic release, as well as the diffuse CO 2 flux from the ground at the summit of the active cone but outside the fumarole area (Fig. 17 shows the site locations). It was particularly interesting to observe that the anomalous values of SO 2 fluxes increased by about one order of magnitude during the same period of the anomalous soil diffuse

ACCEPTED MANUSCRIPT CO2 degassing recorded at the VSCS monitoring station (Fig. 21) located at the summit of La Fossa cone (compare Figs. 24 and 26). The independent acquisition of different geochemical parameters such as CO2 flux from the soil and SO2 flux from the solfataric (volcanic) cloud gave us the same indications, and both revealed an increase in volatiles involved in the degassing processes from a common deep source. The time series in Figure 26 actually represents the first uninterrupted set of telemetric SO2 flux data acquired continuously from the top of a closed volcanic system during

PT

solfataric activity.

RI

5. Stromboli Island

SC

Stromboli volcano is the northernmost island of the Aeolian Arc (Fig. 1) and is characterized by open-system degassing and persistent mild explosive activity, which has been episodically

NU

interrupted during recent decades by effusive eruptions in 1985, 2002, 2007, and 2014. The persistent Strombolian activity reflects adelicate dynamic equilibrium (Fig. 24) between the

MA

continuous deep volatile recharge and a consequent Strombolian activity that maintains the pressure equilibrium between the input and output of volatiles in the shallow volcanic system

PT E

D

(Inguaggiato et al., 2011a).

When the increases in the frequency and energy of summit crater explosions are not able to relieve the overpressure of the plumbing system, new fractures open on the flanks of the volcanic

CE

edifice. The opening of new fractures and the ensuing lava flows restore the dynamic equilibrium of the Stromboli plumbing system, changing the volcanic activity from normal Strombolian to

AC

effusive (Inguaggiato et al., 2011a). Apart from being extensively released through the central conduit, a small proportion of the Stromboli volcanic gases are diffusively emitted through the soil and interact with the basal aquifer while following fault-controlled pathways (Inguaggiato et al., 2016). Widespread soil degassing areas, sometimes associated with thermal anomalies, are located in the summit area as well as on the flanks throughout the volcanic edifice (Inguaggiato et al., 2013).

5.1 Monitoring the Thermal Aquifer of Stromboli Island

ACCEPTED MANUSCRIPT The first studies of the shallow thermal aquifer of Stromboli were performed in the mid-1990s, and aimed at identifying the area most affected by magmatic degassing by utilizing chemical and isotopic markers in both water and dissolved gases (Carapezza & Federico, 2000). The northeast sector of the Island, where the village of Strombolilies, hosts a thermal saline aquifer (with a temperature of about 40°C) that is mainly y fed by seawater and CO 2-rich fluids, as suggested by their dissolved C contents (Fig. 25). The pristine composition of this aquifer, where meteoric

PT

waters mix with seawater, has been deeply modified by the leaching of rocks and by the dissolution of a CO2-rich gas phase (Grassa et al., 2008). Resistivity surveys have provided evidence

RI

of water-saturated high-porosity rocks in the area of Scari (Finizola et al., 2006), while selfpotential data have indicated an unconfined aquifer in the same area, characterized by a head

SC

gradient of 50 m/km (Revil et al., 2011). The isotopic compositions of both dissolved CO2 (δ13C) and He (3He/4He ratio) have provided insights into the magmatic source of these gases (Capasso et

NU

al., 2005). The systematic monitoring of a few thermal wells that started in 1999 with a monthly sampling frequency has yielded signals predictive of impending energetic explosive events. The

MA

2002-2003 eruption gave the opportunity to observe changes in the chemical and isotopic compositions of the dissolved gas that heralded the onset of the eruption by several months

D

(Carapezza et al., 2004, Federico et al., 2008). Those authors considered that the shallow aquifer,

PT E

located at the base of the cone, would not be sensitive to the dynamics of the shallow part of the feeding system (at altitudes above 500 m), where Strombolian explosions originate. During normal Strombolian activity, the basal aquifer undergoes a steady-state input of magmatic gas. In

CE

contrast, the basal aquifer is also affected by an enhancement in the ascent of volatiles from the deepest parts of the feeding system when a change in the eruptive activity is going to occur, as

AC

observed several months before the 2002-2003 eruption. In particular, five months before the onset of the eruption, an increase in dissolved CO2 content was recorded at the four monitored thermal wells. The magmatic isotopic signature of the dissolved CO2 accounted for an enhanced gas-driven magma supply rate and gas overpressure within the conduits (Federico et al., 2008). Capasso et al. (2005) developed a model for the chemical and isotopic fractionation of magmatic CO2 and He during their ascent toward the surface and partial condensation in hydrothermal aquifers. Both δ13C of CO2 and the He/CO2 ratios measured in thermal waters result from fractionation processes related to the different solubilities in water of He, CO2, and the 13C and 12C isotopes (Fig. 26). This process would modify the initial composition, increasingly when the gas flux from depth is smaller. Conversely, He isotopes are considered to be unaffected by fractionation in

ACCEPTED MANUSCRIPT water, whereas they undergo fractionation upon magma degassing (Inguaggiato & Rizzo, 2004b). During the recent eruptions, in particular those in 2002-2003, 2007, and 2014, the increases in 3

He/4He (R/Rac) measured in groundwater are consistent with a gas phase in equilibrium with a

deep primitive magma that did not experience any extensive degassing (Rizzo et al., 2015) (Fig. 27). R/Rac values of 4.20±0.15 are typical of a period of normal Strombolian activity, whereas values as high as 4.6 would characterize the pure low porphyritic magma (Martelli et al., 2014). A

PT

higher input of gas-charged primitive magma, indicated by higher CO2 contents and R/Rac values measured in dissolved gases, would coexist with both frequency and energy increases in

SC

5.2 Monitoring the C/S Ratio in the Plume

RI

Strombolian explosions at the summit crater.

Volcanic plume emissions were the target of repeated aircraft-based measurements performed

NU

from the 1980s to the next two decades (Allard et al., 1994, 1999, 2000). The measurement techniques included UV (COSPEC) remote sensing of the SO2 plume flux, onboard analysis of the

MA

in-plume gas concentrations of SO2, CO2, and H2O, and in-laboratory analysis of trace elements in the particulate matter. These aircraft-based measurements were very expensive, and were

D

gradually replaced by ground-based plume investigations involving filter packs (Allard et al., 2000;

PT E

Aiuppa et al., 2009), diffusive samplers (Aiuppa & Federico, 2004), OP (open path)-FTIR spectroscopy (Burton et al., 2007a), and multiple gas sensors (Aiuppa et al., 2009). The different methods applied for chemical investigations of plume gases have revealed that passive degassing

CE

is characterized by a broad chemical composition and either short-term oscillations or long-term variations (on the order of weeks or months). The plume emissions are predominantly composed

AC

of water vapor (Allard et al., 1999), averaging ≈80 mol% (Burton et al., 2007a; Aiuppa et al., 2010a,b), followed by CO2, SO2, and HCl in decreasing order of abundance. SO2 represents the main S species by far, with SO2/H2S molar ratios ranging from 14 to 17 (Aiuppa et al., 2005a). The SO2/HCl ratio ranges from 0.3 to 3.1, with an exceptionally high value of about 9 measured for the paroxysm that occurred on April 5, 2003 (Allard et al., 2008 and references therein). The CO2/SO2 ratio during passive degassing ranges widely, from 0.9 to 26 (Aiuppa et al., 2010a). A permanent monitoring system based on the Multi-GAS technique (Aiuppa et al., 2007) has been operating in the summit crater area of Stromboli since May 2006 and is regularly measuring the in-plume CO2 and SO2 concentrations. Since the first years of monitoring (which included the 2007 Stromboli eruption), it has been evident that the volcanic gases emitted immediately prior to

ACCEPTED MANUSCRIPT and during the effusive eruption were remarkably different in composition (i.e., richer in CO2) from those typical of ordinary Strombolian activity (Allard et al., 2008; Aiuppa et al., 2009) (Fig. 28). Remarkable short-period variations (on a timescale of seconds) in volcanic gas compositions at Stromboli were first documented by high-frequency FTIR measurements (Burton et al., 2007b). Those authors demonstrated that during the short-lived Strombolian explosions, the volcanic gas phase was richer in CO2 than was the bulk (quiescent) plume. Since CO2 is significantly less soluble

PT

in basaltic melts than are H2O, S, and Cl, and thus is deeply exsolved, it was concluded that the gas slugs feeding the explosions have a deeper provenance (0.8–2.7 km below the summit vents) than

RI

those feeding the ordinary (mild Strombolian) activity. The anomalously high CO 2/SO2 plume ratios measured before the onset of the 2007 eruption (daily average of CO 2/SO2≈24) have been

SC

interpreted therefore as evidence of an enhanced supply of deeply derived CO2-rich gas bubbles into the shallow-plumbing system, triggering the transition from mild Strombolian to anomalous

NU

(effusive to highly explosive) volcanic activity (Aiuppa et al., 2009). Furthermore, Aiuppa et al. (2010) discussed a limited data set of simultaneous measurements of

MA

CO2, SO2, and H2O recorded from July to December 2008 using the Multi-GAS system updated to measure H2O concentration, according to the method proposed by Shinohara et al. (2008). Those

D

authors showed that Stromboli’s synexplosive gas phase is richer in CO2 (11–50%; mean 26%) and

PT E

poorer in H2O (48–88%; mean, 73%) than the bulk plume that is passively released by the volcano’s open vents between the explosions (mean CO2 and H2O of 15% and 82%, respectively), thus confirming the previous insights of Burton et al. (2007). The mechanisms controlling such

CE

temporal changes in Stromboli’s gas emissions have been explored by comparing Multi-GAS measurements with the model results derived from an equilibrium saturation code and based on

AC

the melt inclusion record of the abundance of volatiles in magma (Aiuppa et al., 2010). Those authors concluded that the compositional features of Stromboli’s quiescent and synexplosive gas emissions result from the mixing of gases that are persistently sourced from different depths of the plumbing system. In particular, volatiles are contributed by (i) the degassing of the porphyritic magma filling the upper (<1 km) dyke-conduit system and (ii) CO2-rich gas bubbles originating at a depth of >4 km (pressure > 100 MPa) in the plumbing system (Fig. 29). The contributions of these two gas sources change over time, causing the striking variability in Stromboli’s gas emissions. These fluctuating contributions are related to their relative strengths, which may reflect changes in the magma-gas content or convection rate, or the fluxing of CO2-rich gas bubbles from greater depth, controlled by tectonics. Regardless of the cause of these variable contributions, the CO 2-

ACCEPTED MANUSCRIPT enriched signatures of surface emissions reveal the increasing supply of gas bubbles ascending from depth, and this can potentially herald large-scale deeply sourced paroxysms.

5.3 Monitoring the Plume SO2 Fluxes The chemical composition of volcanic gas has been used to monitor the status of volcanic activity and to evaluate the degassing processes ongoing at depth below volcanoes. Chemical analyses of

PT

sampled fumarole gases yield the composition of these gases, and this allows evaluations of the conditions of the underlying magmatic systems. The direct sampling of fumarole gases in the

RI

summit areas, particularly in volcanoes exhibiting explosive activities such as Stromboli Island, is dangerous for volcanologists. This situation has led to the realization in recent years that is

SC

necessary to strengthen summit gas measurements using remote-sensing techniques. Measurements of the gas flux from active volcanoes are of great importance for monitoring

NU

volcanic activity.

SO2 is generally the main S species in a high-temperature volcanic gas plume, and its emission

MA

rate is considered a reference parameter for monitoring the degassing and eruptive activities. The first studies to estimate the amount of S species emitted during Strombolian activity were carried

D

out at Stromboli Island by Millan and Hoff (1978), Stoiber et al. (1983), and Allard et al. (1994).

PT E

They measured SO2 plume fluxes using a UV COSPEC. The SO2 plume fluxes were measured intermittently using a COSPEC during 1980–2004 (Allard et al., 1994, and references therein; Allard et al., 1999, 2000; Ripepe et al., 2005; Burton et al., 2008), after which the DOAS method was

CE

applied (Edner et al., 1994; McGonigle et al., 2003). The monitoring frequency increased during the 2002-2003 effusive eruption, and the reference

AC

value of SO2 flux during periods of normal Strombolian activity has been fixed at 250±50 t day–1 (Burton et al., 2007b). After the effusive eruption during 2002-2003, a network of scanning UV spectrometers was installed at Stromboli to measure the SO2 flux in the plume of the summit craters. The FLux Automatic MEasurement (FLAME) network (Burton et al., 2004) comprises four scanners located on the flanks of the Stromboli edifice (see Fig. 25) that since 2004 have been connected via a Wi-Fi network to the main volcanological observatory of Stromboli (COA) (Burton et al., 2009). The continuous measurements of SO2 fluxes were integrated with new campaigns of mobile SO2 flux measurements that started on June 27, 2006 using a Mini-DOAS device using the sky radiation as the light source (passive DOAS). Monthly measurements were carried out by performing several

ACCEPTED MANUSCRIPT traverses within 1 day (i.e., usually about five) in a boat beneath the plume, making measurements with a Mini-DOAS instrument (USB2000). The plume speed was assumed to be equal to the wind speed as measured by a weather station (STR02) located in the summit area at Pizzo Sopra la Fossa, at an altitude of about 950 m a.s.l. The SO2 flux was 190±50 t day–1 (Inguaggiato et al., 2011a) during 2006-2007, with the exception of the effusive period from February 27 to April 4, 2007. On January 19, 2007, the first increase in

PT

SO2 flux (up to 730 t day–1) was noted, after which the value returned to the background (on January 22). On February 28, 2007 the flux showed a further increase, reaching 1,280 t day–1,

RI

which was 1 day after the effusive activity began (Fig. 30). During the effusive period the flux showed values (600±200 t day–1) that are generally higher than the background range. A sudden

SC

increase up to 2900 t day–1 was observed on March 14, just 1 day before the paroxysm on March 15, 2007. During this last paroxysm, pyroclastic material was launched from craters down

NU

to low altitudes of the island, burning part of the Mediterranean vegetation and destroying volcanic monitoring equipment and the helicopter landing area. The SO 2 flux decreased after this

MA

event, but it remained around 500 t day–1 until after the end of the effusive eruption on April 2, 2007, when it decreased to the values typical of normal Strombolian activity (around 200 t day–1) (Inguaggiato et al., 2011a; Fig. 33). A further decrease in SO2 flux was recorded in September 2007,

PT E

D

down to 70 t day–1.

On the basis of these promising results, and thanks to the NOVAC project we were able to improve the plume SO2 network, installing two UV-scanning DOAS prototypes (Mark-II) (Kern,

CE

2009) in March 2007 in collaboration with Heidelberg University. The two instruments were installed respectively in the northeast (Saibbo) and southern (Punta Lena) sides of the island

AC

(Fig. 28). The acquired SO2 flux data confirm the good performances of the UV-scanning DOAS equipment. The continuous SO2 data for 2009-2010 are reported in Fig. 303 (blue symbols) together with the discontinuous data (red symbols) acquired by mobile DOAS during 2006–2008 (Inguaggiato et al., 2011a). The SO2 flux recorded continuously by the new NOVAC network, ranging from 80 to 480 t day–1 during 2009-2010, shows coherence with the variations observed in the other geochemical parameters (soil CO2 flux) that were also acquired continuously. More recently, Tamburello et al. (2012) carried out further investigations of magmatic active degassing processes, such as Strombolian explosions and puffing, using UV SO2 camera equipment. Coupling geophysical to geochemical data on a volcano at 1 Hz, the first measurements of explosive gas masses from Stromboli, was carried out by McGonigle at al., 2009, Mori and Burton,

ACCEPTED MANUSCRIPT 2009 and Delle Donne et al., 2016. The SO2 plume flux is now monitored continuously by the FLAME network using UV DOAS and managed by INGV, and the acquired data are communicated in real time to the Italian Civil Protection Department (Burton et al. 2009). Moreover, on the basis of the good performance of UV camera an additional network was installed and integrated in the geochemical continuous monitoring system at Stromboli.

PT

5.4 Monitoring Soil CO2 Fluxes The studies of soil diffuse CO2 emissions have highlighted two high-permeability zones located at

RI

the summit (Pizzo Sopra La Fossa) and in the peripheral (Scari) areas of Stromboli volcano (Fig. 25; Carapezza & Federico, 2000; Brusca et al., 2004; Inguaggiato et al., 2013). Furthermore,

SC

Inguaggiato et al. (2013) estimated that the total CO2 output was 394 t day–1, taking into account the summit and peripheral areas and including both plume fluxes and soil degassing over the

NU

whole island (Fig. 31a). The main contributor to degassing processes was the fluids discharged from the summit area (377 t day–1): 350 t day–1 from the plume and 27 t day–1 from diffuse

MA

degassing (Fig. 31a). Scari and the lower flanks mainly accounted for the diffuse CO 2 degassing in peripheral areas (17 t day–1). The summit area therefore supplies more than 90% of the total CO 2

D

output from the island, suggesting that monitoring the fluids discharged from the summit area is the best way for detecting changes in volcanic activity.

PT E

The continuous monitoring of CO2 fluxes (Fig. 25) from the soil at Pizzo Sopra la Fossa (station STR02) and from the Scari area (station STR01) has been performed on an hourly basis since 2000

CE

(Brusca et al., 2004; Carapezza et al., 2004; Madonia et al., 2008; Inguaggiato et al., 2011, 2017) by means of an accumulation chamber (West Systems; Chiodini et al., 1998). Data are transmitted to the COA Civil Protection volcano observatory at Stromboli via Wi-Fi (stations STR02 and STR01),

AC

from where it is sent to the INGV-Palermo geochemical monitoring center via a virtual private network. The soil summit degassing recorded at station STR02 during the entire period (2000– 2015) showed an almost unimodal distribution of log(CO2) values, even if different mean values of CO2 fluxes were recorded during each subperiod (Inguaggiato et al., 2017) of the different effusive eruptions (Fig. 31b,c). The daily average values (24 measurements/day) of CO2 fluxes recorded in the summit area (at station STR02) during 2000–2015 are plotted in Figure 31d. The CO2 fluxes showed a wide range of values, from 2,000 to 85,000 g m–2 day–1, with an average of around 8,500 g m–2 day–1 for the period. Moreover, the three effusive eruptions periods of 2000-2003, 2007, and 2014 were

ACCEPTED MANUSCRIPT characterized (Fig. 31d) by large increases in soil CO2 degassing, indicating volatile overpressures and disequilibrium states in the volcanic system (Inguaggiato et al., 2016). The histogram of all of the CO2 flux data shows a distribution that is almost unimodal, with log(CO2) values of around 3.9 (≈8,500 g m–2 day–1), but a consistent (≈30% of total values) rightside residual tail showing values above 10,000 g m–2 day–1 (Fig. 31c,d) . The summit soil degassing recorded at station STR02 during the past 16 years highlights the

PT

importance of this parameter for monitoring the activity of Stromboli volcano. In particular, the changes from Strombolian to effusive activity were preceded by significant changes of up to one

RI

order of magnitude in the diffuse gas emissions.

For better understanding the soil CO2 degassing from the summit area, Inguaggiato et al. (2017)

SC

calculated the monthly average for the STR02 data set over the past 16 years of observations. The monthly averaging process filtered out rapid variations (spikes) of the CO 2 fluxes and revealed the

NU

main degassing trends. Figure 34d shows that the first effusive period (2002-2003) is characterized by the highest CO2 fluxes, of over 20,000 g m2 day–1, decreasing to around 7,000 g m2 day–1 after

MA

the effusive eruption (in July 2003). The highest values of CO2 fluxes can be attributed to large amounts of volatiles arriving inside the plumbing system of the volcanic edifice and to the

D

continuous refilling of magma batches after the last eruption that occurred in 1985. For example,

PT E

the soil CO2 fluxes recorded at station STR02 showed a long-term increasing trend from July 2005 to 2015 (arrow in Fig. 31d). The positive trend in the CO2 flux reflects the continuous refilling of a CO2-rich magma batch that restored the pre-2002 eruption conditions of the system,

CE

characterized by a high CO2 partial pressure. The effusive eruption that occurred in August 2014 confirmed the interpretation of the

AC

increasing trend of CO2 flux, with most of the values exceeding 10,000 g m2 day–1 being recorded during the last period of observation (from 2011 to 2015) (Rizzo et al., 2015; Inguaggiato et al., 2017).

The review of the long-term data set of CO2 flux from soils acquired using the continuous monitoring network in the anomalous degassing area of Stromboli volcano has supplied further clues for the type of geochemical surveillance required to evaluate changes of the volcanic activity 6. Panarea Island The island of Panarea is another volcanic edifice of the subduction-related Aeolian arc in the Southern Tyrrhenian Sea, Italy. With an area of 56 km2, it is the summit of a seamount rising from the seafloor, 2000 m below the sea level (Gabbianelli et al., 1986 and 1990; Favalli et al., 2005).

ACCEPTED MANUSCRIPT Most of the outcropping rocks show an acid calcalkaline composition and less abundant products with high-K calcalkaline and shoshonitic compositions (Calanchi et al., 2002). About 3 km East of Panarea Island, a group of islets and reefs (Dattilo, Panarelli, Lisca Bianca, Bottaro, Lisca Nera and Le Formiche; Fig. 32) forms the subaerial remnant of an old volcanic centre, surrounding a submerged area, located at a relatively shallow depth (<35 m b.s.l.). The gas and thermal water discharges from the seafloor have been observed since historical times (Italiano and Nuccio, 1991

PT

and references therein; see Fig. 33) and old reports talk about a “bollitore,” namely boiling water, at the sea surface. Undersea fluid releases have been the only observable clues of an active

RI

volcanism, since less than 10,000 yrs before present, at the time of the most recent eruptive

SC

activity, forming the dome of Basiluzzo (Gabbianelli et al., 1990).

NU

6.1 Submarine fluids emissions: chemical and isotopic characterization The chemical and isotope composition of the thermal fluids discharged from the submarine fumarolic field of this area were investigated in the early 1990s by Italiano and Nuccio (1991) and

MA

Calanchi et al. (1995). New studies were carried out after a submarine gas burst event, occurred in the exhalative area on the 3rd of November 2002 (Caliro et al., 2004; Capaccioni et al. 2005 and

D

2007; Caracausi et al. 2005a,b; Chiodini et al. 2006; Tassi et al. 2009 and 2014). The explosive

PT E

output of hydrothermal fluids in November 2002, formed a depression 20 by 8 meters wide and 7 meters deep near the islet of Bottaro (Fig. 32; Caliro et al., 2004). This event offered the chance to develop and test new and more refined sampling methods, as well as new conceptual models of

CE

the system.

The peculiar conditions of the fluid discharges at the sea bottom require us to collect samples by

AC

scuba diving operations (Italiano and Nuccio, 1991) that have been summarized in the appendix 8.8 Submarine fluids sampling techniques. The thermal waters show temperatures ranging between 30 and 130°C, pH values between 3.0 and 5.9, and NaCl contents close to that of seawater (Caracausi et al., 2005a, Tassi et al., 2009). Enrichments in Ca, K, and SiO2 with respect to sea water, as well as depletion in Mg, suggest that intense water-rock interaction probably takes place within the geothermal system. Mixing processes among seawater, seawater concentrated by boiling (slightly enriched in Mg), and a deep, highly-saline end-member (strongly depleted in Mg) suffering extensive water-rock interactions at relatively high temperature, control the chemical composition of the waters (Tassi

ACCEPTED MANUSCRIPT et al., 2009). Na/K, Ca/Na and Ca/K geo-thermometers give equilibrium temperatures with rocks, ranging between 190 and 250°C and water contaminations varying between 75 and 90% (Caracausi et al., 2005a). The gas emissions are largely dominated by CO2, with several percents of H2S and significant concentrations of CH4 and air-derived species, being thus typical hydrothermal fluids (Chiodini and Marini, 1998). The inverse relations between the H2S contents with respect to He and CO2 (Fig.

PT

34), coupled to the usual absence of acidic species (HCl, SO2), suggest that steam condensation and chemical scrubbing affect the composition at the vents (Caracausi et al., 2005a; Chiodini et al.,

RI

2006). Gas geothermometry in the system CO-CO2-CH4 with assumed liquid-gas coexistence, allow

SC

to estimate temperatures of 220 to 280°C for the local geothermal system (Italiano and Nuccio, 1991). The 3He/4He isotope ratios range from 4.2 to 4.6 Ra and δ13CCO2 values result around -2‰

NU

and -3.2‰ (vs. PDB) (Caracausi et al., 2005a, Capaccioni et al., 2007; Heinicke et al., 2009). All of these data, being in agreement with the values expected for the local mantle (Capasso et al., 1997,

MA

Paonita et al., 2002, Martelli et al., 2008 and 2014; Mandarano et al., 2016), address to the clear presence of a magmatic component in the discharged gases.

D

In November 2002, the gas burst dramatically changed the geochemistry of the fumarolic field, even in distal vents from the Bottaro crater. Besides the opening of several new vents and the

PT E

dramatic increases of the gas output from sea bottom (see the following Section), He isotope ratios increased up to 4.4 Ra or higher (Fig. 358; Capaccioni et al., 2007; Heinicke et al., 2009),

CE

standing higher than the values of 4.2 Ra of 1990s (Sano et al., 1989; Italiano and Nuccio, 1991), and both CO/CH4 and H2/CH4 ratios greatly increased too (Fig. 36; Caliro et al., 2004; Chiodini et al., 2006). A massive input of magmatic gases impacted the geothermal system, strongly affecting

AC

the compositional and thermo-baric features (Caliro et al., 2004; Capaccioni et al., 2005; Caracausi et al., 2005a; Chiodini et al., 2006; Capaccioni et al., 2007). To assess the conditions of the hydrothermal system, Caracausi et al. (2005a) proposed a novel geothermobarometric approach based on H2, CO and CH4 equilibrium, with no assumption of redox buffer and liquid-vapor coexistence (Fig. 370). The presence of a two-phase boiling system was assessed rather than assumed, by comparing the gas composition from 1990s to those collected in 2002 and 2003. The authors proposed a new geochemical conceptual model, in which the main fumarolic fluid source is a deep reservoir consisting of a aqueous solution having a high salinity at equilibrium with vapor (∼12 mol% CO2) at temperature and pressure up to 350 °C and

ACCEPTED MANUSCRIPT 16 MPa, respectively. Later on, Taran (2005) calculated 6-20 mol% of CO2 at temperatures of 300350 °C for Panarea fluids, in agreement with previous estimations. According to Caracausi et al. (2005a), the gas composition quenched at these equilibrium conditions up to the surface only in those vents having a high output rate, whereas gases emitted from small and low-flux discharges were affected by re-equilibrium down to temperatures of 150-200 °C. As displayed by Chiodini et al. (2006), the H2/CH4 and CO/CH4 ratios from gases collected just after

PT

the November 2002 event, increased dramatically towards typical values of high-temperature volcanic fumaroles (Fig. 36). Chiodini et al. (2006) suggested two alternative hypotheses: 1) a

RI

pressurization of the hydrothermal reservoir, up to 100 bars (at 300 °C) at atypical oxidizing

SC

conditions due to the magmatic gas input, occurred during the 2002 event, feeding the gas discharges; 2) the sudden release of magmatic gases flashed the existing hydrothermal system at

NU

temperatures up to 450 °C, during the 2002 event. A subsequent study by Capaccioni et al. (2007) interpreted the 2002 event as a convective heat pulse from a deep-seated magmatic body capable

MA

of creating a dry transient zone, where magmatic gases reached the sea bottom. According to Capaccioni et al. (2007), the detection of SO2 and HCl in the gases sampled a few weeks after the event, would be a strong clue for this interpretation. However these species were never detected

D

by other groups who sampled the same vents in the same period (Caliro et al., 2004; Caracausi et

PT E

al., 2005a; Chiodini et al., 2006), so the last interpretation by Capaccioni et al. (2007), remains as an open question.

CE

A remarkable control of the gas geochemistry by the regional tectonic assessment has also been envisaged by Tassi et al. (2014). Highly permeable zones at the intersection of the two main fault systems (NE- and NW-oriented), approximately in the center of the submarine fumarolic field,

AC

allow uprising of fluids from the hydrothermal-magmatic system. Secondary processes have therefore a limited influence on fumarole geochemistry of vents located in these area, so that the chemical composition (H2- and CO-rich) of the deep hydrothermal reservoir is almost preserved up to the surface. At the periphery of this strongly fractured zone, low-temperature gas-water interaction causes that gas discharges are almost completely depleted of the typical hydrothermal species, e.g. H2S, H2, CO and hydrocarbons. According to Tassi et al. (2014), the H2/CO ratios of the hydrothermal fluids have significantly decreased since 2012, possibly suggesting an increase of gas pressure at depth due to gas and energy contribution from the magmatic source to the hydrothermal reservoir. The authors envisage a process similar to that causing the 2002 gas burst.

ACCEPTED MANUSCRIPT However, for gases collected before the 2002 event, Caracausi et al. (2005a) put into evidence the lack of complete equilibrium in the CO2-CO-H2 system, which would strongly affect the H2/CO ratios (Fig. 37). In this view, the decreasing H2/CO ratio could have an opposite meaning, because they would suggest very shallow and low-temperature processes. Thus, any inference on the deep conditions of the hydrothermal reservoir, extrapolated by discontinuous sampling at different

PT

levels of activity (background, anomalous, recovering quiet activity) requires great care.

6.2 Submarine fluids output estimation

RI

Measurement of gas output from an underwater fumarolic field is extremely challenging, because

SC

remote sensing techniques (e.g. FTIR, DOAS) do not work in submarine conditions. Direct methods, which we will detail in the following, involve scuba diving operations or ROV apparatus,

NU

are costly and time-consuming so they can be performed just occasionally. Recently, measurements of dissolved CO2 gradients (by specific probes) around fumarolic fields have

MA

allowed us to estimate volatile output offshore the Campi Flegrei (Di Napoli et al., 2016), although this method would require the proper quantification of intensity and direction of the seawater flow field. At Panarea, changes of gas flow rate have also been continuously detected for several

D

months by a new device monitoring the acoustic emission produced from the gas bubbles of

PT E

submarine vents (Acoustic Bubble Counting, ABCO; Heinicke et al., 2009). However, it is still difficult to perform absolute measurements in this field. Panarea can be considered an interesting case study, concerning the importance of gas output

CE

acquisitions for understanding how an underwater volcanic system works. Flux measurements were carried out in 1985-1987 following a direct method (Italiano and Nuccio, 1991). They were

AC

performed by means of a stainless-steel funnel and an overturned bottle of known volume. To take the output measure, the funnel was put on top of the emission point and the time it took the carried gas to move the water out of the bottle was measured. The volume of the emitted gases was corrected for hydrostatic pressure. An almost constant degassing rate, of the order of 9x106 l day-1 of CO2 (Fig. 35) was found. Measurement campaigns were performed also in 1993, 1996 and 1999 and they confirmed a constant gas output (7–9 ×106 l day-1). The above-mentioned gas burst of November 2002, caused gas output increases of at least two orders of magnitude higher, with respect to the background values (Caliro et al., 2004). The subsequent measurements showed exponential decreases with time, taking the values back to the

ACCEPTED MANUSCRIPT pre-crisis output (Caracausi et al., 2005a). The peak values at Panarea show outgassing rates close to those measured in other active volcanoes (i.e., Vulcano, Phlegrean Fields), addressing to an increased input of fluids and heat from a magma body as being responsible for the crisis. Caracausi et al. (2005a) also suggested that this magma had to be poorly crystallized in order to explain the measured gas flux. An interesting observation was that the CO content in the emitted gases continued to increase for

PT

six months after the crisis, meaning that the geothermal temperature and pressure also increased (Fig. 36; Caliro et al., 2004). Caracausi et al. (2005a) noted that a constant input of deep fluids from

RI

a cooling magma into the geothermal system and a sudden release of gases due to changes in

SC

permeability is not compatible with the total gas output that remained higher after the onset of the crisis and the concurrent increase in T and P. In contrast, a sudden increase in the supply of

NU

deep magmatic gases, causing over-pressurization and opening of the geothermal system, would explain the above observations: an increased contribution of magmatic fluids would have

MA

triggered the P-T increases and high gas fluxes. According to Caracausi et al. (2005a), the increase of T and P for several months after the crisis, even when the gas output was already decreasing, meant that the geothermal system reached its highest capacity to outgas fluids at the onset of the

D

crisis, when the fractures opened, whereas a subsequent lowering of the permeability reduced the

PT E

possibility of the fluids to escape, causing the T-P increase. Subsequently, the decrease of gas output, T and P after July 2002 suggests the end of the anomalous input of magmatic volatiles in

AC

7. Discussion

CE

the hydrothermal system.

The geochemical investigations carried out in recent decades at active volcanoes highlight that an understanding of fluid geochemistry can yield timely interpretations for use in volcanic activity surveillance. Three main aspects have been developed and improved in the last 2 decades: (i) remote data acquisition, (ii) extensive parameter acquisition, and (iii) formulation of geochemical models. 7.1 Remote Data Acquisition Volcanic gases can be monitored via both direct and remote (ground- or satellite-based) measurements. Considering that measurements of volcanic gases can give useful information

ACCEPTED MANUSCRIPT about the dynamics and evolution of magmatic systems (Stoiber et al., 1983; Aiuppa et al., 2002; Allard et al., 2005; Saweyer et al., 2008; Vita et al., 2012), the increased data acquisition frequency achieved in recent decades has played a crucial role. Moreover, the improvements in technology and their availability have led to more reliable and affordable scientific and technological equipment. One of the most significant examples among the ground-based techniques is DOAS, which allows remote measurements of gas fluxes from volcanic plumes (Bobrowski et al., 2003;

PT

McGonigle et al., 2003; 2005; Burton et al., 2004, 2009; Ripepe et al., 2005; Kern, 2009a, b, c; Inguaggiato et al., 2011; Vita et al., 2012, 2014; Christopher et al., 2014).

RI

Moreover, the increasing number of satellites orbiting the Earth in recent decades has resulted in many satellite-based measurements (e.g., temperature, SO2 and CO2 concentrations, and

SC

effusion rate) now being used to follow the activities of large volcanoes that are difficult to reach for on-site sampling. Satellite measurements have also been made at the Aeolian volcanoes, such

NU

as of the effusion rates and thermal flows of the Stromboli eruptions in 2007 and 2014 (Principe et al., 2008; Zakšek et al., 2015). A process of comparison and validation of SO 2 flux data between

MA

ground- and satellite-based measurements has also started at Vulcano Island (Pinardi et al., 2011). In general, ground-based remote measurements make it possible to study and monitor in near-

D

real time huge active volcanoes that are difficult to climb and/or during their paroxysmal activity.

PT E

Meanwhile, satellite-based remote measurements enable to monitoractive volcanoes located in remote areas and to detect in near-real time the onset of eruptions even in volcanoes that are not being monitored directly.

CE

7.2 The Extensive Parameters

AC

The importance of extensive parameters has increased in recent years for evaluating the magma volumes involved before, during, and after eruptive activities and for modeling eruption mechanisms. Moreover, these parameters give strong indications about the kind of volcanic activity based on the degassing style. This section provides some examples taken from the Aeolian Islands. 7.2.1 Open versus closed systems The open conduit of Stromboli volcano and the closed conduit of La Fossa cone (at Vulcano) are characterized by different activities: Strombolian vs solfataric, respectively. The fluxes of volatiles emitted from the summit crater areas of these volcanoes have reflected the different styles of degassing. In particular, direct degassing of volatiles from the magma located in the shallow part

ACCEPTED MANUSCRIPT of the Stromboli edifice has been observed to always show large and rapid variations linked to changes in the deep input on the order of a few days. Conversely, the degassing of volatiles from Vulcano Island has been characterized by large but slower variations in volatile fluxes, linked to indirect magma degassing located at few kilometers below the volcano edifice. The large biphasic hydrothermal system interposed between the magma body and exposed surfaces of the closedconduit volcano partially buffers the heat and gas emissions, and modulates the solfataric activity.

PT

The different statistical properties of the soil CO2 fluxes measured at the monitoring stations reflect these different kinds of degassing: an unimodal distribution characterizes the CO 2 fluxes

RI

recorded at the summit of La Fossa cone (Vulcano Island; Inguaggiato et al., 2012a), while a polymodal distribution characterizes the CO2 fluxes recorded at the top of Stromboli Island

NU

SC

(Fig. 38a,b).

7.2.2 Peripheral Versus Summit Degassing

MA

Further strong support for the validity of extrapolating from estimations of volatile fluxes is the relative distribution of volatile emissions from peripheral and summit areas of a volcanic edifice.

D

The degassing surveys carried out at the islands of Vulcano and Stromboli showed that more than 90% of the CO2 is discharged from the summit area, indicating that the fluids exsolved from the

PT E

magma batch are focused and channeled along preferential pathways that reach the summit areas (Inguaggiato et al., 2012a, 2012b, 2013).

CE

These results can be compared to those from other volcanic systems where the CO2 peripheral and summit degassing has been estimated. Figure 39 confirms similar behavior for open-system degassing volcanoes such as Popocatepetl and Mt Etna: the CO2 summit degassing is one order of

AC

magnitude higher than the CO2 peripheral degassing. Conversely, closed-system degassing volcanoes that exhibit solfataric activity, such as at the islands of Ischia and Pantelleria, show the opposite behavior: the CO2 summit degassing is one order of magnitude lower than the CO2 peripheral degassing (Inguaggiato et al., 2013). Vulcano Island, showing higher CO 2 summit degassing, behaves like an open-conduit volcano even though it is a closed-system volcano exhibiting solfataric activity. An increasing fracturation of the summit area of the cone occurring during recent decades could account for both the short-term variations, namely the geochemical crisis observed after 1996, and the decrease in the emitted fluids detected at the same time in the base area. Thus changes in the ratio between the peripheral and summit gas outputs could

ACCEPTED MANUSCRIPT suggest changes in the type of degassing and aid the interpretation of further evolution of the volcanic activity. 7.2.3 Mass and Energy Outputs (Indirect Estimation of Steam Flux) Combining extensive parameters such as the fluxes of the main gas species with intensive parameters such as the temperature and chemical composition of fluids makes it possible to

PT

estimate the mass output of all of the volatiles emitted from a volcanic system and the associated energy release (Taran, Y. A., & Peiffer, L. (2009); Chiodini et al., 2001, 2014). In particular, a case

RI

study of Vulcano Island that performed a detailed investigation of the SO 2 and CO2 fluxes was

SC

carried out in 2007, and the results were interpreted in combination with the total chemical composition of fumarolic fluids (Inguaggiato et al., 2012a). The acquired data were used to

NU

estimate the mass outputs of the discharged CO2, H2O, N2, He, CO, CH4, Cl, and F. Moreover, the estimated steam fluxes in the (i) plume, (ii) summit, and (iii) peripheral areas were assumed to be a proxy for the total energy discharged by the whole volcano-hydrothermal system. The vapor

MA

released from the fumarole area of La Fossa cone and the associated energy release from Vulcano Island were computed. Taking into account only the latent heat of evaporation of water (2,500 kJ

D

kg–1) and the estimated water flux of about 1,300 t day–1, an energy release of about 3109 kJ day–1

PT E

(39 MW) was estimated for the crater plume (Inguaggiato et al., 2012a). Then, based on the CO 2 fluxes measured in the summit area (outside the fumarole area) and peripheral areas of Vulcano, steam fluxes of around 335 and 105 t day–1, respectively, were estimated, which correspond to

CE

energy releases of about 9.7 and 3 MW in the summit and peripheral areas.

AC

7.3 Geochemical Modeling Early approaches to the monitoring of active volcanoes were based on empirical observations of changes in compositional ratios such as Cl/S and F/Cl in fumarolic fluids, as well as the temperatures of thermal waters. General and robust geochemical models for understanding single observed changes were still missing. During the last 30 years, numerous investigations have characterized volcanic systems from a geochemical point of view. A volcanic system offers different and significant points of observation such as thermal aquifers, anomalous soil degassing, fumaroles, and plumes. On the basis of the detailed and specific studies carried out on these different manifestations in the Stromboli and Vulcano systems, researchers were able to model

ACCEPTED MANUSCRIPT the volcanic systems and to interpret the changes observed in any single parameter within the framework of a general geochemical model. Many geochemical models have been formulated in recent decades with the aim of increasing the understanding of the plumbing system and the relative volatile and energy outputs as discussed in this review article (e.g., Chiodini & Marini, 1998; Inguaggiato et al., 2000, 2011a, 2012a; Nuccio & Paonita, 2001; Paonita et al., 2002, 2013; Aiuppa et al., 2010a, 2010b; Federico et

PT

al., 2010; Diliberto, 2017). Moreover, data on contemporaneous variations in intensive and extensive parameters and their comparisons with physical variations such as seismic activity have

RI

confirmed the importance of the multiparameter approach to volcano monitoring, and has also suggested new perspectives for strengthening the geochemical monitoring of volcanic activity and

SC

for improving the constraints in the construction of accurate geochemical models. Such models are necessary to understand how volcanic systems work and for predicting, at least to some

NU

extent, possible spatial and temporal evolutions of these systems. 8. Conclusions

MA

Finally, on the basis of previous experience and considering the different geochemical models produced, a robust geochemical monitoring network has been designed and installed at the active

D

Aeolian volcanoes (Fig. 40). These geochemical networks make it possible to acquire a wealth of data that are useful for validating and continuously improving dynamic interpretative models.

PT E

Moreover, improved models may identify new observables to be explored, thereby representing a positive feedback mechanism.

CE

The current continuous-monitoring geochemical networks at Aeolian volcanoes currently comprise the following components:

AC

1. A plume geochemical network that measures SO2 fluxes and the C/S ratio. 2. A temperature monitoring network in the summit area. 3. A network for measuring soil CO2 degassing in the peripheral and summit areas. 4. Devices for monitoring dissolved CO2 in the thermal waters. This geochemical network monitors intensive and extensive parameters at a frequency that is sufficiently high for assessing the volcanic activity. Meanwhile, the periodic-monitoring geochemical network performs the following: 1. Periodic sampling (monthly) of the HTFs. 2. Periodic sampling (bimonthly) of thermal waters.

ACCEPTED MANUSCRIPT 3. Periodic surveying (half-yearly) the diffuse CO2 flux in selected zones of the summit and peripheral areas.

9. Appendix: Guide to Geochemical Methodologies

PT

Collecting fluid samples of a volcanic system is the first and one of the more important steps for both research and monitoring activities. Although a considerable amount of information can be

RI

obtained from data collected on natural systems, not all of the sampling sites can be considered representative of the main processes. Site-specific effects as well as the contamination of the

SC

samples may affect the interpretations, potentially resulting in inaccurate models. The levels of magmatic species and their molecular ratios can be determined from the direct

NU

sampling of fumarole gases or by utilizing telemetric observation methods. Collecting fluid samples is the first step in research work, and obtaining useful results often strongly depends on

MA

careful sampling. The main features of a good sampling can be summarized as follows: 1. Choice of sampling sites: Natural volcanic systems require careful preliminary

D

investigations when choosing the sites for collecting fluids that are most representative.

PT E

2. Choice of parameters to investigate: It is also important to choose the parameters to be analyzed that can provide fundamental scientific information for investigating the volcanic system. The chemical composition of gases can give useful information about

CE

the temperature, pressure, and O fugacity of deep systems, while the isotopic compositions of He, C, and water vapor can give useful information about the origin of

AC

these gases and also about the physicochemical processes affecting gases during their ascent toward the surface. 3. Atmospheric contamination: The collected sample must be free of air contamination and should contain only the volcanic fluids. Unfortunately, fumarolic fluids are often affected by mixing among magmatic, hydrothermal, meteoric, and atmospheric components. Moreover, this contamination sometimes occurs at depth, making it impossible to avoid. 4. Re-equilibration and quenching phenomena: During their ascent toward the surface, the temperature of volcanic fluids decreases and the ratio of molecular species can be affected by re-equilibration processes occurring at lower temperatures. Fortunately,

ACCEPTED MANUSCRIPT quenching phenomena often preserve the characteristics of higher temperature equilibria (Le Guern et al., 1982; Gerlach & Casadeval, 1986). 5. Degassing and fractionation processes: Other mechanisms that can change the pristine chemical and isotopic compositions of the ascending fluids are kinetic and equilibrium fractionation processes. For example, the role played by the presence of a shallow groundwater system (Capasso et al., 1997; Capasso & Inguaggiato, 1998; Inguaggiato et

PT

al., 2000, 2005) and the effect of magma degassing processes (Nuccio & Valenza, 1998; Inguaggiato & Rizzo, 2004b) on the chemical and isotopic compositions of the gases

RI

have been studied.

6. Chemical reactions occurring in the plume: The chemical composition of gases in the

SC

plume can be affected by condensation, adsorption, and oxidation processes, which change the original amount of single species and consequently their molecular ratios

NU

(Mori et al., 1995).

MA

9.1 Techniques for Sampling Subaerial Fumaroles

Fumarole gases are generally collected using the alkaline solution method proposed by

D

Giggenbach (1975). A Ti/steel or SiO2/quartz pipe is inserted into the soil inside the fumarole. A

PT E

dewared SiO2 or quartz tube is then inserted into this pipe in order to avoid the condensation of water vapor. This tube is directly connected to an evacuated and preweighed flask filled with 4 M NaOH. The fumarole gases bubble inside the alkaline solution and the water vapor condenses.

CE

Alkaline conditions favor the absorption of CO2 in the solution as CO32-, while HCl, HF, and S species dissolve as anions. In this way almost all of the fluids are adsorbed in the solution while

AC

the noncondensable gases such as H2, N2, CO, and CH4 and rare gases remain in the headspace, thereby being concentrated (up to 100-fold). The isotopic compositions of He and C are determined by also collecting other samples in two-way glass flasks. To determine the stableisotope composition of water vapor (δD and δ18O), a sample of acid condensate is collected using a condenser cooled with ethyl ether connected to the same sampling system (Capasso et al., 1997). The CO2 content of the alkaline solution was analyzed by potentiometric titration, and for Stot, HCl, and HF according to the method described by Sortino et al. (1991). The concentrations of He, H2, O2, N2, CO, and CH4 in the enriched gas in the headspace were analyzed by a gas chromatograph (Clarus 500, Perkin Elmer) equipped with a 3.5-m Carboxen 1000 column and double detector (hot-wire detector and flame ionization detector), with analytical errors lower

ACCEPTED MANUSCRIPT than

3%.

The H2O content was determined by weighing the bottle before and after sampling, taking into account the absorbed amounts of CO2 and acidic species, and finally the composition of the whole fumarolic

gas

was

calculated

accordingly

(Badalamenti

et

al.,

1991).

PT

9.2 Submarine Fluids Sampling Techniques Collecting free gases merely involves placing a stainless-steel funnel on the bubbling gas vent that

RI

is connected to two-way Pyrex bottles tapped with Thorion valves (Fig. 33). The entire sampling

SC

system, filled with seawater, allows the collection of the emitted gases by utilizing water displacement. The funnel can also be connected to a pre-evacuated glass flask that is partially

NU

filled with an alkaline solution (Giggenbach, 1975), in which acidic gases are absorbed and noncondensable species are enriched in the headspace. In this case, contamination by seawater

MA

has to be minimized, and so the silicone-tube connection between the funnel and the collecting glass tube is filled with Milli-Q water and isolated from seawater until the sampler has been connected to the funnel (Capaccioni et al., 2005). Questions related to contamination by seawater

D

also apply to the sampling of thermal waters emitted at the seafloor. The samples to be analyzed

PT E

in the laboratory are collected through a quartz tube inserted 50 cm into the emission orifice and connected to a three-way Pyrex valve (Caracausi et al., 2005b). After rinsing out the system with thermal water, an upside-down glass bottle is filled with the emitted gas before water is injected

CE

(when gas and water are simultaneously discharged; otherwise scuba air can be used). The water is first pumped into a syringe and then injected into the bottle, while the equivalent volume of gas

AC

goes out through a needle until the sampler is full of thermal water. The pH, temperature, and conductivity can nowadays be measured directly at the point of emission using underwater sensors; alternatively, pH and conductivity can be measured as soon as the water sample has been brought to the surface. Once gas and water samples have been collected, the chemical and isotopic characterization is performed in the same way as for conventional subaerial fumarolic and thermal-water samples.

9.3 Chemistry and Physicochemical Parameters of Waters

ACCEPTED MANUSCRIPT The sampling of the cold and thermal waters associated with a volcanic system represents an extremely complex and delicate phase that determines the results of all subsequent operations and therefore strongly affects the result of the total uncertainty analysis. It is therefore very important to ensure that the sample is representative of the whole natural system under investigation and that physical processes such as stratification do not influence the chemistry of the sampled water.

PT

The outlet temperature, electrical conductivity, and pH of the waters were measured at every spring site using a thermometer, conductimeter (ORION 250A+), and pH meter (ORION 250A+),

RI

respectively. Water samples were put into several polyethylene bottles to allow the major components and stable-isotope compositions (δ18O and δD) to be analyzed. The samples for

SC

cations were filtered and acidified with Suprapur(R) HNO3. Alkalinity was determined in situ by titration with 0.1 N HCl, whereas major elements were analyzed in the laboratory using a double-

NU

ion chromatograph (Dionex-Thermo ICS 1100) at an accuracy of ±2%. A column (Dionex CS-12A) with a conductivity suppressor (CSRS 300) was used for the cations (Li, Na, K, Mg, and Ca), while a

MA

column (Dionex AS14A-SC) with conductivity suppressor (ASRS 300) was used for the anions (F, Cl,

D

Br, and SO4).

PT E

9.4 Isotopic Composition of Waters: Analytical Techniques 9.4.1 Isotopic Composition of Oxygen

CE

The isotopic ratio of Oxygen was measured using a mass spectrometer (Thermo Delta V Plus) coupled to a GasBench II system that exploits the principle of the noncondensable gases remaining in the headspace.

AC

The GasBench II system has an universal on-line interface that allows automated measurements of the isotope ratios of small gas samples. The O in the water can be analyzed by adding a certain amount of a mixture of He gas and CO2. This is left to equilibrate in the system for about 18 hours, and then the O that remains in the water in excess with respect to that introduced for the exchange and subsequent analysis of the CO2 present in the headspace will determine the isotopic ratio of O in the water. Using a gentle stream of He, the CO2 in the headspace of the sample container continuously passes through a Valco sampling port. Multiple analyses can be achieved by switching the contents of the sample loop into a gas chromatography column every 90 s.

ACCEPTED MANUSCRIPT 9.4.2 Isotopic Composition of Deuterium A mass spectrometer (Delta Plus XP) coupled to a TC/EA reactor was used to determine the Deuteriumisotope ratio. A properly calibrated autosampler takes an aliquot of the sample (0.8 μl) and injected it into the reactor. The reactor is set to 1450°C, which is a temperature that allows water pyrolysis. Once the two H and O atoms are separated, the H atom is sent to the mass SMOW; the uncertainties were ±0.1% for δ18O and ±1% for δD.

RI

9.5 Chemical Composition of Dissolved Gases

PT

spectrometer and determined. The isotopic values for the waters are expressed in δ‰ versus V-

SC

Dissolved gases were sampled and analyzed according to the method described by Capasso and Inguaggiato (1998), which is based on the equilibrium partition of gas species between a liquid and

NU

a gas phase after introducing a host gas (Ar) into the sample. The analysis was performed using a gas chromatograph (Perkin Elmer Clarus 500) equipped with a double detector (thermal

MA

conductivity detector [TCD] and a flame ionization detector [FID] with a methanizer) and Ar as the carrier gas. He, H2, O2 N2, and CO2 were measured using the TCD detector, while CH4 and CO were

D

determined using the FID detector coupled to the methanizer.

PT E

9.6 Isotopic Composition of Dissolved Gases 9.6.1 C-Isotope Composition of TDIC

CE

A fast and completely automated procedure was used to determine the δ13C of total dissolved inorganic C (TDIC) in water (δ13CTDIC) (Capasso et al., 2005). This method is based on the

AC

acidification of water samples transforming the TDIC species into CO2. Water samples are directly injected by syringe into vials with screw caps that have a pierceable rubber septum. A GasBench II system was used both to flush pure He into the vials and to automatically dispense a fixed amount of H3PO4. Full-equilibrium conditions between the produced CO2 and water are reached at a temperature of 70°C (±0.1°C) in less than 24 hours. C-isotope ratios (13C/12C) were measured using a mass spectrometer (Delta V Plus) connected on-line to the GasBench II system.

9.6.2 Isotopic Composition of He

ACCEPTED MANUSCRIPT The rapid and accurate technique used to determine the dissolved He-isotope ratios in groundwaters (Inguaggiato & Rizzo, 2004) is based on the extraction and subsequent equilibrium of dissolved gases in an added host gas phase. Ultrapure N 2 is placed in glass flasks (250 cc) containing water samples, and were hermetically sealed after being collected. After shaking in an ultrasonic bath for 10 minutes, an aliquot of the separated gas phase was removed from the flask for MS analysis. The elemental and isotopic composition of He as well the 4He/20Ne ratios were

PT

determined by separately admitting He and Ne into a split-flight-tube mass spectrometer (Helix SFT). The measured elemental and isotopic compositions of He were calibrated using the 3

RI

atmospheric standard. The analytical error of the He isotope analysis was less than 0.3%. The He/4He ratios were corrected for the atmospheric contamination on the basis of their 4He/20Ne

SC

ratios (Sano & Wakita, 1985), and are reported as R/Ra values (where Ra is 1.39×10–6). The sampled gases were analyzed using a gas chromatograph (Perkin Elmer Clarus 500) equipped with

NU

a double detector (a TCD plus a FID with a methanizer) using Ar as the carrier gas and a 3-m packed column (Restek Shincarbon ST). He, H2, O2, N2, and CO2 were measured using the TCD

MA

detector, while CH4 and CO were determined using the FID detector coupled to the methanizer.

D

9.7 Soil Degassing Processes

PT E

The total flux measured from the soils is produced by both diffusive and advective degassing processes. Lower values of ΦCO2 are driven by diffusion processes and are proportional to

CE

concentration gradient expressed by the first law of Fick (Fick, A. (1855).): Φd = –νD(dC/dx)

AC

where ν and D represent the soil porosity and the diffusion coefficient, respectively. The minus sign indicates that the molecules will move down a concentration gradient. In contrast, high values of flux are driven by advective processes that involve a force such as pressure gradient dP/dx. Advective flux Φa is well described by Darcy’s law (Darcy, H.,1856; Whitaker, S., 1986): Φa = K/μ(dP/dx) where K is the specific permeability of the soil and μ is the viscosity of the fluid. The total degassing flux of volatiles from the soils in geothermal and volcanic systems is the result of the combination of these two kinds of fluxes in a different percentage on the basis of the power of degassing.

ACCEPTED MANUSCRIPT The dynamic accumulation chamber modified and adapted to measure soil CO2 fluxes in geothermal and volcanic areas by Chiodini et al. (1998) and marketed by West Systems was utilized in the program of geochemical surveillance at the Aeolian volcanoes. This method was universally accepted by the scientific community for investigating soil CO 2 fluxes in volcanic areas both in both general surveys and also for continuous geochemical monitoring. The West System portable flux meter utilized in this discontinuous field campaign was equipped with an LI-COR

PT

detector containing an IR sensor for CO2. The IR sensor is based on an IR differential absorption; that is,the correlation of intensity of IR radiation within a certain band of wavelengths of a

RI

standard sample gas. Each molecule absorbs a fraction of the incident radiation, and the

SC

absorption bands characterize every molecule in an unique way. 9.8 Principles of Method for Measuring Plume Ratios

NU

9.8.1 Multi-GAS

MA

The Multi-GAS system hosts a Gascard II IR spectrometer for CO2 determination (calibration range, 0–4000 ppmv; accuracy, ±2%; resolution, 0.8 ppmv) and two electrochemical sensors that are specific to SO2 (model SO2-S-100, Membrapor: calibration range, 0–100 ppmv; accuracy, 2%;

D

resolution, 0.5 ppmv) and H2S (model H2S-S-50, Membrapor: calibration range, 0–50 ppmv;

PT E

accuracy, 2%; resolution, 0.05 ppmv) (Aiuppa et al., 2005b). The volcanic gas plume is actively pumped during the measurements (at an average flow rate of 0.6 l m–1) into IR and electrochemical cells that work in series. Both sensors are connected to a data-logger board

CE

enabling data capture and logging. The Multi-GAS system is operated daily at Stromboli for four cycles per day, each lasting 30 minutes. During this operation, data are captured from the sensors

AC

every 9 s, for a total of 200 determinations of CO2 and SO2 concentrations during each 30-minute cycle. At the end of each cycle, a radio link is used for automatic data transfer from the remote Multi-GAS system to the base station in Palermo, where data are analyzed. CO2 and H2O have been measured simultaneously with an LI-840 NDIR closed-path spectrometer (CO2 measurement range, 0–3000 ppm; accuracy, ±1.5%; H2O measurement range, 0–80 ppt; accuracy, ±1.5%) (Shinohara et al., 2008), while a sensitive electrochemical sensor was used for SO2 (model 3ST/F, Cod.TD2D-1A, City Technology: calibration range, 0–30 ppmv; repeatability, 1%).

9.8.2 FTIR spectroscopy

ACCEPTED MANUSCRIPT OP-FTIR spectroscopy makes it possible to analyze the different amounts of chemical species (e.g., HCl, HF, SO2, CO2, H2O, CO, and OCS) emitted in the volcanic plume based on their absorption in the IR region (Edmonds et al., 2002), without altering their natural course during the ascent into the atmosphere. Grutter et al. (2003) used OP-FTIR spectroscopy to measure trace gases over Mexico City. Their report was the first on the concentration profiles of C2H2, C2H4, C2H6, C3H8, and CH4 in that region, and correlations between the profiles and wind directions were assessed to

PT

determine the main contributors to the measured profiles. The source of IR radiation used in the measurements can be of natural origin, such as the sun, moon, a lava flow, hot rocks, exposed

RI

magma, or thermal emissions of the volcanic plume itself, or artificial light such as that from an IR lamp (Oppenheimer et al., 1998, 2011). The first determination of HCl and SO2 in volcanic gas

SC

using an FTIR spectrometer was performed by Mori et al. (1993) during a study of a dome lava extrusion of Unzen Volcano.

NU

Other gases including H2O, CO2, CO, COS, SO2, HF, and SiF4 were measured using remote FTIR spectroscopy (Mori & Notsu, 1997; Burton et al., 2000; Mori & Notsu, 2008; Stremme et al., 2011).

MA

Major chemical species that are present in volcanic gas are typically measured using remote FTIR spectroscopy, except for H2S, which is hidden by large H2O absorption, and H2, which is does not

D

absorb IR (Notsu & Mori, 2010). A telescope-based FTIR spectral radiometer was used to study the

PT E

compositions of volcanic gases at several active volcanoes in Japan by Mori et al., 2008. IR radiation from hot lava was used as the IT source for one of the volcanoes monitored, IR radiation from the hot ground surface was used for some others, and scattered solar light for the rest. The

CE

observations suggest that HCl/SO2 and HF/HCl ratios are the most promising parameters for reflecting volcanic activity among various parameters observable in remote FTIR measurements

AC

(Notsu & Mori, 2010).

9.8.3 DOAS

DOAS (Platt, 1994; Platt & Stutz, 2008) is a new alternative remote-sensing method used for quantifying the concentrations of different trace gases. It is based on the principles of absorption spectroscopy (Bouguer–Beer-Lambert law) by collecting UV spectra using scattered sunlight from the sky (Bobrowsky et al., 2003; Edmonds et al., 2003; Vita et al., 2012). The passive DOAS technique makes it possible to quantify columns of different volcanic gases emitted from active volcanoes (SO2, NO2, and BrO) by collecting UV spectra using scattered sunlight from the sky

ACCEPTED MANUSCRIPT (Bobrowsky et al., 2003; Edmonds et al., 2003; Vita et al., 2012). Active DOAS uses an artificial light source, and has been applied for monitoring a large numbers of trace gases (e.g., NO 2, NO, HCHO, SO2, O3, and CS2) and many aromatic hydrocarbons since its first applications in the 1970s (Platt and Perner, 1983; Stutz and Platt, 1996; Kern, 2009; Kern et al., 2006, 2009a, 2009b). In the framework of the NOVAC project, a scanning DOAS system was designed to measure volcanic gas emissions by UV absorption spectroscopy by the Optical Remote Sensing Group in the Department

PT

of Radio and Space Science at Chalmers University of Technology, Göteborg, Sweden. The single system comprises one spectrometer from Ocean Optics Company, an embedded personal

RI

computer, a GPS receiver, a fiber, and a telescope. Technical aspects of continuous monitoring with scanning DOAS systems such as instrument design, station networking, and measurement

SC

geometries were described in Zhang (2005), Galle et al. (2010), and Johansson (2009). Data analysis was significantly improved by identifying and considerably reducing the two main sources

NU

of error: the wind speed at the plume height and radiative transfer effects (Kern et al., 2009a). Continuous measurements of SO2 fluxes were integrated with campaigns of mobile SO2 flux

MA

measurements that started on June 27, 2006 using the passive Mini-DOAS system with sky radiation as the light source. The monthly measurements involved several measurement traverses

D

per day (i.e., usually about five) from a boat performing transects beneath the plume with MiniDOAS instruments (USB2000), consisting of a UV spectrometer manufactured by Ocean Optics

PT E

(spectral range, 245–400 nm; resolution, ≈0.7 nm) and a vertically pointing telescope with a 7mrad field of view, with a circular-to-linear converter, and four 200-µm fiber bundles that connect

CE

the telescope to the fiber. A USB cable connects the spectrometer to a laptop computer to provide power and a means of data transfer. Software control of the USB2000 instrument was achieved J scripts

executed

in

DOASIS

software

(https://doasis.iup.uni-

AC

using

heidelberg.de/bugtracker/projects/doasis/) to save and analyze spectra, providing real-time concentration readings. The geographic coordinates of each spectrum were obtained using a handheld GPS receiver. Details of the DOAS routines used and the flux calculations can be found in Galle et al. (2003).

9.9 Continuous Monitoring Procedures 9.9.1 Testing Procedures

ACCEPTED MANUSCRIPT Any new monitoring device for volcanic surveillance purposes needs a period of testing that is approached as a real research activity, before being a civil protection service tool. The research/test period must have a defined duration that depends on the state of activity. It should last at least 2 years in order to define the first reference range of acquired data necessary to interpret the further evidence also on an empirical basis and to register the possible annual cycle. Moreover, the activities performed during the research/test period should aim to address the

PT

following basic aspects: 1. Selecting the new parameters to be registered. Some parameters are expected to

RI

increase the observational network (e.g., physical and chemical determinations, gas flux, and heat flux), while others (e.g., voltage, ambient temperature, moisture, and ambient

SC

pressure) supply technical support that is important for ensuring acceptable data

2. Choosing the new monitoring site.

NU

quality.

3. Defining the relationship between the selected site and the entire observational

MA

network. This can be achieved by performing cross-correlations among spatial and temporal variations of different series data, including the newly recorded ones.

D

4. Verifying if the monitoring system fits well with the environmental conditions, and

PT E

evaluating the cost-to-benefit ratio of maintenance. However, after collecting data during the test period, the cost-to-benefit ratio for each new monitoring device should reduce with time.

management.

CE

Paying attention to these points reduces the risks of failures and incorrect decisions about data

AC

9.9.2 Data Annotation Process While INGV-Palermo is the data repository, a website is the best choice for making the data available to a wide user community, and so we are currently working on this based on a specific data policy at the institutional level. The radio transmission of packet data is used as the data annotation process, where each data packet consists of a sequence of serial bits including a header and payload checksum. The continuous monitoring system is structured in a central unit, which polls the peripheral stations, more peripheral stations connected via radio transceivers or cell-phones, and one or more subcenters that poll the devices connected via radio. A data logger stores data at the remote site

ACCEPTED MANUSCRIPT at predefined intervals (e.g., every hour, every 2 h, or every day) according to the state of the system. A central server located at INGV-Palermo acts as the master station that interrogates the data loggers to download the data packets from their memory and processes the transmitted data strings to extract information (e.g., channel, time, and measurement data). The data are automatically saved to a text file (*.csv), and are made available in the form of tables. The communication with the peripheral stations occurs via a RF modem (at 600–1200 baud) or

PT

cellphone (at 9600 baud). The series of data of geochemical parameters from the monitoring network set up for the surveillance of volcanic activity are acquired within different time windows

RI

depending on the variability shown by the parameter and on the level of activity (i.e., intereruptive periods, periods of unrest, or periods of eruptive activity). The automatic acquisition

SC

can be changed remotely by the responsible of the network.As an example, the data of HTFs from Vulcano Island are acquired and stored at the Vulcra and Vulcra2 peripheral stations (Fig. 15a). An

NU

intermediate position on Lipari Island acts as a radio link and retransmits data to the central unit located at INGV-Palermo. The interval between automatic measurements is usually programmed

MA

as 1 hour. All of the monitoring data acquired since the beginning of the monitoring activity have been analyzed and integrated, as shown in Fig. 13a,b. The latest data analysis procedure began in

D

2009, which yielded a homogeneous series of temperature data, of which 179 MB were used to

PT E

determine the daily mean temperature during 1990–2016 at four high-temperature monitoring sites (range, 250–450°C) located at the top of the active La Fossa cone (Vulcano, Italy). The recorded temperatures of fumaroles were validated by comparison with other data sets. The

CE

ambient temperature and supply voltage were considered first, followed by rainfall data and discontinuous measurements from all of the fumaroles, including also the chemical and isotopic

AC

compositions of gases associated with steam release. A personal computer was used as the master station for some new stations during the initial acquisition period until data processing was arranged to being included in the central server. New stations and new monitoring devices have been added to the monitoring network after several years of testing. 9.10 Monitoring Methods for Evaluating the Thermal Output 9.10.1 HTF Monitoring The time series of temperatures of HTFs is of great interest in any volcanic area because the temporal variations measured at the surface are directly related to the heat and mass outputs that

ACCEPTED MANUSCRIPT represent geochemical information about the deeper sources and so also reflect changes in the thermodynamic conditions. The time-series analysis of long-term monitoring of ground surface temperatures—if this is included as part of wider observations—could also contribute to the evaluation of global warming, with inferences on its effect on the natural environment, especially given that the data have been measured for more than 2 decades. The first system for automatically and continuous measuring geochemical parameters of

PT

volcanic activity at the Aeolian Islands was started in June 1984 (Carapezza et al., 1984; Badalamenti et al., 1986;). The stations were installed in the Lipari–Vulcano system, and the

RI

original monitored sites were located within the known fumaroles areas at the top of La Fossa cone and, in Levante Bay (Figs.2,15a,b), and onthe island of Lipari (north-east).

SC

The extreme environmental conditions of La Fossa cone have always represented a considerable challenge for both researchers and technicians, due to the presence of acidic gases, high moisture

NU

levels, and heavy rainy periods alternating with very dry seasons. Moreover, no power supply is available at the top of the active cone, and the output temperatures ranged from normal ambient

MA

values to very high values, with a maximum recorded value of 550°C and a temperature of 670°C measured in another vent near to the monitored site. The data were collected for the surveillance

D

of volcanic activity, and have been periodically renewed by the research group belonging to INGV

PT E

(ex Istituto di Geochimica dei Fluidi), the Gruppo Nazionale per la Vulcanologia and the Consiglio Nazionale delle Ricerche. Our opinion is that before changing to using innovative techniques, they should be tested by comparing the new results with the previous ones, in order to ensure

CE

continuity of data acquisition. Therefore, while we have renewed some aspects of the monitoring systems, we did not abandon the former monitoring sites nor chosen to adopt new procedures so

AC

as to not interrupt such a long record of data. We confirmed this to be the correct monitoring procedure by making multidisciplinary comparisons of the long-term data series. The reducing capacity (RC) and temperature of fumarole emissions were the original parameters acquired continuously for geochemical monitoring. Carapezza et al. (1984) applied the fuel cell method to measure the RC of volcanic gases. Many authors expected temporal variations of the RC of fumarole emissions to be directly correlated with changes in seismic and volcanic activity, since the RC of a fluid mixture reflects the H2 content of gas emissions (Carapezza et al., 1980; Sato, 1978; Sato & McGee, 1982; Wakita, 1996; Wakita et al., 1980). The acquisition of RC for volcanic surveillance at Vulcano was stopped after a few years due to the high maintenance requirements of the sensors, which were often affected by drift errors. However, a direct

ACCEPTED MANUSCRIPT relationship between changes in RC and volcano-tectonic events had been observed (i.e., Oskarsson, 1984; Giammanco et al., 1998). Di Martino et al. (2013) recently tested a more-specific detector using new experimental procedures and data acquired in a different volcanic area. The monitoring network at La Fossa cone (Vulcano) has been implemented progressively as new tools have become available and their efficacy was tested for the specific environmental conditions encountered in the field. The stations located within the main fumaroles are equipped

PT

with an in-house data logger that stores the measurements made by a thermocouple fixed at depth of 50 cm; the sensor is protected by a steel case designed to ensure a reasonable technical

RI

life (1–2 years) under severe operating conditions involving high temperatures and mixtures of water vapor and strong acidic species as HF, HCl, and H2SO4. Temperatures measured every hour

SC

are stored in the local memory of the logger and transmitted via a GSM modem or radio link to the operating center at INGV-Palermo.

NU

Sensor breakdowns, power outages, and interruptions to the transmission lines were the most common failures during the monitoring period. The outlet temperatures are measured by

MA

chromel–alumel thermocouples (with a sensitivity of approximately 41 μV °C–1) in stainless-steel and Teflon probes at the vents exhibiting temperatures higher and lower than 300°C, respectively.

D

Each probe is inserted into the steaming vent at a depth of 0.5 m in order to optimize the contact

PT E

between the sensor and ascending fluids and to reduce external disturbances. The electronic devices of the monitoring system are kept away from the fumarole emissions, being placed inside stations Vulcra and Vulcra2. The measurement accuracy is maximized by applying the cold-

CE

junction compensation (CJC, https://www.ia.omron.com/support/glossary/meaning/144.html) technique, and good thermal contact is ensured between the input connectors and the measuring

AC

instrument by performing regular maintenance. CJC is an automation applied by an electronic process that considers the voltage produced by temperature variations at the cold joint; in this configuration, “cold” refers to the ambient temperature while “hot” refers to the temperature at the fumarole output. Any loss of the normally good thermal contact at the cold junction— frequently due to extreme weather conditions around the fumarole—will result in disturbances in the time series of temperature data that may mask the real temperature variation of the fumaroles. Such erroneous data are identified by comparing the fumarole records with the ambient temperature variations, and any uncertain values are removed from the time-series data until the necessary maintenance fieldwork is performed. During technical maintenance of the system, episodic temperature measurements are made near to the monitored locations using

ACCEPTED MANUSCRIPT chromel–alumel thermocouples in portable devices (error <±1%). Non-contact IR sensors (called pyrometers) have been tested since 2011 with the results compared to the thermocouple measurements. Pyrometer measurements are based on the radiative heat transfer process. The measurements were made close to the ground surface (at height of 0.5 m above the Earth surface ), and they indicated the same temperatures as the direct measurements made below the ground level, just at some vents with particularly high steam outputs. This was an expected result, since

PT

the measurement probe is inserted at a fixed depth (0.5 m below the ground level) where the high flux keeps the temperature constant along the shallow soil profiles, in order to minimize the

RI

external disturbances and to avoid undesirable short-term effects. Measurements made using noncontact measurements (e.g., IR sensors) can accurately reflect the temperature of the

SC

fumarole output only when the multiphase system in the shallow soil is in thermal equilibrium with the pressurized steam.

NU

Additional geochemical investigations have been performed since 1984 at the island depending on the available resources. All of the updated field data, included those obtained at many more

MA

acquisition points, have been taken into account to interpret and implement the results provided by continuous monitoring of temperature (Diliberto 2017).

D

9.10.2 Results from Temperature Monitoring of High Temperature Fumaroles

PT E

The validated data of High Temperature Fumaroles (HTFs) indicate temperatures ranging from 250°C to 450°C in the longest time series (Fig. 13a,b). The delayed validation process has identified

CE

some secondary gaps of data in the time series, since in some periods the field maintenance was difficult. Whether some technical noises appeared in the monitoring records, such as short-term random variations and defective CJCs, the incorrect temperatures were deleted during the

AC

validation process. The sampling interval finally chosen for the validation of the long-term monitoring is one value per day. The data have been used to follow the variations of overheated steam emissions at Vulcano (Aeolian Islands, Italy). The variation of temperature is directly correlated with variations of heat and energy released from the magmatic system, and is one evident surface effect of the continuous advection of fluids that affects such an active geodynamic system. In some cases the geochemical crisis involved the base area of the cone, as revealed by multiparameter comparisons (Diliberto et al., 2002 and references therein; Granieri et al., 2006; Cannata et al., 2012; Diliberto, 2017), increasing the gas hazard in Vulcano Porto village and more generally suggesting an increased probability of volcanic activity. Some major trends lasting more

ACCEPTED MANUSCRIPT than 4 years have been observed in the recorded temporal variations. Application of the fast Fourier transform to subsets of data revealed the main cyclic components in recent years (1998 to 2011) and simplified the comparison among various other time-series data (Diliberto, 2013). This data validation process demonstrated that the continuous monitoring of fumaroles can yield real-time data while also minimizing the need to access the extreme volcanic environment. The surface temperature is sensitive to many influencing factors, but the relationships with the hidden

PT

energy sources (hydrothermal, magmatic, and geodynamic) that we want to study can be clarified by applying a multiparameter approach. The long-term monitoring allowed comparisons with

RI

different subsets of data and confirmed the direct relationship with the changes in magmatic gas release, while comparisons with other geochemical and geophysical parameters highlighted a

SC

common mechanism underlying for the main temporal variations. Moreover, in a region (i.e., the southern Tyrrhenian Sea) characterized by large releases of seismic energy, the recorded change

NU

in temperatures showed interesting temporal relationships with local volcano-tectonic activity (Aubert et al., 2008; Milluzzo et al., 2010; Cannata et al., 2012; Madonia et al., 2013). The

MA

monitoring of fumarole temperatures indicates that, in the thermal area of La Fossa cone, external agents such as rain and barometric perturbations only exert minor and short-term effects on the

D

temperatures of HTFs. In contrast, the ensemble of temperatures recorded in the time series for

PT E

1991–2015 define the temporal variation of the hydrothermal energy released from the summit area of the active cone. The output temperature of HT fumaroles is temporally related to many other geophysical and geochemical variables, even when the physical conditions of the fluids are

CE

highly unstable, responding promptly to even minor external perturbations. The temperature data supplied by the monitoring system described here can be analyzed from different point of views.

AC

The same monitoring system could be installed at other quiescent volcanoes to follow any change in intensity and extension of thermal anomalies, related to hydrothermal and or magmatic steam outputs. This monitoring could be a starting point to setting up a surveillance system at a volcanic observatory for remotely following the evolution of closed-conduit volcanic systems. Moreover, well-defined monitoring networks would help the calibration of IR thermal images and of different remote-sensing techniques for the interpretation of changes observed at the Earth’s surface. The ensemble of temperatures recorded from 1991 in the time series herald a change in the partitioning of the energy release among the summit area of the active cone and the peripheral areas. This change occurred after March 1996 and it has been verified in models of fluid geochemistry (Taran, 2011; Paonita et al., 2013). The changes revealed by the hydrothermal

ACCEPTED MANUSCRIPT release can be temporally related to the thermal output. By utilizing a continuous monitoring network we minimize the need to access unsafe areas while continuing to record direct information about local changes in emissions related to the natural convection of vapors and gases. Multiparameter comparisons with other data series (e.g. chemical and isotopic compositions of fluids, diffuse CO2 flux, deformation, and rate of energy release evaluated on local

PT

and regional scales) could suggest the next evolution of this volcanic system.

RI

9.10.3 Results From Steam Heated Soil Monitoring

SC

A research project performed from 2005 to 2007 verified a practical method for monitoring the heat flux from the ground in the closed-conduit volcanic system of La Fossa cone. The response of

NU

this new measurement tool, and the conditions for evaluating the soil heat flux based on temperature measurements, were verified also by experimental tests performed under controlled

MA

conditions (Diliberto, 2007). Aubert (1999) introduced the theoretical background for the evaluation and monitoring of thermal release, starting from ground temperatures and testing the method for the south-eastern flank of Mt Etna . The new monitoring system consists of a data

D

logger that once per hour stores the ground temperatures at six points lying on a shallow vertical

PT E

profile of soil along the main direction of diffuse heat flux. To estimate the heat flux from the measured temperature gradients, the most important conditions required at the site are (i) the surface temperature of the site being lower than boiling point and (ii) the heat flux being only in

CE

the direction of the monitored profile (Z direction and depth over Z1 in Fig. 17), while the monitoring profile should get close to or reach the boiling point. The success of this monitoring

AC

approach is crucially dependent on the site location (Diliberto et al., 2017). The best site for monitoring the heat flux is where the condensation zone is just beneath the monitored profile; such areas can be defined to be steam heated soils (SHS), and according to the direct relationships with other volcanological features, they are worthy of greater attention in scientific investigations of volcanic activity (Aubert & Alparone, 2000; Aubert et al., 2008; Diliberto 2011, 2013, 2017; Cannata et al., 2012). The sketch of the SHS soil profile in Figure 17 shows the main layer where different heat transfer regimens prevail, and the main contribution to the vertical balance of the heat flux can be considered. The first condensation level is indicated as Z1, which is at the bottom of the layer where heat production occurs by vapor condensation. In this intermediate layer the heat transfer

ACCEPTED MANUSCRIPT is partially conductive and partially convective [(Cd+Cv)HT in Fig. 17]. Below the Z2 level at ambient pressures, the temperature of the ground is buffered by the boiling point of the water vapor (i.e., 100°C; 1 bar), which is the main fluid phase in the porous structure, and here heat transfer occurs mainly via convection (Cv HT in Fig. 17). In conclusion, when the two above-described necessary conditions are fulfilled at a monitoring site, the heat flux be calculated in the upper part of the profile (i.e., above the Z1 level), starting

PT

from the temperature profile, since conductive transfer (CdHT in Fig. 17) represents the main component of the heat flux along the direction profile. This condition is fulfilled only if there is a

RI

very high linear temperature gradient (R2≈1 in Figs. 17a; 18a,b; 19). The limit of the conductive layer (Z1) calculated by the best-fit equation coincides with the depth in the ground where the soil

SC

is at boiling point. Above this limit the simplified heat-flow equation (φH) can be applied up to the

NU

ground level (Aubert, 1999): (1) φH = λ(t5 – t2)/(zt5 – zt2)

MA

where λ (W m−1 °C−1) is the thermal diffusivity of the ground, t5 and t2 are the temperatures (°C) measured at different levels in the ground, R2dt/dz<0.999 indicates the linear relationship among

D

the variables under consideration, and zt5 and z t2 are the reference depths (in meters, see Fig. 17).

PT E

If the steam flux increases consistently, the condensation level of the SHS may reach the ground surface, even at the monitoring site. In that case the simplified heat-flow equation cannot be applied and the heat flux from the surface cannot be evaluated using temperature profiles, since

CE

all of the monitored levels will reach boiling point and there will be no diffusive heat transfer. However, the particular condition of an anomalously large heat release can still be followed when

AC

boiling temperature values are registered in the complete temperature profile; see for example the data in Figures 18a and 18b, that were recorded during a geochemical crisis.

Authors thanks the colleagues of INGV Palermo for their useful help during the field campaigns. The chemical and isotopic analysis were performed thanks to Geochemical Laboratories of INGV Palermo. This work was supported by Italian Civil Defence (DPC Italia).

10. References Aiuppa, A., Bertagnini, A., Métrich, N., Moretti, R., Di Muro, A., Liuzzo, M., & Tamburello, G. (2010a). A model of degassing for Stromboli volcano. Earth and Planetary Science Letters, 295(1), 195-204.

ACCEPTED MANUSCRIPT Aiuppa A, Burton M, Muré F, Inguaggiato S (2004) Intercomparison of volcanic gas monitoring methodologies performed on Vulcano Island, Italy. Geophys Res Lett 31:L02610. doi:10.1029/2003GL018651 Aiuppa, A., Burton, M., Caltabiano, T., Giudice G., Gurrieri S., Liuzzo, M., Murè, F., Salerno, G., (2010b) Unusually large magmatic CO2 gass emission prior to a basaltic paroxysm. Geophysical research letters, vol. 37, L17303

PT

Aiuppa, A., Dongarrà, G., Capasso, G., Allard, P., 2000. Trace elements in the thermal groundwaters of Vulcano Islands (Sicily). J. Volcanol. Geotherm. Res. 98: 189-207.

RI

Aiuppa A, Federico C (2004) Anomalous magmatic degassing prior to the 5th April 2003 paroxysm on Stromboli. Geophys Res Lett L14607. doi:10.1029/2004GL020458

SC

Aiuppa A, Federico C, Giudice G, Gurrieri S, Valenza M (2006) Hydrothermal buffering of the SO2/H2S ratio in volcanic gases: Evidence from La Fossa Crater fumarolic field, Vulcano Island. Geophys Res Lett 33:L21315, doi:10.1029/2006GL027730

NU

Aiuppa, A., C. Federico, A. Paonita, G. Pecoraino, and M. Valenza (2002) S, Cl and F Degassing as an Indicator of Volcanic Dynamics: The 2001 Eruption of Mount Etna, Geophys. Res. Lett., 29(11), doi:10.1029/2002GL015032, 2002.

MA

Aiuppa A, Federico C, Giudice G, Giuffrida G, Guida R, Gurrieri S, Liuzzo M, Moretti M, Papale P (2009) The 2007 eruption of Stromboli volcano: Insights from real-time measurement of the volcanic gas plume CO2/SO2 ratio. Journal of Volcanology and Geothermal Research, 182(3), 221-230.

PT E

D

Aiuppa A, Inguaggiato S, McGonigle AJS, O'Dwyer M, Oppenheimer C, Padgett MJ, Rouwet D, Valenza M (2005a) H2S fluxes from Mt. Etna, Stromboli, and Vulcano (Italy) and implications for the sulfur budget at volcanoes. Geochim Cosmochim Ac 7:1861-1871

CE

Aiuppa, A., Federico, C., Giudice, G., Gurrieri, S., (2005b) Chemical mapping of a fumarolic field: La Fossa Crater; Vulcano Island (Aeolian Island, Italy) Geophysical research letters, vol. 32, L13309

AC

Aiuppa, A., Moretti, R., Federico, C., Giudice, G., Gurrieri, S., Liuzzo, M., Papale, P., Shinohara, H., Valenza, M., 2007. Forecasting Etna eruption by real time evaluation of volcanic gas composition. Geology 35 (12), 1115–1118. doi:10.1130/G24149A. Allard P. (1983) The origin of hydrogen, carbon, sulphur, nitrogen and rare gases in volcanic exhalations: evidence from isotope geochemistry. In: Tazieff H., Sabroux Jc (eds) Forecasting Volcanic Events. Elsevier, Amsterdam, 337-386 Allard, P., Aiuppa, A., Loyer, H., Carrot, F., Gaudry, A., Pinte, G., ... & Dongarrà, G. (2000). Acid gas and metal emission rates during long‐lived basalt degassing at Stromboli volcano. Geophysical Research Letters, 27(8), 1207-1210. Allard P., Burton M., Murè F.(2005). Spectroscopic evidence for a lava fountain driven by previously accumulated magmatic gas. Nature, 433 (2005), pp. 407-410 Nature 433, 407-410 (27 January 2005) | doi:10.1038/nature03246

ACCEPTED MANUSCRIPT Allard, P., Aiuppa, A., Burton, M., Caltabiano, T., Federico, C., Salerno, G., & La Spina, A. (2008). Crater gas emissions and the magma feeding system of Stromboli volcano. The Stromboli Volcano: An Integrated Study of the 2002-2003 Eruption, 65-80. Allard, P., Carbonnelle, J., Metrich, N., Loyer, H., & Zettwoog, P. (1994). Sulfur output and magma degassing budget of Stromboli volcano. Nature, 368(6469), 326-330. Allard, P., Aiuppa, A., & Loyer, H. (1999). Airborne determinations of gas and trace-metal fluxes in volcanic plumes of Mt. Etna and Stromboli. Monitoring Volcanic Risks by Remote Sensing.

RI

PT

Alparone S., Cannata A., Gambino S., Gresta S., Milluzzo V., Montalto P. (2010) Time–space variation of volcano-seismic events at La Fossa (Vulcano, Aeolian Islands, Italy): new insights into seismic sources in a hydrothermal system. Bull. Volcanol., 72, 803-816, doi: 10.1007/s00445-010-0367-6

SC

Aubert M. (1999) Practical evaluation of steady heat discharge from dormant active volcanoes: case study of Vulcarolo fissure _Mount Etna, Italy. Journal of Volcanology and Geothermal Research, v. 92, p. 413–429.

NU

Aubert M. and Alparone S., (2000) Hydrothermal convective flux variation related to a seismo –tectonic crisis in the La Fossa of Vulcano (Italy). C.R.Acad.Sc.Paris, 330, 2000, 603-610.

MA

Aubert M., Diliberto S., Finizola A, Chébli, Y. (2008). Double origin of hydrothermal convective flux variations in the Fossa of Vulcano (Italy). Bull. Volcanol., Vol. 70, pp. 743-751: DOI 10.1007/s00445-0070165-y.

PT E

D

Badalamenti, B., Falsaperla, S., Neri, G., Nuccio, P.M., Valenza, M., 1986. Confronto preliminare tra dati sismici e geochimica nell’area Lipari–Vulcano. Consiglio Nazionale Ricerche Gruppo Nazionale Vulcanologia, Bollettino, 1:37–47

CE

Badalamenti, B., Gurrieri, S., Hauser, S., Parello, F., Valenza, M., 1988. Soil CO2 output in the island of Vulcano during the period 1984–1988: surveillance of gas hazard and volcanic activity. Rend. Soc. Ital. Mineral. Petrol. 43, 893–899.

AC

Badalamenti B., Chiodini G., Cioni R., Favara R., Francofonte S., Gurrieri S., Hauser S., Inguaggiato S., Italiano F., Magro G., Nuccio P. M., Parello F., Pennisi M., Romeo L., Russo M., Sortino F., Valenza M., and Vurro F. (1991) Special field workshop at Vulcano (Aeolian Islands) during summer 1988: Geochemical results. Acta Vulcanol., 1, 223-227. Barberi F., Neri G., Valenza M., and Villari L. (1991) 1987-1990 unrest at Vulcano. Acta Vulcanol., 1, 95-106. Baubron C., Allard P., and Toutain J.P. (1990) Diffuse volcanic emissions of carbon dioxide from Vulcano Island, Italy. Nature, 344, 51-53, doi: 10.1038/344051a0 Baubron J.C., Allard P., Sabroux J.C., Tedesco D. and Toutain J.P. (1991) - Soil gas emanations as precursory indicators of volcanic eruptions. J. Geol. Soc. Lond. 148: 571-576. Bischoff J.L., Rosenbauer R.J., Fournier R.O. (1996) The generation of HCl in the system CaCl2-H2O: Vapourliquid relations from 380-500°C. Geochim. Cosmochim. Acta 60, 7-16. Bobrowski N, Hönninger G, Galle B, Platt U (2003) Detection of bromine monoxide in a volcanic plume. Nature 423:273-276

ACCEPTED MANUSCRIPT Bolognesi, L. (1997). A tentative correlation between seismic activity and changes in the composition of thermal waters on Vulcano Island, Italy. Geothermics, 26(3), 379-392. Bolognesi, L. (2000). Earthquake-induced variations in the composition of the water in the geothermal reservoir at Vulcano Island, Italy. Journal of volcanology and geothermal research, 99(1), 139-150.

PT

Bolognesi, L., D’Amore, F. (1993) Isotopic variation of the hydrothermal system on Vulcano Island, Italy. Geochim. Cosmochim. Acta, 9, 2069-2082, doi: 10.1016/0016-7037(93)90094-D.

RI

Brusca, L., Inguaggiato, S., Longo, Madonia, P., Maugeri, R., 2004. The 2002-2003 eruption of Stromboli (Italy): Evaluation of the volcanic activity by means of continuous monitoring of soil temperature, CO2 flux, and meteorological parameters. Geochem. Geophys. Geosys 5(12), doi: 10.1029/2004GC000732. issn: 1525-2027.

NU

SC

Bukumirovic, T., Italiano, F., Nuccio, P.M. (1997) The evolution of a dynamic geological system: The support of a GIS for geochemical measurements at the fumarole field of Vulcano, Italy. J. Volcanol. Geotherm. Res., 79, 253-263, doi: 10.1016/S0377-0273(97)00032-2.

MA

Burton MR, Oppenheimer C, Horrocks LA, Francis PW (2000) Remote sensing of CO2 and H2O emission rates from Masaya volcano, Nicaragua. Geology 28: 915–918. Burton, M., Allard, P., Muré, F., & La Spina, A. (2007a). Magmatic gas composition reveals the source depth of slug-driven Strombolian explosive activity. Science, 317(5835), 227-230.

PT E

D

Burton MR, Mader HM, Polacci M (2007b) The role of gas percolation in quiescent degassing of persistently active basaltic volcanoes. Earth Planet Sci Lett 264:46-59 Burton, M.R.; Caltabiano, T.; Murè, F.; Salerno, G.; Randazzo, D. SO2 flux from Stromboli during the 2007 eruption: Results from the FLAME network and traverse measurements. (2009) J. Volcanol. Geotherm. Res. 182 (2009) 214–220

CE

Burton, M.; Caltabiano, T.; Salerno, G., Mure, F.; Condarelli, D.; 2004. Automatic measurements of SO2 flux on Stromboli using a network of scanning ultraviolet spectrometers Geophys. Res. Abstr. 6, 03970

AC

Calanchi N., Capaccioni B., Martini M., Tassi F. and Valentini L. (1995) Submarine gas-emission from Panarea Island (Aeolian Archipelago): distribution of inorganic and organic compounds and inferences about source conditions. Acta Vulcanol. 7, 43–48. Calanchi N., Peccerillo A., Tranne C. A., Lucchini F., Rossi P. L., Kempton P., Barbieri M. and Wu T. W. (2002) Petrology and geochemistry of volcanic rocks from the island of Panarea: implications for mantle evolution beneath the Aeolian island arc (southern Tyrrhenian sea). J. Volcanol. Geotherm. Res. 115, 367–395. Caliro, S., Caracausi, A., Chiodini., G., Ditta., M., Italiano, F., Longo, M., Minopoli., C., Nuccio, P.M., Paonita, A., Rizzo, A., (2004) Evidence of a recent input of magmatic gases into the quiescent volcanic edifice of Panarea, Aeolian Islands, Italy. Geophys. Res. Lett. 31, L07619.

ACCEPTED MANUSCRIPT Cannata A., Diliberto S., Alparone S., Gambino S., Gresta S., Liotta M., Madonia P., Milluzzo V., Aliotta M., Montalto P. (2012) Multiparametric approach in investigating volcano-hydrothermal systems: the case study of Vulcano (Aeolian Islands, Italy). Pure App. Geophys., 169, 167-182, doi:10.1007/s00024-0110297-z. Capaccioni, B., Tassi, F., Vaselli, D., Tedesco, D., Rossi, P.M.L., (2005) The November 2002 degassing event at Panarea Island (Italy): the results of a 5 months geochemical monitoring program. Ann. Geophys. 48, 755–765.

PT

Capaccioni, B., Tassi, F., Vaselli, O., Tedesco, D., Poreda, R., (2007) Submarine gas burst at Panarea Island (southern Italy) on 3 November 2002: a magmatic versus hydrothermal episode. J. Geophys. Res. 112, B05201.

SC

RI

Capasso, G., Carapezza, M. L., Federico, C., Inguaggiato, S., and Rizzo, A., 2005, Geochemical monitoring of the 2002-2003 eruption at Stromboli volcano (Italy): precursory changes in the carbon and helium isotopic composition of fumarole gases and thermal waters: Bull.Volcanol., v. 68, p. 118-134.

NU

Capasso G., D’Alessandro W., Favara R., Inguaggiato S., Parello F. (2001) Interaction between the deep fluids and the shallow groundwaters on Vulcano Island (Italy). J. Volcanol. Geotherm. Res., 108, 187198, doi: 10.1016/S0377-0273(00)00285-7.

MA

Capasso G., Dongarra G., Hauser S., Favara R., Valenza M. (1992) Isotope composition of rain water, well water and fumarolic steam on the island of Vulcano, and their implications for volcanic surveillance. J. Volcanol. Geotherm. Res., 49, 147-155, doi: 10.1016/0377-0273(92)90010-B.

PT E

D

Capasso G., R. Favara, and S. Inguaggiato. (1997) Chemical features and isotopic composition of gaseous manifestations on Vulcano Island (Aeolian Islands, Italy): an interpretative model of fluid circulation. Geoch.et Cosmoch.Acta 61 (16):3425-3440. doi: 10.1016/S0016-7037(97)00163-4.

CE

Capasso, G., Favara, R., Inguaggiato, S., 2000. Interaction between fumarolic gases and thermal groundwaters at Vulcano Island (Italy): evidences from chemical composition of dissolved gases in waters. J. Volcanol. Geotherm. Res. 102: 309-318.

AC

Capasso, G., Favara, R., Francofonte, S., Inguaggiato, S. (1999) Chemical and isotopic variations in fumarolic discharge and thermal waters at Vulcano Island (Aeolian Island, Italy) during 1996: Evidence of resumed volcanic activity. J. Volcanol. Geotherm. Res., 88, 167-175, doi: 10.1016/S0377-0273(98)00111-5. Capasso, G., Federico, C., Madonia, P., & Paonita, A. (2014). Response of the shallow aquifer of the volcanohydrothermal system during the recent crises at Vulcano Island (Aeolian Archipelago, Italy). Journal of volcanology and geothermal research, 273, 70-80. Capasso, G., Inguaggiato, S., 1998. A simple method for the determination of dissolved gases in natural waters. An application to thermal waters from Vulcano Island. Appl. Geochem. 13, 631-642. Caracausi, A., Ditta, M., Italiano, F., Longo, M., Nuccio, P.M., Paonita, A., Rizzo, A., (2005a) Changes in fluid geochemistry and physico-chemical conditions of geothermal systems caused by magmatic input: the recent abrupt outgassing off the island of Panarea (Aeolian Islands, Italy). Geochim. Cosmochim. Acta 69, 3045–3059.

ACCEPTED MANUSCRIPT Caracausi, A., Ditta, M., Italiano, F., Longo, M., Nuccio, P.M., Paonita, A., (2005b) Massive submarine gas output during the volcanic unrest off Panarea Island (Aeolian Arc, Italy): Inferences for explosive conditions. Geochem. J., 39, 459-467. Carapezza, M., Dongarrà, G., Hauser, S., Longinelli, A., 1983. Preliminary isotopic investigations on thermal waters from Vulcano Island, Italy. Miner. Petrogr. Acta 27: 221–232. Carapezza, M., Nuccio, P.M., and Valenza, M. (1980), Geochemical precursor of earthquakes. In: High Pressure Science & Technology. (Eds. Vodar and Marteau) (Pergamon, Oxford 1980) pp. 90-103.

PT

Carapezza M., Nuccio P.M., Valenza M. (1981) Genesis and evolution of the fumaroles of Vulcano (Aeolian Islands, Italy): A geochemical model. Bull. Volcanol., 44, 547-563.

RI

Carapezza, M., Badalamenti B., Valenza M. (1984) Geochemical surveillance of the Aeolian Islands by a radio-linked computerized continuous monitoring, Codata Bulletin, Paris.

NU

SC

Carapezza M. L., Diliberto I. S. (1993) Helium and CO2 soil degassing. Data Related to Eruptive Activity, Unrest Phenomena, and Other Observation on the Italian Active Volcanoes in 1991. Vulcano and Stromboli. Acta Vulcanol., 3, 273-276.

MA

Carapezza, M. L., & Federico, C. (2000). The contribution of fluid geochemistry to the volcano monitoring of Stromboli. Journal of Volcanology and Geothermal Research, 95(1), 227-245.Carapezza, M. L., S. Inguaggiato, L. Brusca, and M. Longo (2004), Geochemical precursors of the activity of an open-conduit volcano: The Stromboli 2002-2003 eruptive events, Geophys. Res. Lett., L07620, doi:10.1029/2004GL019614.

PT E

D

Cardellini, C., G. Chiodini, and F. Frondini (2003), Application of stochastic simulation to CO2 flux from soil: Mapping and quantification of gas release, J. Geophys. Res., 108, 2425, doi:10.1029/2002JB002165, B9.

CE

Chiodini, G., Liccioli, C., Vaselli, O., Calabrese, S., Tassi, F., Caliro, S., ... & D'Alessandro, W. (2014). The Domuyo volcanic system: An enormous geothermal resource in Argentine Patagonia. Journal of volcanology and geothermal research, 274, 71-77.

AC

Chiodini, G., Allard, P., Caliro, S., Parello, F. (2000) 18O exchange between steam and carbon dioxide in volcanic and hydrothermal gases: Implications for the source of water. Geochim. Cosmochim. Acta, 64, 2479-2488, doi: 10.1016/S0016-7037(99)00445-7. Chiodini, G., Caliro, S., Caramanna, G., Granieri, D., Minopoli, C., Moretti, R., Perotta, L., Ventura, G., (2006) Geochemistry of the submarine gaseous emissions of Panarea (Aeolian Islands, southern Italy): magmatic vs. hydrothermal origin and implications for volcanic surveillance. Pure Appl. Geophys. 163, 759–780. Chiodini, G., Frondini, F., Cardellini, C., Granieri, D., Marini, L., Ventura, G., 2001. CO2 degassing and energy release at Solfatara volcano, Campi Flegrei, Italy. J. Geophys. Res. 106, 16213^16221. Chiodini, G., Cioni, R., Guidi, M., Marini, L., Panichi, C., Raco, B., Taddeucci, G, 1996. Geochemical surveillance at Vulcano island from 1993 to 1995. Acta Vulcanol. 8: 193-197.

ACCEPTED MANUSCRIPT Chiodini G., Cioni R., Marini L., and Panichi C. (1995) Origin of fumarolic fluids of Vulcano Island, Italy and implications for volcanic surveillance. Bull. Volcanol., 57, 99-110, doi: 10.1007/BF00301400. Chiodini G., Cioni R., Falsaperla S., Montalto A., Marini L., Guidi M. (1992) Geochemical and seismological investigations at Vulcano (Aeolian Islands) during 1978-1989. J. Geophys. Res., 97, 11025-11032, doi: 10.1029/92JB00518.

PT

Chiodini G., Cioni R., and Marini L. (1993) Reactions governing the chemistry of crater fumaroles from Vulcano Island, Italy, and implications for volcanic surveillance. Appl. Geochem., 8, 357-371, doi: 10.1016/0883-2927(93)90004-Z.

RI

Chiodini, G., Marini, L., 1998. Hydrothermal gas equilibria: the H2O–H2–CO2–CO–CH4 system. Geochim. Cosmochim. Acta 62, 2673–2687. http://dx.doi.org/10.1016/S0016-7037(98)00181-1.

NU

SC

Chiodini, G., D. Granieri, R. Avino, S. Caliro, A. Costa, and C. Werner (2005), Carbon dioxide diffuse degassing and estimation of heat release from volcanic and hydrothermal systems, J. Geophys. Res., 110, B08204, doi:10.1029/2004JB003542.

MA

Christopher T., Edmonds M., Taisne B., Odbert H., Costa, A. Hards V. and Wadge G. Periodic sulphur dioxide degassing from the Soufrière Hills Volcano related to deep magma supply.(2014). Geological Society, London, Special Publications, 410, 123-141, 5 September 2014,https://doi.org/10.1144/SP410.11

D

Cioni R., D’Amore F. (1984) A genetic model for the crater fumaroles of Vulcano Island (Sicily, Italy). Geothermics, 13, 375-384, doi: 10.1016/0375-6505(84)90051-8.

PT E

Clocchiatti R., Del Moro A., Gioncada A., Joron J.L., Mosbah M., Pinarelli L., Sbrana A. (1994a) Assessment of a shallow magmatic system: the 1888-90 eruption, Vulcano Island, Italy. Bull. Volcanol., 56, 466-486, .

CE

Clocchiatti R., Gioncada A., Mosbah M., Sbrana A. (1994b) Possibile deep origin of sulfur output at Vulcano (Southern Italy) in the light of melt inclusion studies. Acta Vulcanol., 5, 49-53.

AC

Cortecci, G., Ferrara, G., Dinelli, E. (1996) Isotopic time-variations and variety of sources for sulphur in fumaroles at Vulcano island, Aeolian archipelago, Italy. Acta Vulcanol., 8, 147-160. Cortecci, G., Dinelli, E., Bolognesi, L., Boschetti, T., & Ferrara, G. (2001). Chemical and isotopic compositions of water and dissolved sulfate from shallow wells on Vulcano Island, Aeolian Archipelago, Italy. Geothermics, 30(1), 69-91. Darcy, H. (1856). Les fontaines publiques de la ville de Dijon. Paris: Dalmont. De Astis G, La Volpe L, Peccerillo A, Civetta L (1997) Evoluzione vulcanologica e magmatologica dell’isola di Vulcano. In Progetto Vulcano: Risultati dell’attività di Ricerca 1993-95 (eds. La Volpe L., Dellino P., Nuccio PM., Privitera E., Sbrana A.), Felici Editore, Pisa, pp 155-177. De Astis, G., Lucchi, F., Dellino, P., La Volpe, L., Tranne, C.A., Frezzotti, M.L., Peccerillo, A., 2013. Geology, volcanic history and petrology of Vulcano (central Aeolian archipelago). In: Lucchi, F., Peccerillo, A., Keller, J., Tranne, C.A., Rossi, P.L. (Eds.), The Aeolian Islands Volcanoes. Geological Society, London, Memoirs 37, pp. 181–349.

ACCEPTED MANUSCRIPT De Astis, G., Ventura, G., Vilardo, G., 2003. Geodynamic significante of the Aeolian volcanism (Southern Tyrrhenian Sea, Italy) in light of structural, seismological, and geochemical data. Tectonics 22 (4), 1040– 1057. Del Pezzo, E., & Martini, M. (1981). Seismic events under Vulcano, Aeolian Islands, Italy. Bulletin of Volcanology, 44(3), 521-525.

PT

Di Liberto V, Nuccio PM, Paonita A (2002) Genesis of chlorine and sulfur in fumarolic emissions at Vulcano Island (Italy): assessment of pH and redox conditions in the hydrothermal system. J. Volcanol. Geotherm. Res., 116, 137-150, doi: 10.1016/S0377-0273(02)00215-9.

SC

RI

Diliberto I.S., Gurrieri S., Valenza M. (2002) Relationships between diffuse CO2 emissions and volcanic activity on the island of Vulcano (Aeolian Islands, Italy) during the period 1984-1994. Bull. Volcanol., 64, 219-228, doi: 10.1007/s00445-001-0198-6. Diliberto 2007 INGV-DPC Project V3_5 – Vulcano Section 2. Report from individual Research Units, p.1-7

NU

Diliberto I.S. (2011) Long-term variations of fumaroles temperatures on Vulcano Island (Italy). Ann. Geophys., 54, 175-185, doi: 10.4401/ag-5183.

MA

Diliberto I.S. (2013) Time series analysis of high temperature fumaroles monitored on the island of Vulcano (Aeolian Archipelago, italy). J. Volcanol. Geotherm. Res., 264, 150-163.

PT E

D

Diliberto, I.S. (2017) Long-term monitoring on a close conduit volcano. A 25 Years long time-series of temperatures recorded at La Fossa Cone (Vulcano Island, Italy), ranging from 250 °C up to 520 °C. VHSTaran special issue of Journal of Volcanology and Geothermal Research 10.1016/j.jvolgeores.2017.03.005

CE

Diliberto I.S., Gagliano Candela E, Morici S., Pecoraino G., Bellomo S., Bitetto M., Longo M. Article title: Changes in heat release by hydrothermal circulation monitored during an eruptive cycle on Mount Etna (Italy) Bulletin of Volcanology

AC

Di Martino, R. , Camarda, M., Gurrieri, S., Valenza, M. (2013) Continuous monitoring of hydrogen and carbon dioxide at Mt Etna - Chemical Geology 357(357):41-51 Di Napoli, R., et al. (2016), Hydrothermal fluid venting in the offshore sector of Campi Flegrei caldera: A geochemical, geophysical, and volcanological study, Geochem. Geophys. Geosyst., 17, doi:10.1002/2016GC006494. Dongarrà, G., Hauser, S., Capasso, G., & Favara, R. (1988). Characteristics of variations in water chemistry of some wells from Vulcano island. Rend Soc It Mineral Petrol, 43, 1123-1131. Edmonds M., Herd R. A., Galle B. (2003) Automated, high time-resolution measurements of SO2 flux at Soufrière Hills Volcano, Montserrat Bulletin of Volcanology, 2003, Volume 65, Number 8, Page 578, Edmonds M.; Oppenheimer C., Pyle D.M., Herd R.A., Thompson G. (2003) SO2 emissions from Soufrie're Hills Volcano and their relationship to conduit permeability, hydrothermal interaction and degassing regime Journal of Volcanology and Geothermal Research, Volume 124, Issue 1, 2003, Pages 23-43, ISSN 0377-0273, http://dx.doi.org/10.1016/S0377-0273(03)00041-6.

ACCEPTED MANUSCRIPT Edmonds, M., Pyle, D. & Oppenheimer,C., 2002: HCl emissions at Soufrière Hills Volcano, Montserrat, West Indies, during a second phase of dome building: November 1999 to October 2000.- Bul. Volcanol. 64: 21-30. Edmonds M, Oppenheimer C (2001) A model for degassing at the Soufrière Hills Volcano, Montserrat, West Indies, based on geochemical data. 186:159-173

PT

Edner, H., S. Ragnarson, S. Svanberg, E. Wallinder, R. Ferrera, R. Cioni, B. Raco, and G. Taddeucci, Total fluxes of sulfur dioxide from the Italian Volcanoes Etna, Stromboli and Vulcano measured by differential absorption lidar and passive differential optical absorption spectroscopy, J. Geophys. Res.,99, 18,827– 18,838, 1994.

SC

RI

Favalli, M., Karátson, D., Mazzuoli., R., Pareschi, M.T., Ventura, G. (2005) Volcanic geomorphology and tectonics of the Aeolian archipelago (southern Italy) based on integrated DEM data. Bull. Volcanol. 68, 157–170.

MA

NU

Favara Rocco, Salvatore Giammanco, Salvatore Inguaggiato, Giovannella Pecoraino,(2001) Preliminary estimate of CO output from Pantelleria Island volcano (Sicily, Italy): evidence of active mantle degassing, Applied Geochemistry, Volume 16, Issue 7, 2001, Pages 883-894, ISSN 0883-2927 http://dx.doi.org/10.1016/S0883-2927(00)00055-X.

PT E

D

Federico, C., Madonia, P., Cusano, P., & Petrosino, S. (2013). Groundwater geochemistry of the Mt. Vesuvius area: implications for volcano surveillance and relationship with hydrological and seismic signals. Annals of Geophysics, 56(4), S0447. doi: http://dx.doi.org/10.4401/ag-6453

CE

Federico, C., Capasso, G., Paonita, A., & Favara, R. (2010). Effects of steam-heating processes on a stratified volcanic aquifer: Stable isotopes and dissolved gases in thermal waters of Vulcano Island (Aeolian archipelago). Journal of Volcanology and Geothermal Research, 192(3), 178-190.

AC

Federico C. Brusca L., Carapezza M. L., Cigolini, C. Inguaggiato S., Rizzo A. and Rouwet D. (2008) Geochemical Prediction of the 2002–2003 Stromboli Eruption From Variations in CO2 and Rn Emissions and in Helium and Carbon Isotopes. American Geophysical Union.10.1029/182GM1 DOI: 10.1029/182GM11 Federico Cinzia , Alessandro Aiuppa, Rocco Favara, Sergio Gurrieri, Mariano Valenza,(2004).Geochemical monitoring of groundwaters (1998–2001) at Vesuvius volcano (Italy), Journal of Volcanology and Geothermal Research, Volume 133, Issue 1, 2004, Pages 81-104, ISSN 0377-0273, http://dx.doi.org/10.1016/S0377-0273(03)00392-5. Fick, A. (1855). Poggendorff's Annalen der Physik und Chemie. 94, 59-86.

Frazzetta G., La Volpe L. (1991) Volcanic history and maximum expected eruption at ‘‘La Fossa di Vulcano’’ (Aeolian Island, Italy). Acta Vulcanol., 1, 107- 113.

ACCEPTED MANUSCRIPT Fulignati, P., Gioncada, A., Sbrana, A., 1996. Hydrothermal alteration in the subsoil of Porto di Levante, Vulcano (Aeolian Islands, Italy). Acta Vulcanol. 8 (2): 129-138. Gabbianelli G., Gillot P. V., Lanzafame G., Ligi M., Postpischl D. and Rossi P. L. (1986) Controllo strutturale nell’evoluzione vulcanica di Panarea (Isole Eolie). Italian National Res. Council Internal Report, pubbl. # 239, Catania, Italy. Gabbianelli G., Gillot P. Y., Lanzafame G., Romagnoli C. and Rossi P. L. (1990) Tectonic and volcanic evolution of Panarea (Aeolian Islands, Italy). Mar. Geol. 92, 313–326.

PT

Galle B, Oppenheimer CM (2003) Automated, high time-resolution measurements of SO2 flux at Soufrière Hills Volcano, Montserrat. Bull Volcanol 65:578-586

SC

RI

Galle B, Oppenheimer C, Geyer A, McGonigle AJS, Edmonds M, Horrocks LA (2003) A miniaturised ultraviolet spectrometer for remote sensing of SO2 fluxes: A new tool for volcano surveillance. J. Volcanol. Geotherm. Res. 119:241– 254

MA

NU

Galle, B., Johansson, M. Rivera, C., Zhang, Y., Kihlman, M., Kern, C., Lehmann, T., Platt, U., Arellano, S., and Hidalgo, S. (2010) Network for Observation of Volcanic and Atmospheric Change (NOVAC)—A global network for volcanic gas monitoring: Network layout and instrument description. J. Geophys. Res., 115, D05304, doi:10.1029/2009JD011823, 2010

PT E

D

Gambino S., Guglielmino F. (2008) Ground deformation induced by geothermal processes: a model for La Fossa Crater (Vulcano Island, Italy). J. Geophys. Res., 113, B07402, doi: 10.1029/2007JB005016.

Gerlach TM, Nordlie BE (1975) The C-O-H-S gaseous systems. Part II: temperature, atomic composition, and molecular equilibria in volcanic gases. Am J Sci 275:377-394.

CE

Gerlach TM, Casadeval TJ., (1986). Evaluation of gas data from high-temperature fumaroles at Mount St. Helens, 1980-1982. J. Volcanol. Geoth. Res. 28:107-140

AC

Gerlach TM (1993) Oxigen buffering of Kilauea volcanic gases and the oxygen fugacity of Kilauea basalt. Geochim Cosmochim Acta 57:795-814 Giammanco S., S. Inguaggiato, and M. Valenza. (1998) Soil and fumarole gases of Mount Etna: geochemistry and relations with volcanic activity. J.Volcanol.Geoth.Res. 81 (3-4):297-310,. Giggenbach W.F. (1975) A simple method for the collection and analysis of volcanic gas samples. Bull. Volcanol., 39, 132-145. Giggenbach W.F. (1980) Geothermal gas equilibria. Geochim Cosmochim Acta 44:2021-2032. Giggenbach, W.F. (1996), Chemical composition of volcanic gases, in Monitoring and Mitigation of Volcano Hazards, edited by R. Scarpa and R.I. Tilling, Springer-Verlag, Berlin, pp 221–256.

ACCEPTED MANUSCRIPT Gioncada A., Mazzuoli R., Bisson M., Pareschi MT. (2003) Petrology of volcanic products younger than 42 ka on the Lipari-Volcano complex (Aeolian Islands, Italy): an example of volcanism controlled by tectonics. J Volcanol. Geotherm. Res., 122, 191-220, doi: 10.1016/S0377-0273(02)00502-4.

PT

Granieri, D., Carapezza M.L., Chiodini G., Avino R., Caliro S., Ranaldi M., Ricci T., Tarchini L. (2006) 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? Geophys. Res. Lett., 33, L13316, doi: 10.1029/2006GL026460.

RI

Grassa, F., Inguaggiato, S., & Liotta, M. (2008). Fluid geochemistry of Stromboli. The Stromboli Volcano: An Integrated Study of the 2002-2003 Eruption, 49-63.

SC

Grutter M, Flores E, Basaldud R, Ruiz-Suárez LG (2003) Open-path FTIR spectroscopic studies of the trace gases over Mexixo City. Atmos. Oceanic Opt. 16(3): 232-236.

MA

NU

Harris A., Alparone S., Bonforte A., Dehn J., Gambino S., Lodato L., Spampinato L. (2012) Vent temperature trends at the Vulcano Fossa fumarole field: the role of permeability. Bull. Volcanol. doi: 10.1007/s00445-012-0593-1.

D

Heinicke, J., Italiano, F., Maugeri, R., Merkel, B., Pohl, T., Schipek, M., Braun, T., (2009) Evidence of tectonic control on active arc volcanism: the Panarea–Stromboli tectonic link inferred by submarine hydrothermal vents monitoring (Aeolian arc, Italy). Geophys. Res. Lett. 36, L04301.

PT E

Inguaggiato, S., Vita F., Cangemi M., Mazot A., Sollami A., Calderone L., Morici S., Jacome Paz M. P. (2017) Stromboli volcanic activity variations inferred by fluids geochemistry observations: sixteen years of continuous soil CO2 fluxes monitoring (2000-2015) Chemical Geology (doi: 10.1016/j.chemgeo.2017.01.030)

CE

Inguaggiato S., Londoño J. M., Chacón Z., Liotta M., Gil E., Diego Alzate (2016) The hydrothermal system of Cerro Machín volcano (Colombia): New magmatic signals observed during 2011–2013 doi: org/10.1016/j.chemgeo.2016.12.020

AC

Inguaggiato, C., Vita, F., Diliberto, I.S., Calderone, L. (2016) The role of the aquifer in soil CO 2 degassing in volcanic peripheral areas: A case study of Stromboli Island (Italy). Chemical geology, doi.org/10.1016/j.chemgeo.2016.12.017 Inguaggiato S., M.P. Jacome Paz, A. Mazot, H. Delgado Granados, C. Inguaggiato, F. Vita (2013) CO2 output discharged from Stromboli Island (Italy). Chemical Geology (2013) http://dx.doi.org/10.1016/j.chemgeo.2012.10.008 Inguaggiato, S., Mazot, A., Diliberto, I.S., Inguaggiato, C., Madonia, P., Rouwet, D., Vita, F. 2012a. Total CO2 output from Vulcano island (Aeolian Islands, Italy). Geochemistry, Geophysics, Geosystems 13 (2). http://dx.doi.org/10.1029/2011GC003920 (Q02012).

ACCEPTED MANUSCRIPT Inguaggiato S., Calderone L., Inguaggiato C., Mazot A., Morici S., Vita F. (2012b) Long-time variation of soil CO2 fluxes at summit crater of Vulcano (Italy). Bull. Volcanol., 74, 1859-1863, doi: 10.1007/s00445-0120637-6. Inguaggiato S, Vita F., Rouwet D., Bobrowski N., Morici S., Sollami A. (2011a) Geochemical evidence of the renewal of volcanic activity inferred from CO2 soil and SO2 plume fluxes: the 2007 Stromboli eruption (Italy) Bull Volcanol DOI: 10.1007/s00445-010-0442-z

PT

Inguaggiato S., Mazot A., Ohba T. (2011b) Monitoring of active volcanoes: The geochemical approach Annals of Geoph. Vol. 54, 2, 2011 doi: 10.4401/ag-5187

RI

Inguaggiato S, Shinohara Hiroshi and Fisher T. (2011c) Geochemistry of volcanic fluids Bullettin of Volcanology (2011), Vol 73, Nuber 4, Pages 3693-371

SC

Inguaggiato, S., Martin-Del Pozzo, A. L., Aguayo, A., Capasso, G., and Favara, R., 2005, Isotopic, chemical and dissolved gas constraints on spring water from Popocatepetl (Mexico): evidence of gas-water interaction magmatic component and shallow fluids: J.Volcanol.Geoth.Res., v. 141, p. 91-108.

NU

Inguaggiato, S., Taran, Y. A., Grassa, F., Capasso, G., Favara, R., Varley, N., and Faber, E., 2004a, Nitrogen isotopes in thermal fluids of a forearc region (Jalisco Block, Mexico): evidence for heavy nitrogen from continental crust: Geochemistry, Geophysics, Geosystems, v. 5, p. Q12003.

MA

Inguaggiato, S. and Rizzo, A., 2004b, Dissolved helium isotope ratios in ground-waters: a new technique based on gas-water re-equilibration and its application to Stromboli volcanic system: Applied Geochem., v. 19, p. 665-673.

PT E

D

Inguaggiato S., G. Pecoraino, and F. D'Amore. (2000) Chemical and isotopical characterization of fluid manifestations of Ischia Island (Italy). J.Volcanol.Geoth.Res. 99:151-178. Italiano F. and Nuccio P. M. (1991) Geochemical investigations of submarine exhalations to the east of Panarea, Aeolian Islands, Italy. J. Volcanol. Geotherm. Res. 46, 125–141.

CE

Italiano F., Nuccio P.M. (1997) Variazione del rapporto isotopico 3He/4He nelle esalazioni fumaroliche di Vulcano. In Progetto Vulcano: Risultati dell’Attività di Ricerca 1993-95 (eds. La Volpe L., Dellino P., Nuccio P.M., Privitera E., Sbrana A.). Felici Editore, Pisa, pp. 124-127.

AC

Italiano, F., G. Pecoraino, and P. M. Nuccio (1998) Steam output from fumaroles of an active volcano: Tectonic and magmatic-hydrothermal controls on the degassing system at Vulcano (Aeolian arc). J. Geophys. Res., 103, 29829-29842, doi: 10.1029/98JB02237. Liotta M., A. Paonita, A. Caracausi, M. Martelli, A. Rizzo, R. Favara (2010) Hydrothermal processes governing the geochemistry of the crater fumaroles at Mount Etna volcano (Italy). Chemical Geology 278 (2010) 92–104 doi:10.1016/j.chemgeo.2010.09.004 Liotta, M., L. Brusca, F. Grassa, S. Inguaggiato, M. Longo, and P. Madonia (2006), Geochemistry of rainfall at Stromboli volcano (Aeolian Islands): Isotopic composition and plume-rain interaction, Geochem. Geophys. Geosyst., 7, Q07006, doi:10.1029/2006GC001288. Louban, I., Bobrowski, N., Rouwet, D., Inguaggiato, S., & Platt, U. (2009). Imaging DOAS for volcanological applications, B. Volcanol., 71, 753–765.

ACCEPTED MANUSCRIPT

Johansson M, Galle B, Rivera C, Zhang Y (2009) Tomographic Reconstruction of Gas Plumes Using Scanning DOAS. submitted to Bull. Volcanol. DO7s00445-009-0292-8 Kern, C., Sihler, H., Vogel, L., Rivera, C., Herrera, M., Platt, U. (2009b) Halogen oxide measurements at Masaya Volcano, Nicaragua using active long path differential optical absorption spectroscopy. Bull. Volcanol. 71:659-670

PT

Kern, C., Detschmann, T., Vogel, L., Wohrbach, M., Wagner, T., Platt, U. (2009a) Radiative transfer corrections for accurate spectroscopic measuremens of volcanic gas emissions. Bull. Volcanol 72:233247

RI

Kern C (2009) Spectroscopic measurements of volcanic gas emissions in the ultra-violet wavelength region. Ph.D. thesis, University of Heidelberg, Heidelberg

SC

Kern C., Trick S., Rippel B., Platt U. (2006) Applicability of light-emitting diodes as light sources for active differential optical absorption spectroscopy measurements. Appl. Opt. 45:2077-2088

MA

NU

Leeman WP., Tonarini S., Pennisi M., Ferrara G. (2005) Boron isotope variations in fumarolic condensates and thermal waters from Vulcano Island, Italy: implications for evolution of volcanic fluids. Geochim. Cosmochim. Acta, 69, 143-163, doi: 10.1016/j.gca.2004.04.004. Le Guern, F., Gerlach TM, Nohl A., (1982). Field gas chromatograph analysis of gases from a glowing dome at Merapi volcano, Java, Indonesia, 1977,1978,1979, J Volcanol. Geotherm. Res. 14: 223-245.

PT E

D

Madonia, P., Capasso, G., Favara, R., Francofonte, S., & Tommasi, P. (2015). Spatial Distribution of Field Physico-Chemical Parameters in the Vulcano Island (Italy) Coastal Aquifer: Volcanological and Hydrogeological Implications. Water, 7(7), 3206-3224.

CE

Madonia, P., Cusano, P., Diliberto, I. S., & Cangemi, M. (2013). Thermal anomalies in fumaroles at Vulcano island (Italy) and their relationship with seismic activity. Physics and Chemistry of the Earth, Parts A/B/C, 63, 160-169.

AC

Madonia P., Pecoraino G., Sammarco C. (2011) Georeferenced cartography dataset of the La Fossa crater fumarolic field at Vulcano Island (Aeolian Archipelago, Italy): conversion and comparison of data from local to global positioning methods. Ann. Geophys.-Italy, 54, 3; doi: 10.4401/ag-4941. Madonia, P., Brusca, L., Inguaggiato, S., Longo, M. and Morici, S. (2008) Variations of Soil Temperature, CO2 Flux, and Meteorological Parameters, in The Stromboli Volcano: An Integrated Study of the 2002-2003 Eruption (eds S. Calvari, S. Inguaggiato, G. Puglisi, M. Ripepe and M. Rosi), American Geophysical Union, Washington, D. C.. doi: 10.1029/182GM22 Magro G. (1997) La composizione isotopica dell’He e dell’Ar nei prodotti di Vulcano: Inclusioni fluide e gas fumarolici. In Progetto Vulcano: Risultati dell’Attività di Ricerca 1993-95 (eds. La Volpe L., Dellino P., Nuccio P.M., Privitera E., Sbrana A.), Felici Editore, Pisa, pp. 128-132. Mandarano M., Paonita A., Martelli M., Viccaro M., Nicotra E., Millar I.L. (2016) Revealing magma degassing below closed-conduit active volcanoes: Geochemical features of volcanic rocks versus fumarolic fluids at Vulcano (Aeolian Islands, Italy). Lithos, 248-251, 272-287.

ACCEPTED MANUSCRIPT M. Martelli, A.L. Rizzo, A. Renzulli, F. Ridolfi, I. Arienzo, A. Rosciglione (2014) Noble-gas signature of magmas from a heterogeneous mantle wedge: The case of Stromboli volcano (Aeolian Islands, Italy). Chemical Geology, Volume 368, 2014, Pages 39-53, ISSN 0009-2541, http://dx.doi.org/10.1016/j.chemgeo.2014.01.003.

PT

Martelli, M., Nuccio, P.M., Stuart, F.M., Di Liberto, V., Ellam, R.M., (2008) Constraints on Mantle Source and Interactions from He–Sr Isotope Variation in Italian Plio-Quaternary Volcanism (Q02001. doi:10.1029/2007GC001730 ISSN: 1525–2027).

SC

RI

McGonigle, A. J. S.; Inguaggiato, S.; Aiuppa, A.; Hayes, A. R.; Oppenheimer,C. Accurate measurement of volcanic SO2 flux: Determination of plume transport speed and integrated SO2 concentration with a single device. (2005) Geochem. Geophys. Geosys doi:10.1029/2004GC000845 ISSN: 1525-2027

NU

McGonigle, A. J. S.; Oppenheimer, C.; Hayes, A. R.; Galle, B.; Edmonds, M.; Caltabiano, T.; Salerno, G.; Burton, M.; Mather, T. A. Sulphur dioxide fluxes from Mount Etna, Vulcano, and Stromboli measured with an automated scanning ultraviolet spectrometer. (2003) J. Geophys. Res, VOL. 108, NO. B9, 2455, doi:10.1029/2002JB002261

MA

McGonigle, A.J.S.; Aiuppa, A.; Ripepe, M.; Kantzas, E.P.; Tamburello, G. Spectroscopic capture of 1 Hz volcanic SO2 fluxes and integration with volcano geophysical data. Geophys. Res. Lett. 2009, 36, L21309, doi:10.1029/2009GL040494.

PT E

D

McGonigle, A.J.S., Aiuppa, A., Giudice, G., Tamburello, G., A. J., Hodson, Gurrieri S. (2008) Unmanned aerial vehicle measurements of volcanic carbon dioxide fluxes. Geophysical research letters, vol. 35, L06303

CE

Métrich, N., Bertagnini, A., Landi, P., Rosi, M., Belhadj, O., 2005. Triggering mechanism at the origin of paroxysms at Stromboli (Aeolian Archipelago, Italy): The 5 April 2003 eruption. Geophys. Res. Lett., 32, L10305, doi:10.1029/2004GL022257. Millan, M.M., Hoff, R.M. (1978) Remote sensing of air pollutant by correlation spectroscopy instrumental response characteristics. Atmospher Environm 12:853-864

AC

Milluzzo V., Cannata A., Alparone S., Gambino S., Helleweg M., Montalto P., Cammarata L., Diliberto I.S., Gresta S., Liotta M., Paonita A. (2010) Tornillos at Vulcano: clues to the dynamics of the hydrothermal system. J. Volcanol. Geotherm. Res., 198, 377-393, doi: 10.1016/j.jvolgeores.2010.09.022.

Mori T., Notsu K., Tohjima Y., Wakita H., Nuccio P.M. and Italiano F. (1995) Remote detection of fumarolic gas chemistry at Vulcano, Italy, using an FT-IR spectral radiometer. Earth and Planet. Sci. Lett., 134, 219224 Mori T., Notsu K (2008) Temporal variation in chemical composition of the volcanic plume from Aso volcano, Japan, measured by remote FT-IR spectroscopy. Geochem. J. 42: 133– 140. Mori T, Notsu K (1997) Remote CO, COS, CO2, SO2, HCl detection and temperature estimation of volcanic gas. Geophys. Res. Lett. 24: 2047–2050.

ACCEPTED MANUSCRIPT Mori T, Notsu K, Tohjima Y, Wakita H (1993) Remote detection of HCl and SO2 in volcanic gas from Unzen volcano, Japan. Geophys. Res. Lett. 20: 1355–1358. Mori T.; Burton M. Quantification of the gas mass emitted during single explosions on Stromboli with the SO2 imaging camera. J. Volcanol. Geotherm. Res. 2009, 188, 395–400, doi:10.1016/j.jvolgeores.2009.10.005.

PT

Morner, N.A. and G. Etiope (2002): Carbon degassing from the lithosphere, Global Planet. Change, 33 (1/2), 185-203

RI

Notsu K, Mori T (2010) Chemical monitoring of volcanic gas using remote FT-IR spectroscopy at several active volcanoes in Japan. Applied Geochemistry 25: 505–512.

SC

Nuccio, P.M., Paonita, A., Sortino, F. (1999) Geochemical modeling of mixing between magmatic and hydrothermal gases: The case of Vulcano, Italy. Earth Planet. Sci. Lett. 167, 321-333, doi: 10.1016/S0012-821X(99)00037-0.

MA

NU

Nuccio P. M., Paonita A. (2000) Investigation of the He solubility in H2O-CO2-bearing silicate liquids at moderate pressure II: the Extended Ionic Porosity (EIP) model. Earth Planet. Sci. Lett. 183, 499-512, doi: 10.1016/S0012-821X(00)00215-6.

D

Nuccio, P.M., Paonita, A. (2001) Magmatic degassing of multicomponent vapours and assessment of magma depth: Application to Vulcano Island (Italy). Earth Planet. Sci. Lett., 193, 467-481, doi: 10.1016/S0012-821X(01)00512-X.

PT E

Nuccio P. M. and Valenza M. (1998) Magma degassing and geochemical detection of its ascent. In WaterRock Interaction WRI-9 (eds. G. B. Arehart and J. R. Hulston), pp. 471–474. Balkema, Rotterdam, the Netherlands.

CE

Oppenheimer, C., Francis, P., Burton, M., Maciejewski, A.J.H. & Boardman, L., 1998: Remote measurement of volcanic gases by Fourier transform infrared spectroscopy.- Applied Physics. B, Lasers and Optics, 67(4): 505–515.

AC

Oppenheimer, C., Scaillet, B. & Martin, R., 2011: Sulfur Degassing from Volcanoes: Source Conditions, Surveillance, Plume Chemistry and Earth System Impacts.- Rev. Mineral. Geochem 73: 363-421. Oskarsson, N. (1984) Monitoring of fumarole discharge during the 1975–1982 rifting in Krafla volcanic center, North Iceland. J. Volcanol. Geoth. Res. 22, 97–121. O’Dwyer. M., Padgett, M. J., McGonigle, A. J. S., Oppenheimer, C., Inguaggiato, S., (2003) Real- time measurement of volcanic H2S, and SO2 concentrations by UV spectroscopy. Geophysical research letters, vol 30, NO. 12, 1652 Panichi C., Noto P. (1992) Isotopic and chemical composition of water, steam and gas samples of the natural manifestations of the Island of Vulcano (Aeolian Arc, Italy). Acta Vulcanol., 2, 297-312. Paonita A., Gigli G., Gozzi D., Nuccio P.M., Trigila R. (2000) Investigation of He solubility in H2O-CO2 bearing silicate liquids at moderate pressure: an experimental method. Earth Planet. Sci. Lett., 181, 595-604, doi: 10.1016/S0012-821X(00)00215-6.

ACCEPTED MANUSCRIPT Paonita A., Favara R., Nuccio P.M., Sortino F. (2002) Genesis of fumarolic emissions as inferred by isotope mass balances: CO2 and water at Vulcano Island, Italy. Geochim. Cosmochim. Acta, 66(5), 759-772, doi: 10.1016/S0016-7037(01)00814-6. Paonita, A., Federico, C., Bonfanti, P., Capasso, G., Inguaggiato, S., Italiano, F., Madonia, P., Pecoraino, G., Sortino, F. (2013) The episodic and abrupt geochemical changes at La Fossa fumaroles (Vulcano Island, Italy) and related constraints on the dynamics, structure, and compositions of the magmatic system. Geochim. Cosmochim. Acta 120, 158–178.

PT

Paonita A., Caracausi A., Iacono-Marziano G., Martelli M., A. Rizzo (2012) Geochemical evidence for mixing between fluids exsolved at different depths in the magmatic system of Mt Etna (Italy). Geochim. Cosmochim. Acta, 84, 380-394, doi: 10.1016/j.gca.2012.01.028.

SC

RI

Papale P., Moretti R., Barbato D. (2006) The compositional dependence of the saturation surface of H2O+CO2 fluids in silicate melts. Chem. Geol., 229, 78-95, doi: 10.1016/j.chemgeo.2006.01.013.

NU

Peccerillo A., Frezzotti M.L., De Astis G., Ventura G. (2006) Modeling the magma plumbing system of Vulcano (Aeolian Islands, Italy) by integrated fluid-inclusion geobarometry, petrology, and geophysics. Geol., 34(1), 17-20, doi: 10.0030/G22117.1.

MA

Pecoraino, G., Brusca, L., D'alessandro, W., Giammanco, S., Inguaggiato, S., & Longo, M. (2005). Total CO2 output from Ischia Island volcano (Italy). Geochemical Journal, 39(5), 451-458.

D

Pering, T.D.; Tamburello, G.; McGonigle, A.J.S.; Aiuppa, A.; Cannata, A.; Giudice, G.; Patan , D. High time resolution fluctuations in volcanic carbon dioxide degassing from Mount Etna. J. Volcanol. Geotherm. Res.2014, 270, 115–121, doi:10.1016/j.jvolgeores.2013.11.014.

PT E

Pinardi G., Gaia Pinardi, R. Campion, M. Van Roozendael, C. Fayt, J. van Geffen, B. Galle,S. Carn, P. Valks, M. Rix, S. Hildago, J. Bourquin, G. Garzon, S. Inguaggiato, F. Vita (2011) COMPARISON OF VOLCANIC SO2 FLUX MEASUREMENTS FROM SATELLITE AND FROM THE NOVAC NETWORK in Final Report of NOVAC Project. Editors: Bo Galle, Hugo del Gado, Gustavo Garzon, Laif Vogel and Ulrich Platt.

AC

CE

Piochi M., De Astis G., Petrelli M., Ventura G., Sulpizio R., Zanetti A. (2009) Constraining the recent plumbing system of Vulcano (Aeolian Arc, Italy) by textural, petrological, and fractal analysis: The 1739 A.D. Pietre Cotte lava flow. Geochem. Geophys. Geosys., 10(1), doi:10.1029/2008GC002176. Platt, U., and D. Perner, "Measurements of atmospheric trace gases by long path differential UV/visible absorption spectroscopy", in Optical and Laser Remote Sensing, edited by D.A. Killinger, and A. Mooradien, pp. 95-105, Springer Verlag, New York, 1983. Platt, U., "Differential optical absorption spectroscopy (DOAS)", Chem. Anal. Series, 127, 27 - 83, 1994. Platt U, Stutz J (2008) Differential Optical Absorption Spectroscopy - Principles and Applications. Springer, Berlin Heidelberg New York, pp 1-597 Ripepe, M., Marchetti, E., Ulivieri, G., Harris, A., Dehn, J., Burton, M., Caltabiano, T., Salerno, G., 2005. The Effusive to explosive transition during the 2002 eruption of Stromboli volcano. Geology 33 (5), 341– 344.

ACCEPTED MANUSCRIPT Rizzo, A. L., Federico, C., Inguaggiato, S., Sollami, A., Tantillo, M., Vita, F., ... & Liuzzo, M. (2015). The 2014 effusive eruption at Stromboli volcano (Italy): inferences from soil CO2 flux and 3He/4He ratio in thermal waters. Geophysical Research Letters, 42(7), 2235-2243. Rizzo A., Grassa F., Inguaggiato S., Liotta M., Longo M., Madonia P., Brusca L., Capasso G., Morici S., Rouwet D., Vita F. (2008) Geochemical evaluation of the volcanic activity changes observed during 2007 eruption at Stromboli (Aeolian islands, Italy) JVGR, J. Volcanol. Geotherm. Res. 182:246–254

PT

Sano, Y., and H. Wakita (1985) Geographical distribution of 3He/4He in Japan: Implications for arc tectonics and incipient magmatism. J. Geophys. Res., 90, 8729-8741, doi: 10.1029/JB090iB10p08729.

RI

Sano, Y., Wakita, H., Italiano, F., Nuccio, M., (1989) Helium isotopes and tectonics in southern Italy. Geophys. Res. Lett. 511–514.

SC

Sato, M. (1978) Oxygen fugacity of basaltic magmas and the role of gas-forming elements. Geophys. Res. Lett. 5, 447–449.

NU

Sato, M., McGee, K.A. (1982) Continuous monitoring of hydrogen on south flank of Mount St. Helens. In: Lipman, P.W., Mullineaux, D.R. (Eds.), The 1980 eruption of Mount St. Helens. Washington, Geological Survey Professional Paper, 1250, pp. 209–219.

MA

Sawyer, G.M., Oppenheimer, C., Tsanev, V.I. & Yirgu, G., 2008b: Magmatic degassing at Erta ‘Ale volcano, Ethiopia.- J. Volcanol. Geotherm. Res. 178: 837–846.

D

Schöpa, A., Pantaleo, M., Walter, T.R. (2011) Scale-dependent location of hydrothermal vents: Stress field models and infrared field observations on the Fossa Cone, Vulcano Island, Italy. J. Volcanol.Geotherm. Res., 203, 133–145, http://doi:10.1016/j.jvolgeores.2011.03.008

PT E

Shinohara, H., Aiuppa, A., Giudice, G., Gurrieri, S., & Liuzzo, M. (2008). Variation of H2O/CO2 and CO2/SO2 ratios of volcanic gases discharged by continuous degassing of Mount Etna volcano, Italy. Journal of Geophysical Research: Solid Earth, 113(B9).

CE

Sommaruga, C., 1984. Le ricerche geotermiche svolte a Vulcano negli anni '50. Rendiconti della Società Italiana di Mineralogia e Petrologia 39, 355-366.

AC

Sortino F., Inguaggiato S., Francofonte S. (1991) Determination of HF, HCl, and total sulfur in fumarolic fluids by ion chromatography. Acta Vulcanol., 1, 89-91. Stoiber, R.E., Malinconico L.L., Williams S.N. (1983) Use of the correlation spectrometer at volcanoes. In: Tazieff H, Sabroux JC (eds) Forecasting volcanic events. Elsevier, Amsterdam, pp 425-444 Stremme, W., Ortega, I., Siebe C., Grutter M., 2011: Gas composition of Popocatépetl Volcano between 2007 and 2008: FTIR spectroscopic measurements of an explosive event and during quiescent degassing. Earth and Planetary Science Letters 301 (2011) 502–510 Stutz, J., and U. Platt (1996)., Numerical Analysis and Estimation of the Statistical Error of Differential Optical Absorbtion Spectroscopy Measurements with Least-Squares methods, Appl. Opt., 35, (30), 6041-6053

ACCEPTED MANUSCRIPT Tamburello G., Aiuppa, A., Kantzas, E.P., McGonigle, A.J.S., M. Ripepe. (2012) Passive vs. active degassing modes at an open-vent volcano (Stromboli, Italy) Earth and Planetary Science Letters 359–360 (2012) 106– 116 Tamburello, G.; Kantzas, E.P.; McGonigle, A.J.S.; Aiuppa, A.; Giudice G. (2011) UV camera measurements of fumarole field degassing (La Fossa crater, Vulcano Island). J. Volcanol. Geotherm. Res. 2011, 199, 47-52, doi:10.1016/j.jvolgeores.2010.10.004.

PT

Taran, Y. A., & Peiffer, L. (2009). Hydrology, hydrochemistry and geothermal potential of El Chichón volcano-hydrothermal system, Mexico. Geothermics, 38(4), 370-378.

RI

Taran, Y., (2005) A method for determination of the gas–water ratio in bubbling springs. Geophys. Res. Lett. 32, L23403.

SC

Taran Y. (2011) N2, Ar, and He as a tool for discriminating sources of volcanic fluids with application to Vulcano, Italy. Bull. Volcanol., 73, 395-408, doi: 10.1007/s00445-011-0448-1.

NU

Taran YA, Esikov AD, Cheshko AL (1986) Deuterium and Oxygen-18 in waters of Mutnovsky geothermal area. Geochem Int 4:458-468

MA

Taran YA, Fischer TP, Pokrovsky B, Sano Y, Armienta MA, Macías JL (1998) Geochemistry of the volcanohydrothermal system of El Chichón Volcano, Chiapas, Mexico. Bull Volcanol 60:436-449 Taran Y., Pokrovsky BG., Dubik YM. (1989) Isotopic composition and origin of water in andesitic magmas. Dokl. Earth Sci., 304, 440-443.

PT E

D

Tassi, F., Capaccioni, B., Caramanna, G., Cinti, D., Montegrossi, G., Pizzino, L., Quattrocchi, F., Vaselli, O., 2009. Low-pH waters discharging from submarine vents at Panarea Island (Aeolian Islands, southern Italy) after the 2002 gas blast: origin of hydrothermal fluids and implications for volcanic surveillance. Appl. Geochem. 24, 246–254.

CE

Tassi, F., Capaccioni, B., Vaselli O. (2014) Compositional spatial zonation and 2005–2013 temporal evolution of the hydrothermal-magmatic fluids from the submarine fumarolic field at Panarea Island (Aeolian Archipelago, southern Italy). J. Volcanol. Geotherm. Res., 277, 41-50.

AC

Tedesco D. (1995) Fluid geochemistry at Vulcano Island: a change in volcanic regime or fluctuations in the mixing of different systems? J. Geophys. Res. 100, 4157-4167, doi: 10.1029/94JB02595. Tedesco D., Nagao K. (1996) Radiogenic 4He, 21Ne and 40Ar in fumarolic gases on Vulcano: Implication for the presence of continental crust beneath the island. Earth Planet. Sci. Lett., 144, 517-528, doi: 10.1016/S0012-821X(96)00196-3. Tedesco D., Scarsi P. (1999) Intensive gas sampling of noble gases and carbon at Vulcano (Southern Italy). J. Geophys. Res., 104, 10499-10510, doi: 10.1029/1998JB900066. Toshiya Mori, Kenji Notsu, Yasunori Tohjima, Hiroshi Wakita, P. Mario Nuccio, Francesco Italiano. (1995) Remote detection of fumarolic gas chemistry at Vulcano, Italy, using an FT-IR spectral radiometer. Earth and Planetary Science Letters 134 (1995) 219-224

ACCEPTED MANUSCRIPT Ventura, G., 2013. Kinematics of the Aeolian volcanism (Southern Tyrrhenian Sea) from geophysical and geological data. In: Lucchi, F., Peccerillo, A., Keller, J., Tranne, C.A., Rossi, P.L. (Eds.), The Aeolian Islands Volcanoes. Geological Society, London,Memoirs 37, pp. 3–11 doi: http://dx.doi.org/10.1144/M37.2. Ventura, G., Vilardo, G., Milano, G., Pino, N.A., 1999. Relationship among crustal structure, volcanism and strike-slip tectonics in the Lipari-Vulcano volcanic complex (Aeolian Islands, Southern Tyrrhenian Sea, Italy). Phys. Earth Planet. 116, 31–52.

PT

Vita, F., Kern, C., and Inguaggiato S. (2014) “Development of a portable active long-path differential optical absorption spectroscopy system for volcanic gas measurements” Journal of Sensor and Sensor Systems, 3, 355–367, 2014 doi:10.5194 jsss-3-355-2014 http://www.j-sens-sens-syst.net/3/355/2014/jsss-3-3552014.html

SC

RI

Vita, F., Inguaggiato, S., Bobrowski, N., Calderone L., Galle, B., Parello, F. (2012) “Continuous SO2 flux measurements for Vulcano Island, Italy” Annals of Geophysics 55, 2, 2012; doi: 10.4401/ag-5759 Wakita, H., 1996. Geochemical challenge to earthquake prediction. Proc. Natl. Acad. Sci. U. S. A. 93, 3781–3786.

NU

Wakita, H., Nakamura, Y., Kita, I., Fujii, N., Notsu, K., 1980. Hydrogen release: new indicator of fault activity. Science 210, 188–190.

MA

Werner C, Cardellini C (2006) Comparison of carbon dioxide emissions with fluid upflow, chemistry, and geologic structures at the Rotorua geothermal system, New Zealand. Geothermics 35:221–238

PT E

D

Whitaker, S. (1986) Flow in Porous Media I: A Theoretical Derivation of Darcy’s Law. Transport in Porous Media, 1, 3-25, doi:10.1007/BF01036523

CE

Zakšek K., Hort M., and Lorenz E. (2015) Satellite and Ground Based Thermal Observation of the 2014 Effusive Eruption at Stromboli Volcano. Remote Sens. 2015, 7, 17190–17211; doi:10.3390/rs71215876

AC

Zhang Y. (2005) Wireless sensor network for volcano monitoring. Masters thesis, Chalmers University of Technology.

Captions Figure 1: Aeolian Archipelago sketch map modified from Peccerillo et al 2013; b) Geodynamic setting; c) South Italy map. Figure 2: a) Vulcano Map with location of thermal wells (light blue symbols), UV scanning-DOAS network (white symbols), CO2 flux network (blu symbols), Temperature crater stations (red symbols); b) Inset of crater area.

ACCEPTED MANUSCRIPT Figure 3: Chemical variations in well VH2 in Vulcano Island and comparison with the variations in the fumarolic activity. a) Cl contents; b) pCO2 and c) temperature (t) in well VH2. The grey area represents the CO2 concentration in the fumarole FA (modified from Capasso et al., 2014).

PT E

D

MA

NU

SC

RI

PT

Figure 4: Cl vs SO4 amount in Vulcano thermal wells Samples from wells VH2, VH3, and VC3 collected during certain peculiar periods are distinguished. The mixing line (cyan) links a saline endmember, represented by well VH3 (Federico et al., 2010), and a chiefly meteoric endmember. The black arrow points to the sea water (s.w.) composition. The curve indicates results obtained from the steam-heating model proposed by Federico et al. (2010), according to single-step (s.s.) and multistep (m.s.) models at 80 °C and for different compositions of the deep vapor (model 1: [HCl] = 0.2%, [CO2] = 20%; model 2: [HCl]= 0.1%, [CO2]=5%). Dots on the curves indicate different ϕ/σ ratios. According to this model, the pristine composition of the sampled thermal waters is modified by both the input of chemicals and enthalpy from the ascending fumarolic vapor and the consequent boiling and steam separation, instead of the simple mixing proposed by previous models. Figure 5: Interpretative model of the steam-heating process in La Fossa volcano (modified from Federico et al., 2010; not in scale). The shallow thermal aquifer only receives vapor separating from the 400 °C hydrothermal aquifer, with a variable mass rate; in the proximity of central conduits, the shallowest aquifer directly receives (e.g. in a single step - s.s.) high-enthalpy fluid . High temperature Cl-SO4-rich water derives from single step condensation of high enthalpy fluids, while CO2-rich "peripheral waters" from multistep (m.s.) condensation (e.g. the high enthalpy fluids enters an overlying water bodies, causing further boiling and vapour separation). During periods of enhanced volcanic activity, such as the 1988–1992 crisis, the monophasic zone extended far from central conduits and the 400 °C hydrothermal aquifer was almost completely vaporized; close to the conduits, the thermal aquifer received a larger amount of a vapor richer in CO2, HCl and S, similar to those emitted by crater fumaroles.

AC

CE

Figure 6: Chronograms of a) CO2 and b) He contents, and c) He (expressed as R/Ra, where R=3He/4Hesample and Ra=3He/4Heair) and d) carbon (expressed as δ13CCO2 ‰ units vs PDB) isotopic composition of main fumarolic vents at La Fossa. The periods of simultaneous increases of CO 2 and He contents, lasting some months, mark the so called “geochemical crises” (see text).

Figure 7: Chronogram of a) emission temperature, b) total sulfur content and c) HCl concentration for the crater fumaroles. It is evident the dramatic decrease of acidic species through the 19891993. Figure 8: Time changes of the pH of the deep hydrothermal system below La Fossa cone and total steam output from the crater fumarolic field (modified from Di Liberto et al., 2002; steam output data from Italiano et al., 1997). Figure 9: Binary correlation between a) CO2 and N2 and b) CO2 and He. The best fit straight lines suggest that mixing between magmatic (CO2-He-N2-rich) and hydrothermal (CO2-He-N2-poor) endmembers occurs.

ACCEPTED MANUSCRIPT

PT

Figure 10: He/CO2 ratio versus CO2 concentration in fumarolic gases (modified from Paonita et al., 2013). Fumaroles have been grouped according to their sampling periods. The hyperbolic curves in a) and b) depict two-endmember mixing between magmatic (M1 and M2) and hydrothermal (Hy) fluids (thin curves for 1988–1996, thick curves for 2004). M1 and M2 magmatic endmembers refer to 1988-96 and 2004 periods, respectively, therefore showing a unambiguous change in the composition of the magmatic fluids over time. The estimated magmatic He/CO 2 values and related errors were 1.20 ± 0.15×10-5 for 1988–96 and 1.6 ± 0.2×10-5 for November 2004–January 2005 (see Paonita et al., 2013). The regression of these mixing equations to data was performed with the constrain of the hydrothermal endmember having 3 mol% CO2.

MA

NU

SC

RI

Figure 11: Relationship between a) 3He/4He and CO2 content and b) δ13CCO2 and CO2 in crater fumaroles (modified from Paonita et al., 2013). In a) curves for mixing were computed by using a magmatic endmember having the average 3He/4He of 1996 data (black thin curve) and the average 3 He/4He of 2004 crisis data (thick grey curve). In b) the black thin curve is the main mixing path which fit the 1988–1996 data, while the thick grey curve depicts the best fit of mixing for 2004 crisis. The regression of these mixing equations to data was performed with the constrain of the hydrothermal endmember having 3 mol% CO2. The estimated magmatic δ13CCO2 values and related errors were 0.5 ± 0.20‰ for 1988–96 and 0.0 ± 0.25‰ for November 2004–January 2005. In b) we show the open-system degassing path of gas phase for different melts starting from an initial gas phase with 30 mol% CO2 and δ13CCO2 of 0.5‰ (as derived from the mixing line for 1988– 1996 data). The path of latite is the most probable one to explain the isotope changes of the endmember magmatic fluid.

CE

PT E

D

Figure 12: 3He/4He ratio (corrected for air contamination) vs. He content in fluid inclusions hosted in olivine and pyroxene minerals from rock samples representative of the volcanic history of Vulcano Island. The shaded area represents the 3He/4He range for the magmatic endmember feeding the fumarolic gases (5.2-6.2 R/Ra; Paonita et al., 2013). Only basalt of SOM and latite of ROV have 3He/4He ratio comparable to the fumarolic fluids.

AC

Figure 13: a)Daily average values from the near real time records of temperatures of the fumarole release, validated after 1991; b) Yearly average fumarole temperature calculated from data recorded on a steaming fracture of the Upper Zone (TRZ). The error bars result from the maximum and minimum values recorded year by year on a large sample of values. Figure 14: A mosaic of IR thermal image showing the temperature distribution in and around the main fumaroles area from Schopa et al., (2011). The authors indicated the direction of vent alignments (O, C, R) with respect to the crater centre: O stands for oblique, C for circumferential and R for radial trends.

Figure 15: a) Satellite image of the top of La Fossa Cone (source Google Earth, 22/05/2014), the square and rhombic symbols are for the monitoring stations while TIS and TRZ(1-3) are the locations of thermal probes inside fumarole vents; b) DEM of the central part of the island of

ACCEPTED MANUSCRIPT Vulcano showing the village of Vulcano Porto and the active cone of La Fossa; c) Inset, showing the areas covered by figs. a,b. Fig. 16: Comparison between two different parameters by independent acquisitions carried out in the framework of the geochemical surveillance program at Vulcano Island.

RI

PT

Figure 17: a) Sketch model of the monitored soil profile with indication of the prevailing heat transfer processes, as described in the legend (from Diliberto et al., 2017); b) Model of temperature gradient at BTL station: Between Z1 and the surface the conductive transfer (CdHT) dominates, and δT/δZ is linear; Below the Z2 level at ambient pressures, the temperature of the ground is buffered by the boiling point of the water vapour (i.e. 100°C; 1 bar) that is main fluid phase in the porous structure, the main heat transfer is convective (CvHT). (modified from Diliberto et al., 2017)

SC

Figure 18: a) Heat flux (HF) evaluated by SHS profile of temperature. The linear correlation (R2) among the temperatures values of the monitored profile is reported on the right axis. Low linear 2 correlation values (R <0.98) indicated critical condition for HF monitoring in SHS: The red arrows

MA

NU

(underestimated HF) underline a first critical condition due to the uprise of condensation level to the soil surface; The red dots, above the last HF peak, underline a second critical condition, due to rainfall infiltration in the SHS ground; b) Temperature record of the SHS profile during a geochemical crisis, started in November 2004.

AC

CE

PT E

D

Figure 19: Heat flux variation (f) in watts (w · m-2) from the newest monitored station (VSCS in figure 17a) installed in a SHS area, in May 2015. In the same graph the time variation of air temperature °C (grey curve, scale on the left axis) and the R square (coeff. of correlation) showing the value of linear regression between temperature values and depth (blue curve, scale on the right axis).

Figure 20: Total CO2 output of Vulcano island estimated (482 t day-1). The contributions of the different main areas of degassing have been reported. (a) Crater area (453 t day -1), 362 and 91 t day-1, respectively, for plume and summit soil degassing. (b) Total peripheral soil degassing (20 t day-1). (c) Bubbling gases, main for the Istmo area (3.6 t day-1). (d) Total carbon (HCO3+CO2 dissolved) of aquifer (5.1 t day-1). Figure 21: CO2 fluxes monitoring data (daily average) acquired by VCS station located on the summit area of Vulcano Island, from 2007 to 2010. Yellow line represents the weekly mobile average.

ACCEPTED MANUSCRIPT Figure 22: Location map of the UV-scanning DOAS network to measure the SO2 Plume of Vulcano Island Figure 23: Vulcano SO2 fluxes monitoring data. Figure 24: Stromboli sketch map of deep fluids interacting with aquifers and surficial fluids i) aquifer, ii) soils, iii) plume. Modified from Inguaggiato et al. 2011. Peripheral CO 2 fluxes station (STR01); Summit CO2 fluxes station (STR02).

RI

PT

Figure 25: Stromboli map with location of thermal wells (blue symbols), UV scanning-DOAS FLAME network (yellow symbols), UV scanning-DOAS Novac 2 network (light blue symbols), CO2 flux network (red symbols), C/S plume crater stations (green symbols).

D

MA

NU

SC

Figure 26: He/CO2 vs. δ13C of dissolved CO2 scatter plot in the Stromboli aquifer (modified from Capasso et al., 2005). The progressive 13C-depletion paralleled by increasing He/CO2 ratios in dissolved gases are attributed to Rayleigh-type distillation processes. The red boiling curve represents the fractionation process occurring during vapor separation from a boiling aquifer at a temperature of 130°C. The average chemical and isotopic composition of a crater fumarole (δ13C = −1.8‰, He/ CO2 = 3×10-6; Finizola and Sortino 2003) is chosen as initial conditions. Thereafter, starting from three different fractions of CO2 separated from the boiling aquifer, three curves are drawn, accounting for a fractionation process during dissolution in shallower aquifers at a temperature of 40◦C and pH 6.1. The green solid curve starts from 2%, the dashed line from 4% and the dotted line from 10% of separated vapor. Points along the curves represent the fraction of residual vapor during dissolution in water.

CE

PT E

Figure 27 Time series of the 3He/4He ratio (in units of Rc/Ra) of dissolved thermal wells, performed during 2008–2014. The gray area indicates the standard deviation (±0.15 Ra) of the mean value (4.2 Ra, black dashed line) calculated for 2002–2014 (modified from Rizzo et al., 2015). The red area represents the eruption that lasted from 6 August 2014 to 17 November 2014. Figure 28: C/S ratio data of Stromboli plume in the period 2006-2008.

AC

Figure 29: Schematic illustrative section of Stromboli’s shallow plumbing system in the 0-4 km depth range (modified from Aiuppa et al., 2009). The modelled CO2/SO2 ratios over the same depth-pressure range (calculated by the equilibrium saturation model proposed by Moretti et al., 2003) are shown by vertical bars (for model runs with CO2TOT of 0.2 wt% and 2.1 wt%, respectively). The shallow dyke-conduit system (Chouet et al., 2003) is filled with a high-porphyric (HP) magma emitted as scoriae during persistent Strombolian activity. Assuming that gas-melt separation occurs at the top of this system (at atmospheric pressure), the CO 2/SO2 ratio in the degassed magmatic gas phase would be 0.8-8 (as suggested by the modelled CO2/SO2 ratios). On the other hand, CO2/SO2 ratios as high as 26, recurrently measured in the period December 2006 – April 2007, would be produced by degassing at a depth range 1.5-3.5 km below summit vents. This coincides with the volcano-crust interface, which has been interpreted (Allard et al., 2008; Burton et al. 2007) as a key structural and geological transition in Stromboli’s feeding system, also favouring the formation of gas slugs sustaining Strombolian explosions during “normal” activity at

ACCEPTED MANUSCRIPT Stromboli (Burton et al., 2007). The location of hypocenters of deep volcano-tectonic earthquakes (Patanè et al., 2007) supports the presence of a magma storage zone below 3 km depth, occupied by a low-porphyric (LP) magma (Francalanci et al., 1999; Métrich et al., 2005). The increasing gas contributions from this deep magma storage zone would be the cause of the anomalous composition of the volcanic gas plume emitted at Stromboli in the pre- and syn-eruptive period.

PT

Figure 30: Stromboli Plume-SO2 flux long-time variations expressed in t d-1. The yellow bar represent the 2007 effusive eruption period. The continuous SO 2 data (2009-2010) acquired with UV-scanning DOAS NOVAC II are reported (blue symbols) together with the discontinuous data (red symbols) acquired with mobile DOAS during 2006-2008 (Inguaggiato et al. 2011).

SC

RI

Figure 31: a) Stromboli Soil CO2 fluxes from summit and peripheral areas; b) Cumulated probability of the STR02 log CO2 fluxes of the three sub-period 2000-2004, 2005-2010, 2011-2015; C) Hystogram of log CO2 fluxes of STR02; d) CO2 fluxes recorded at STR02 from 2000 to 2015. Yellow ovals indicate the effusive eruptions. (modified from Inguaggiato et. al 2017).

NU

Figure 32: Map of Panarea island and the group of islets eastward. The colored line shows the area where the main underwater gas emissions are located.

MA

Figure 33: Scuba diving operations for a) gas output measurements and b) gas sample collections from a submarine vent near Lisca Bianca; c) the impressive gas column emitted by the 7 m deep crater that formed during the 02 November 2002 gas burst, close to the Bottaro islet.

PT E

D

Figure 34: Relationships between a) H2S vs. CO2 and b) H2S vs. He for the underwater fumaroles offshore Panarea (modified from Caracausi et al., 2005a). The arrows are a path of selective gas dissolution in seawater computed by using an open system model at 80°C, 3 MPa, 96.5 mol% CO 2 and 3.2 mol% H2S in the pristine gas.

AC

CE

Figure 35: Chronograms of a) He isotope composition, b) geotemperature and c) gas output for the Panarea fumaroles. The gas burst event of November 2002 is marked, while the reported values before this event are representative for the end of 1980s and the beginning of 1990s. He isotope data from Caracausi et al. (2005a) and Heinicke et al. (2009); gas output from Caliro et al. (2004) and Caracausi et al. (2005a); geo-temperatures have been computed by the CO-CO2-CH4 geo-thermometer in presence of a boiling brine (see Caracausi et al., 2005a). Figure 36: H2/CH4 and CO/CH4 relationships for the Panarea gases (modified from Caliro et al., 2004). The theoretical grid for hydrothermal gases was calculated in the temperature range 100373°C and at PCO2 between 0.1 and 100 bar, by assuming coexistence with pure liquid water and typical redox buffer of hydrothermal environments. Volcanic gases and hydrothermal gases areas are indicated. Figure 37: Plot of Log(CO/H2) vs. 2Log(CO2/CO) – 0.5Log(CO2/CH4)(modified from Caracausi et al., 2005a). Isotherms and isopleths of water fugacity are displayed (black lines), as well as the vapour coexisting with 1 wt% NaCl boiling liquid (blue line), computed at a CO2/H2O ratio of the vapour phase equal to 0.13. Data collected during the gas burst show a trend in agreement with the

ACCEPTED MANUSCRIPT coexisting vapour-liquid line, whereas gases collected before and after the burst event are largely scattered, suggesting a lack of chemical equilibrium.

PT

Figure 38: Statistical approach to the volatiles degassing of Vulcano and Stromboli characterized by different activity: a) Solfataric and b) Strombolian respectively (modified from Inguaggiato et al. 2011, 2012a). The volatiles fluxes emitted from the summit crater areas of these volcanoes with different volcanic activity, reflect different style of degassing. An uni-modal distribution characterized the CO2 fluxes recorded at the summit of La Fossa cone and a poli-modal distribution characterized the CO2 fluxes recorded on top of Stromboli island

SC

RI

Figure 39: CO2 peripheral degassing vs CO2 Summit degassing on active volcanoes. R= CO2 Summit/CO2 Peripheral

AC

CE

PT E

D

MA

NU

Figure 40: Geochemical Automated monitoring systems installed on Aeolian active volcanoes.

PT E

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

Figure 1

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

PT E

D

MA

NU

SC

Figure 2

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

PT E

D

MA

Figure 3

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

PT E

D

Figure 4

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

PT E

Figure 5

PT E

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

Figure 6

PT E

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

Figure 7

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

PT E

D

Figure 8

PT E

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

Figure 9

AC

Figure 10

CE

PT E

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

PT E

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

Figure 11

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

PT E

D

Figure 12

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

PT E

Figure 13

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

PT E

Figure 14

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

PT E

D

Figure 15

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

PT E

Figure 16

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

PT E

D

Figure 17

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

PT E

Figure 18

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

PT E

Figure 19

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

PT E

D

MA

Figure 20

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

PT E

D

MA

Figure 21

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

PT E

D

MA

Figure 22

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

PT E

D

MA

NU

Figure 23

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

PT E

Figure 24

PT E

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

Figure 25

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

PT E

D

MA

Figure 26

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

PT E

D

MA

NU

SC

Figure 27

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

PT E

D

MA

NU

SC

Figure 28

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

PT E

D

Figure 29

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

PT E

D

Figure 30

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

PT E

Figure 31

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

PT E

D

MA

Figure 32

PT E

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

Figure 33

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

PT E

D

MA

NU

Figure 34

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

PT E

Figure 35

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

PT E

Figure 36

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

PT E

D

MA

Figure 37

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

PT E

D

Figure 38

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

PT E

D

MA

Figure 39

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

PT E

D

Figure 40