Crustal dynamics of Mount Vesuvius from 1998 to 2005: Effects on seismicity and fluid circulation

Crustal dynamics of Mount Vesuvius from 1998 to 2005:

Effects on seismicity and fluid circulation

Paolo Madonia,1 Cinzia Federico,1 Paola Cusano,2 Simona Petrosino,2

Alessandro Aiuppa,3 and Sergio Gurrieri1

Received 7 June 2007; revised 18 February 2008; accepted 17 March 2008; published 30 May 2008.

[1] This paper presents the results of hydrogeochemical and seismological studies carriedout at Mount Vesuvius during the period June 1998 to December 2005. Hydrogeochemicaldata show the occurrence of slowly varying long-term variations in the total dissolvedsalts and bicarbonate contents of the groundwaters, accompanied by a general decline inwater temperatures. The temporal distributions of air temperature and rainfall in theVesuvius area suggest that these variations do not depend on changes in the hydrologicalregime. The changes in the geochemical parameters are accompanied by slight variationsin both the seismicity rate and energy release. A further relationship between seismicactivity and fluid discharge rate is highlighted by a particular episode that occurred inAugust 2005, when a soil thermal anomaly was observed a few weeks before theoccurrence of a very shallow earthquake. Moment tensor analysis of this earthquakesuggests that the most plausible source mechanism is a shear faulting combined with theopening of tensile crack. This feature is often observed in volcanic areas and it is usuallyrelated to fluid- and/or gas-driven rock fracturing. The observed seismological,hydrological, and geochemical temporal changes are interpreted not as changes of thevolcanic system but in terms of an external forcing as identified in the variation of theregional and local stress field acting on the volcano. This study has inferences ontothe evaluation of the state of activity of volcanic systems and the eventual detection ofunrest phenomena.

Citation: Madonia, P., C. Federico, P. Cusano, S. Petrosino, A. Aiuppa, and S. Gurrieri (2008), Crustal dynamics of Mount Vesuvius

from 1998 to 2005: Effects on seismicity and fluid circulation, J. Geophys. Res., 113, B05206, doi:10.1029/2007JB005210.

1. Introduction

[2] Fluids circulating in volcanic areas (gas, vapor, orliquid water) result from the mixing of volcanic, hydrother-mal, and shallow meteoric components [Pilipenko, 1989;Giggenbach et al., 1990; Fischer et al., 1997; Lewicki et al.,2000; Brusca et al., 2001; Capasso et al., 2001]. Geochem-ical anomalies are hence attributable to three possiblecauses: (1) variations of the deep source (i.e., the volcanicsystem), whose detection is the main challenge of anymonitoring activity; (2) changes in soil permeability dueto the dynamics of the stress field (both at local andregional/subregional scales) acting on the volcanic edificethat may influence the mixing ratio between the differentend-members [Rojstaczer and Wolf, 1992; Rojstaczer et al.,1995; Toutain et al., 1997; Poitrasson et al., 1999; Johnson

et al., 2000; Claesson et al., 2004; Cutillo et al., 2006]; and(3) modifications to the hydrological regime that maychange the dilution of the deep components with variableamounts of meteoric waters.[3] Seismicity is an important indicator of the state of a

volcanic system. The stress field acting on the volcanoaffects seismological parameters such as the seismicity rateand energy release, and hence temporal variations of theseparameters could be indicative of changes in the stressregime. Moreover, the location of the hypocenters and focalmechanisms provide useful insight into the source processesacting inside the volcano. Therefore all of these parametersneed to be monitored in order to better understand thedynamics of a volcano.[4] The seismic activity, hydrological regime, and geo-

chemistry of underground fluids thus represent three cate-gories of measurable parameters, which describe a uniqueshallow process acting in the crust. Crustal loading could beresponsible for recent episodes of apparent volcanic unrest,characterized by either increased seismicity (as in the caseof the seismic crises of October and November 2005 atTanaga volcano [Coombs et al., 2007]) or changes in thehydrothermal manifestations (as in the thermal crisis thatoccurred in the Norris Geyser Basin, Yellowstone NationalPark, in 2003 [United States Geological Survey, 2003]),

JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 113, B05206, doi:10.1029/2007JB005210, 2008

1Sezione di Palermo, Istituto Nazionale di Geofisica e Vulcanologia,Palermo, Italy.

2Osservatorio Vesuviano, Sezione di Napoli, Istituto Nazionale diGeofisica e Vulcanologia, Naples, Italy.

3Dipartimento di Chimica e Fisica della Terra ed Applicazioni alleGeorisorse e ai Rischi Naturali, Universita degli studi di Palermo, Palermo,Italy.

Copyright 2008 by the American Geophysical Union.0148-0227/08/2007JB005210$09.00

B05206 1 of 12

whose interpretation remained ambiguous. In the LongValley caldera (California), the stress field and the propa-gation of seismic waves produced by a series of earthquakeshad effects on the water level of several wells, either directlyor through the solicitation of a hydrothermal aquifer[Roeloffs et al., 2003]. Crustal processes can be constantlyactive in volcano-tectonic (VT) areas, even during non-eruptive or low-seismicity rate periods. Indeed, they repre-sent the background level in active VT areas.[5] The prolonged quiescent state (since 1944) and the

huge amount of available geophysical and geochemical dataon the Somma-Vesuvius volcanic complex makes it asuitable site for observing eventual crustal phenomena andcomparing seismological and geochemical data sets. Theseismic activity has been monitored continuously by aseismic network managed by the Istituto Nazionale diGeofisica e Vulcanologia (INGV) Osservatorio Vesuvianosince 1972, while a geochemical surveillance programmanaged by the Palermo branch of INGV started in 1998.Here we report on seismological, hydrological, and geo-chemical data acquired at Mount Vesuvius during the period1998–2005, which were analyzed for a common drivingmechanism underlying the observed anomalies.

2. Background Information on the Study Area

2.1. Geological Setting

[6] The Somma-Vesuvius complex (1281 m high abovesea level (asl) and 10 km wide) is a stratovolcano within theCampanian Plain (southern Italy), a depression bordered byTertiary and Mesozoic carbonate massifs and filled withvolcaniclastic and sedimentary deposits, at the intersectionof two regional tectonic fault systems (NW–SE/NNW–SSE and NNE–SSW/NE–SW). The first regional structureinvolves the northwestern sector of the Somma edifice forabout 1.5 km and probably represents the surface trace ofthe SW-dipping fault that dislocates the sedimentary base-ment. The second regional structure involves the northeast-ern sector of the Somma edifice and extends in thesouthwestern sector of Mount Vesuvius.[7] Besides these two regional structures, there are also

local eruptive fractures aligned in the E–W and N–Sdirections inside the Somma caldera and on the southernflank of the volcano. Mount Vesuvius developed within awide gravimetric anomaly located in the central part of theCampanian Plain, named the Acerra graben [Scandone etal., 1991; Marzocchi et al., 1993]. This anomaly is relatedto the subsidence of the carbonate basement, lying at adepth of about 2 km beneath Mount Vesuvius.[8] The volcanic activity of Mount Vesuvius has been

characterized by the alternation between pyroclastic erup-tions (separated by quiescent periods) and open conduitphases, characterized by Strombolian and effusive activity[Santacroce et al., 1994; Cioni et al., 1998]. After the 1631A.D. sub-Plinian eruption, an open conduit phase lasteduntil the 1944 eruption, since when the volcano has been ina state of weak volcanic-hydrothermal activity characterizedby diffuse CO2 degassing and low-temperature fumarolicactivity in the crater area, thermal submarine features, andlow seismic activity [Chiodini et al., 2001; Saccorotti et al.,2002; Frondini et al., 2004].

[9] Seismic tomography [Auger et al., 2001] provides anindication of the internal structure, showing the presence ofa melting zone at a depth of about 8 km. Moreover, there istomographic evidence of a zone exhibiting a high P wavevelocity and roughly cylindrical symmetry with respect tothe crater axis that extends from the surface down to a depthof approximately 2 km, where it encounters the carbonatebasement.

2.2. Vesuvius Seismicity: Spatial Patterns and SourceCharacteristics

[10] The seismicity of Mount Vesuvius consists of VTearthquakes (following the classification of Chouet [1996]),characterized by clear P and S wave arrivals and caused byshear failure mechanisms. Since it has been described inmany papers [e.g., Scarpa et al., 2002, and referencestherein], we briefly summarize the main features of theearthquake spatial distribution and source characteristics.[11] Most earthquakes on Mount Vesuvius occur above

depths of 4 km below sea level (bsl) along the crater axiswith a maximum in the spatial distribution at a depth of 2–3 km, which corresponds to the transition between thevolcanic edifice and the limestone basement [Scarpa etal., 2002; Del Pezzo et al., 2004]. Double-couple (DC)mechanisms are generally associated with Vesuvius earth-quakes. The analysis of a data set recorded during 1993–1995 by Bianco et al. [1998] revealed both strike-slip andnormal/reverse dip-slip focal mechanisms with nodal planesmainly oriented along the NW–SE and NE–SW directions.These authors also demonstrated two main orientations forthe P and T axes: the first are in the NNE–SSW and ESE–WNW directions, respectively, while the second are in theESE–WNW and NNE–SSW directions, respectively.[12] The analysis of a selected data set of earthquakes

recorded in 1989–1999 by Zollo et al. [2002] also revealedtwo classes of fault plane solutions, with the distributions ofP and T axes clustering along the ESE–WNW and N–Sdirections, respectively. These orientations are consistentwith one of the two trends revealed by Bianco et al. [1998].[13] The scaling law of the seismic spectrum obtained for

the Vesuvius earthquakes is non-self-similar [Del Pezzo etal., 2004; Galluzzo et al., 2004], indicating that the sourcedimensions, which are approximately constant (on the orderof 100 m), do not scale with the seismic moment. The stressdrop associated with the rupture mechanisms ranges be-tween 0.1 and 10 MPa. In particular, Del Pezzo et al. [2004]showed that the deepest events have an average stress dropof between 1 and 10 MPa, with the shallowest eventscharacterized by stress drop of up to around 1 MPa. Theseauthors suggested that the shallowest low-stress drop eventsare triggered by increasing pore fluid pressure generated bychanges in the level of the hydrothermal aquifer, whose topis located at about 1 km bsl, beneath the crater. Conversely,high-stress drop seismicity is mainly caused by the regionaltectonic stress release occurring in the prefractured carbon-ate basement, whose top is located at about 2.5 km bsl,beneath the crater.[14] An earthquake of magnitude 3.6 occurred at 0741 UT

on 9 October 1999 at about 4 km bsl, corresponding to thecrater area [Zollo et al., 2002; Del Pezzo et al., 2004]. Theearthquake was felt by more than 2 million people livingwithin the urban area of Naples, and it was followed by a

B05206 MADONIA ET AL.: SEISMICITY AND FLUID FLOW AT VESUVIUS

2 of 12

B05206

sequence of low-magnitude events. The duration magnitudeof this event was the highest recorded since the last eruptionof 1944. The source characteristics of this earthquake wereinvestigated by Zollo et al. [2002] and Del Pezzo et al.[2004], with the latter demonstrating that the event waslocated inside the carbonate basement and was generated bytectonic stress release along a preexisting fracture system.This earthquake had an estimated source radius of 150 m,with a seismic moment of about 1013 N m and acorresponding stress drop on the order of 10 MPa [DelPezzo et al., 2004]. The focal mechanism of the earthquakeshowed a dip-slip mechanism with the strike directionoriented ESE–WNW, almost parallel to the main faultpattern of the region [Del Pezzo et al., 2004; Ventura andVilardo, 1999]. The unconstrained moment tensor inversionprovided no evidence of a significant non-DC component[Zollo et al., 2002].

2.3. Vesuvian Aquifer

[15] The chemistry of Somma-Vesuvius groundwaters hasbeen widely investigated with the aim of elucidating thegas-water-rock interactions occurring in the shallow volca-nic aquifer [Avino et al., 1994; Caliro et al., 1998; Celico etal., 1998; Federico, 1999; Federico et al., 2002]. Theseefforts highlighted the firm dependence of water chemistryon the structural features of the volcano. Characterization ofthe stable isotopes in dissolved gases (CO2 and He) dem-onstrated that magmatic volatiles are actively transported bygroundwaters flowing along the main faults and fractures,namely, the fault systems along the NW–SE and NE–SWdirections that pass through both the volcanic edifice andthe sedimentary basement [Federico et al., 2002]. More-

over, a systematic chemical contrast between groundwatersflowing on the southern and northern sectors of the volcaniccomplex has been demonstrated, with the former beingtypically characterized by higher outlet temperatures, totaldissolved solids, and dissolved CO2 contents. Federico etal. [2002] ascribed the high level of CO2 degassing in thesouthern part of Mount Vesuvius to the geometry of thefractured carbonate basement, whose top becomes progres-sively shallower moving from north to south, where it lies ata depth of only a few hundred meters [Berrino et al., 1998].In contrast, in the northern part of the volcano, a thickvolcaniclastic-clayey pelagic rock succession is interposedbetween the carbonate basement and the shallow volcanicsof Somma-Vesuvius [Ippolito et al., 1973; Aprile andOrtolani, 1979], which limits the ascent of deep gases[Celico et al., 1998].[16] Systematic hydrogeochemical surveillance of Mount

Vesuvius began in May 1998. Sites selected for periodicmeasurements can be distinguished in two groups, withthe first comprising sites 1, 6, 13, 14, and 31 (Figure 1)and being characterized by groundwaters with higherdissolved CO2 contents and total dissolved salts (TDS)and the second comprising the remaining sites having lowersalinity and dissolved gases content. These contrastingcompositions suggest the existence of two separate hydro-logical circuits: (1) a deeper and slower-moving one wheregroundwaters may extensively interact with both deepvolcanic-hydrothermal fluids and host rock and (2) ashallower one that is more affected by seasonal meteoro-logical influences and slight interactions with volcanicfluids [Federico et al., 2002].

Figure 1. Map of the Vesuvius area plotting the locations of earthquakes, stations, and measuring pointsbelonging to the seismic and geochemical surveillance networks.

B05206 MADONIA ET AL.: SEISMICITY AND FLUID FLOW AT VESUVIUS

3 of 12

B05206

[17] In conjunction with the seismic sequence of October1999, important chemical variations were observed in aspring located on the northern flank of the volcano (Olivellaspring (Figure 1)), which were related to the stress-inducedinput of acidic volcanic gases [Federico et al., 2004]. Thissite appears to have been particularly sensitive because ofits very low yield with respect to the main aquifer, whosesalinity decreased a few months before the earthquake,followed by a long-term increasing trend.

3. Instruments and Data Acquisition

[18] The seismicity of Mount Vesuvius is monitored bythe permanent seismic network of the INGV OsservatorioVesuviano, which currently comprises 11 short-period 1 HzMark L4C and Geotech S13 geophones, and 2 broadband(40 s) Guralp CMG40T seismometers (Figure 1). Thesignals are telemetered to the acquisition center in Naples,continuously sampled at 100 Hz, and locally stored on ahard disk. In addition, five digital stations (Lennartz PCM5800) equipped with three-component 1 Hz Lennartz LE-3Dlite sensors also operate in the area in local recordingmode at a sampling rate of 125 Hz. Details of the networkconfiguration are available in work by Castellano et al.[2001].[19] The Vesuvian aquifer has been monitored since

1998, when monthly hydrogeochemical measurementsstarted at 10 private wells and 1 spring (Figure 1). Theparameters measured in situ at the outlet of springs ordrilled wells are water temperature, pH, and redox potentialEh, while major ion contents, dissolved gases, and 18O/16Oratios of groundwater are determined in the laboratory. Inaddition, a permanent device for continuous soil tempera-ture measurement (at the VESA station (Figure 1)) has beenoperating on Mount Vesuvius since April 2005. Soil tem-perature is measured every hour using 12 bit GeminiTinytag Plus loggers (precision ±1.0�C, resolution 0.1�C)with an external probe buried in the ground at a depth of30 cm. The device is located inside the fumarolic fields onthe inner scarp of the northern crater rim.

4. Results

4.1. Long-Term Variations

[20] The temporal distribution of the earthquakes thathave occurred at Mount Vesuvius since 1998, whose loca-tions are shown in Figure 1, was obtained from the seismiccatalog for the OVO station (OVO data are available atwww.ov.ingv.it/download.html). The energy temporal dis-tribution can be calculated from the magnitude by applyingthe Gutenberg-Richter relationship [Lay and Wallace,1995],

Log Eð Þ ¼ 2:9 þ 1:9 MD; ð1Þ

where E is the energy in joules and MD is the durationmagnitude.[21] Both the seismicity rate and corresponding energy

level decreased after the seismic crisis related to the earth-quake on 9 October 1999 (Figure 2). The seismicity wasminimal from May 2001 to July 2004, while the numberof earthquakes increased slightly after August 2004.

The seismicity from January 2000 to December 2005 wasalso characterized by low energy release, with only eightearthquakes with magnitudes greater than 2.5 recordedduring this period: five in 2000, two in 2001, and one in2005. A change-point test [Mulargia and Tinti, 1985] basedon the nonparametric Kologorv-Smirnov statistic [Press etal., 1992] revealed two statistically significant variations inthe seismicity rate corresponding to May 2001 and July2004.[22] Vesuvius groundwaters experienced slowly varying

long-term compositional variations during the period inves-tigated. Figure 3 shows the changes in TDS, bicarbonatecontents, and CO2 contents in representative groundwatersamples from both the western (sites 29 and 31 (Figure 3a))and southeastern (sites 6, 13, and 14 (Figure 3b)) flanks ofMount Vesuvius. The two graphs show increasing trends ingroundwater TDS and bicarbonate contents at most of thesites, starting in the months following the October 1999earthquake and peaking during late 2001 to early 2002. Themagnitudes of these variations differed between the sites,but TDS and bicarbonate contents were on average 10–20%higher during late 2001 than before the 1999 earthquake.After steady values during the period 2001–2003, ground-water TDS and bicarbonate contents declined after thesecond half of 2003, and this continued to 2005(Figure 3). In particular, the TDS decrease at well 29(Figure 3a) was paralleled by slight decreases in the con-tents of HCO3 and dissolved CO2. Bicarbonate contents,which were reasonably constant during 2001 and 2002,averaging about 730 mg/l, decreased to about 650 mg/l in2003 and 2004 and were even lower (down to 580 mg/l)after early 2005. Similarly, CO2 contents showed a dramatic30% decrease in the second half of 2004. Synchronoustrends in bicarbonate contents were also detected in wells14, 13, and 6 on the southern flank of the volcano, where adecrease in bicarbonate contents was observed from late2003 and further during 2004 and 2005 in two wells inTorre Annunziata town (wells 13 and 14).[23] The results are summarized in Figure 2, which shows

the monthly averages of TDS and bicarbonate contents(monthly data with one or more missing values wereexcluded from the calculations). The dashed lines inFigure 2 represent the average values for the periods1998–2000, 2001–2003, and 2004–2005. Figure 2 sug-gests that an increasing dilution of the aquifer was accom-panied by a general decline in well water temperature sincethe end of 2003, although with different amplitudes fordifferent groups of sites. Overall since late 2003, thegroundwaters from the Mount Somma and Mount Vesuviussectors have been characterized by temperature declines of0.2�C and 0.3�C, respectively. As noted above for TDS andbicarbonate content, the groundwater temperature was rel-atively stable until 2003, when the phase of temperaturedecline began. Temperature variation was particularly evi-dent in a subgroup of three wells (wells 26, 27, and 29)located on the western flank of Mount Vesuvius, where thedecrease was as high as 0.6�C.[24] To determine the extent to which the above varia-

tions reflect changes in hydrological parameters, yearlyaverages of air temperature and total rainfall amount inthe Vesuvius area are shown in Figure 4. It is particularlynoteworthy that 2003 was characterized by higher air

B05206 MADONIA ET AL.: SEISMICITY AND FLUID FLOW AT VESUVIUS

4 of 12

B05206

temperatures and lower rainfall relative to the averagehydrological regime. The resulting larger evapotranspirationwould be expected to have reduced the infiltration ofmeteoric waters, causing a positive thermal anomaly inthe aquifer (and likely an increase in groundwater TDS),which is in stark contrast to the observed decrease.

4.2. Short-Term Variations: The August 2005Earthquake and the Associated Thermal Anomaly

[25] On 30 August 2005, the Osservatorio Vesuvianoseismic network recorded an earthquake with a durationmagnitude of 2.8. It was a particularly energetic eventcompared to the average seismicity during that year (themaximum recorded magnitude from January to August was

Figure 2. From the top to the bottom: monthly values of well water temperatures in the Somma sector,the Vesuvius sector, and in three selected wells (26, 27, and 29), TDS contents in groundwaters,bicarbonate contents in groundwaters, energy release, and number of earthquakes during the period1998–2005.

B05206 MADONIA ET AL.: SEISMICITY AND FLUID FLOW AT VESUVIUS

5 of 12

B05206

Figure 3. (a) Time trends of bicarbonate content, dissolved carbon dioxide, and TDS in well 29 andTDS in well 31 and (b) time trends of bicarbonate contents in wells 6, 13, and 14.

B05206 MADONIA ET AL.: SEISMICITY AND FLUID FLOW AT VESUVIUS

6 of 12

B05206

2.0, and the last magnitude 2.8 earthquake had beenrecorded on 8 August 2000). For this event, the spectrumat the OVO station showed main peaks in the 2–8 Hzfrequency band, as typically observed for Mount VesuviusVT earthquakes. This event had an unusually shallowlocation compared to the hypocenter depths of the usualseismic activity (Figure 1). Application of the 3-D gridsearch algorithm NonLinLoc [Lomax et al., 2000] with the3-D velocity model of Scarpa et al. [2002] revealed that thisearthquake was located close to the crater axis at a veryshallow depth of about 300 m asl.[26] Under the hypothesis of a DC mechanism, the scalar

seismic moment, corrected for the effects of the propaga-tion, can be estimated using the following relation [Lay andWallace, 1995]:

M0 ¼4prv3s rW0 exp prf0=vsQð Þ

FYq;f; ð2Þ

where r is the distance from the hypocenter, vs is the averageS wave velocity in the medium between the source and thereceiver, r is the average density, W0 is the level of the low-frequency portion of the S wave displacement spectrum, Fis the free surface operator (equal to 2), Y is the radiationpattern term, and Q is the quality factor at referencefrequency f0 (Table 1). We calculated the amplitudespectrum of the displacement of the S wave phase recordedin the horizontal components at the OVO station andevaluated the low-frequency spectral level W0. The scalarseismic moment M0 for the 30 August earthquakeassociated with the DC mechanism was 2.7 � 1011 N m.[27] According to Brune [1970], the source radius for a

hypothesized circular geometry can be estimated via r =0.37vs/fc, where fc (equal to 5.5 Hz) is the corner frequencyobtained from the intersection of the low-frequency asymp-tote with the high-frequency envelope of the amplitudespectrum [Lay and Wallace, 1995]. We calculated the sourceradius associated with the analyzed earthquake to be on theorder of 100 m.

[28] Finally, we evaluated the static stress drop defined as[Lay and Wallace, 1995]

Ds ¼ 0:44M0

r3: ð3Þ

[29] For the estimated values of seismic moment andsource radius, the stress drop was 0.5 MPa. This low valueis consistent with the usual stress drop range reported byDel Pezzo et al. [2004] for the shallowest earthquakesrecorded at Mount Vesuvius.[30] Earthquakes in volcanic areas are often associated

with fluid- and/or gas-driven rock fracturing yielding non-DC components in the rupture mechanisms, which can berevealed by moment tensor analysis. We performed aseismic moment tensor inversion [Herrmann and Ammon,2002] in order to investigate the source mechanism of the30 August earthquake. We selected a 1.75 s long timewindow that began at the onset of the P wave and applieda band-pass filter of 1–5 Hz to the windowed signals. Wecalculated (1) the pure DC mechanism, (2) the deviatoricmechanism (compensated linear vector dipole (CLVD) andDC), and (3) the full unconstrained mechanism (isotropic(ISO), CLVD, and DC components). The best estimate ofthe location of the source mechanism was at a depth of0.4 km asl for all three inversions, corresponding well withthe 3-D location.[31] We applied the F test [Menke, 1984] to the obtained

solutions to define which of the pure DC, deviatoric, andfull mechanisms best represented the source. The resultssuggest that the best model for the source is given by thefull unconstrained inversion that consists of (1) a 2%

Figure 4. Yearly average air temperatures and total amount of rainfall measured in the Ercolano stationof the Servizio Idrografico during the 1997–2005 period.

Table 1. Parameters Used for the Evaluation of the Seismic

Moment of the Earthquake on 30 August 2005a

r, km W0, cm � s vs, m/s r, g/cm3 f0, Hz Y Q

2.4 1.18 � 10�4 1500 2700 1 0.63 60aHere vs is after Lomax et al. [2001], r and Y are after Zollo et al. [2002],

and Q is after Bianco et al. [1998].

B05206 MADONIA ET AL.: SEISMICITY AND FLUID FLOW AT VESUVIUS

7 of 12

B05206

CLVD, (2) a significant (23%) ISO component, and (3) adominant DC component corresponding to strike-slip mech-anism on a dipping fault plane.[32] We found a good agreement between the observed

signals and the computed signals for the retrieved sourcemodel. Figure 5 compares the real signals and syntheticsand shows the stereographic projection associated with theDC component and the relative fault geometrical parameters(slip = 238�, dip = 74�, and rake = 65�). The trend andplunge values are 117� and 54�, respectively, for the P axis,and 347� and 25�, respectively, for the T axis. We performeda jackknife test [Dreger et al., 2000] to exclude the

influence of each station on the stability of the ISO solutionand the possible dependence of the result on azimuth. Fromthese results, we are confident that the source process isdominated by a fracture mechanism with a significant ISOcomponent, which is related to the opening of a tensilecrack.[33] In order to obtain greater insight into the mechanism

that generated this earthquake, we presented our inversionresults on a source-type plot [Hudson et al., 1989]. In thistype of representation, the ratio of the principal moments(components) defines the position of the point that is

Figure 5. Comparison of observed and predicted seismic signals for the full inversion performed on theearthquake on 30 August 2005. All traces are filtered velocities. Light gray lines correspond to observedtraces, and dark gray lines correspond to synthetics. The nodal planes of the double couple superimposedon the deviatoric component, the isotropic component, and the full mechanism are also shown.

B05206 MADONIA ET AL.: SEISMICITY AND FLUID FLOW AT VESUVIUS

8 of 12

B05206

associated with a given source mechanism. The horizontal(T) and vertical (k) coordinates are defined as

T ¼ 2m1

m3j j ð4aÞ

k ¼ mn

mnj j þ m3j j ; ð4bÞ

where m1 and m3 are the absolutely smallest and largestdeviatoric principal moments, respectively, and mn is theISO moment. According to this definition, the horizontal

coordinate is related to the relative sizes of the DC andCLVD components, while the vertical component is ameasure of the volume changes. The advantage of thisrepresentation is that parameters T and k are independent ofnonunique decompositions into ISO, DC, and CLVDcomponents.[34] Figure 6 shows that the point corresponding to this

earthquake is in a region corresponding to a combinedtensile and shear faulting mechanism, with an angle ofabout 30� between the null axis of the shear fault and thefault plane of the tensile crack. According to this represen-tation, the source mechanism of the analyzed earthquakecould be interpreted as the combination of tensile crackopening and shear dislocation.[35] For the 30 August earthquake, a soil temperature

anomaly was observed in the data recorded at the VESAstation (Figure 7). Between April and July 2005, soiltemperatures displayed only short-term fluctuations withamplitudes of a few degrees Celsius, which we ascribe tothe transient cooling effect of rain events. A very largetemperature increase (with Dt = 50�C) occurred between 2and 17 August 2005, which is tenfold larger than the rain-dependent variations. This event occurred just a few weeksbefore the magnitude 2.8 earthquake of 30 August. Thisstriking temperature anomaly was truncated on 17 August2005 by an exceptional rainfall event that caused a sharpdecline in soil temperatures down to preanomaly values. Inthe light of the above-proposed source mechanism for theearthquake on 30 August 2005, we consider that theobserved temperature anomaly provides evidence of anenhanced vapor flux from the soil because of an increasein fluid pressure that gradually built up until the failurethreshold was exceeded, causing the opening of the crack.The soil temperature decreased from August to lateNovember, as expected for the normal seasonal cycle inthe Northern Hemisphere. The only significant variationsduring this period were due to rain events with amplitudesabout twice those of the previous observations, which isvery typical of the autumnal Mediterranean climate. How-ever, a new anomalous period with an increase in temper-

Figure 6. Source-type plot of k and T parameters. Thesolid lines represent the loci of fixed values for the anglebetween the crack plane and null axis of the share fault. Thedot corresponds to the k and T values calculated for theearthquake on 30 August 2005.

Figure 7. Hourly air temperature, soil temperature at 30 cm depth, and rainfall amount measured atVESA station, located on the Vesuvius crater rim.

B05206 MADONIA ET AL.: SEISMICITY AND FLUID FLOW AT VESUVIUS

9 of 12

B05206

ature started at the end of November, which was inconsis-tent with the expected seasonal trend.

5. Discussion

[36] In volcanic and seismic areas, a cause and effectrelationship has been proposed between seismicity and thechemical and hydrological characteristics of groundwatersystems, even though the underlying mechanisms are still amatter of hypotheses [Rojstaczer and Wolf, 1992; Rojstaczeret al., 1995; Claesson et al., 2004]. Both short- and long-term variations (with timescales ranging from a few days toseveral years) in both geochemical parameters and seismicactivity were observed in the Vesuvius volcanic-hydrological system during 1998–2005, which we interpretas having a common source mechanism.[37] Vesuvius groundwaters experienced slowly varying

long-term compositional variations during the period 1998–2005, characterized by temporal changes in salinity anddissolved carbon contents. As the Vesuvius aquifer ischemically heterogeneous with cold and dilute watersoverlaying saline and gas-charged water bodies [Federicoet al., 2002, 2004], the observed variations probably reflecttemporal changes in the contributions of the two compo-nents. The contribution of the deeper saline water washigher from 2001 to 2003 (Figure 2), with the aquifer beingcharacterized by lower salinity and total carbon contents in1999 and after late 2003. In the latter period, the increasingcontribution of the shallower water appears also to besupported by decreasing well temperatures when air tem-perature and rainfall remained steady.[38] Figure 2 indicates that the described geochemical

trends are accompanied by coeval variations in seismicparameters (i.e., number of events and energy release).The seismic crisis of October 1999 was characterized bydeep, high-stress drop earthquakes that occurred in theprefractured carbonate basement in response to the regionaltectonic stress release [Del Pezzo et al., 2004]. After theOctober 1999 seismic crisis, seismic activity decreased tovery low levels, with shallow hypocenters mainly clusteredwithin the volcanic edifice. These shallow and low-stressdrop earthquakes are considered to be mainly triggered bychanges in the pore fluid pressure, in turn caused either bythe charge/discharge mechanism of the shallow aquifer orby perturbations of the local stress field acting on a smallportion of the volcano. After August 2004, the number ofearthquakes increased slightly.[39] Figure 2 reveals that during periods characterized by

higher seismicity that are related to variations in either theregional tectonic stress field or the pore fluid pressure, thechemistry of the Vesuvian aquifer is apparently moreaffected by the shallow diluted water than the deeper salinewater. This effect, though not systematic and requiringvalidation from further measurements, appears to suggestthat variations in the stress field play a role in controllinggroundwater chemistry. We suggest that this cause andeffect relationship probably results from local permeabilitychanges (i.e., opening or closing of microcracks and frac-tures) triggered by the variations of the stress field that alsocause the earthquakes [Rojstaczer et al., 1995; Miller et al.,1996; Miller and Nur, 2000; Claesson et al., 2004; Cutilloet al., 2006]. The permeability changes would favor the

inflow of shallower waters into tapped wells, whereas thedeeper water would remain partially separate.[40] The absence of any compositional changes in the

aquifer related to the earthquake on 30 August 2005occurred during a phase of slight enhancement of seismicity,which is consistent with its shallow hypocenter (300 m asl)being far above the water table. The very shallow locationand the low stress drop observed for this earthquake suggestthat it was triggered by the pressure increase in the fluidcirculating near to central conduits and essentially ofvolcanic and/or hydrothermal origin under a vapor/gasphase, with the exception of transient episodes of percola-tion of rainwaters. This mechanism of pore pressure in-crease is also supported by the earthquake source dynamicsobtained from the moment tensor analysis, which shows adominant DC mechanism and a significant ISO component.This feature is often observed in volcanic areas, and it isusually related to fluid- and/or gas-driven rock fracturing. Infact, the established mechanism corresponds to shear fault-ing combined with tensile crack opening consistent with avariation of the fluid pressure. The crack opening along thedirection orthogonal to the maximum stress axis, s1, couldhave been enhanced by the increase in pore fluid pressurethat gradually lowered the critical threshold value definedby the Griffith failure criterion [Fyfe et al., 1978], favoringthe rupture. In the absence of other warning signs ofthe reactivation of the hydrothermal-magmatic systems, thispore pressure buildup mechanism must be due to the stressfield acting locally on the volcano. The increase in porefluid pressure in the upper portions of central conduitswould consequentially enhance the upwelling of thermalfluids toward the surface, leading to the thermal anomalyobserved at the VESA station a few days before theearthquake.[41] The direction of s1 can be roughly estimated from

the distribution of the P and T axes for the DC component ofthe moment tensor. Actually, the axis corresponding to themaximum stress lies between the direction of the P waveaxis and the fault plane [McKenzie, 1969]. In the case of theDC mechanism obtained for the analyzed earthquake(Figure 5), this direction (although poorly constrained [seeArnold and Townend, 2007]) is roughly aligned with the N-S axis (considering either the fault or the auxiliary plane)and is compatible with the main microcrack alignmentinferred from a recent study of the seismic anisotropy[Del Pezzo et al., 2004].

6. Conclusions

[42] The temporal patterns of seismic activity and geo-chemical parameters at Mount Vesuvius appear to resultfrom a common mechanism, recognized in the variation ofthe stress field acting on the volcano at both regional andlocal scales. Changes in the stress regime are able toinfluence the discharge rates of fluids within the Vesuviusarea, inducing both variations in the geochemical parame-ters and seismicity. Variations in the direction and intensityof the stress field may induce local permeability changes,which influence soil vapor flux in the crater area and mixingbetween deep and meteoric fluids in the Vesuvian aquifer.Therefore the observed anomalies would not be due tovariations of the deep source (the hydrothermal-magmatic

B05206 MADONIA ET AL.: SEISMICITY AND FLUID FLOW AT VESUVIUS

10 of 12

B05206

system), but instead they would be related to the physicalproperties (i.e., permeability) of the medium (volcanicdeposits) seeped by deep fluids escaping toward the surface.These last considerations are very important to the moni-toring of volcanic activity, for which it is fundamentallyimportant to discriminate early signals of a possible volca-nic unrest from geochemical and geophysical anomaliesinduced by external driving mechanisms, such as tectonicactivity or meteorological phenomena. This is especiallytrue in the Vesuvius area, which is located inside a highlypopulated conurbation and where the last eruption (in 1944)occurred before the development of modern monitoringsystems. The absence of any significant volcanic unrestsince the last eruption has hindered the definition of reliablethresholds of volcanic hazard based on geophysical orgeochemical precursors.

[43] Acknowledgments. Francesca Bianco and Edoardo Del Pezzoare fully acknowledged for their useful suggestions and comments. Theassistance in statistical topics by Lucia Zaccarelli is greatly appreciated.John Townend, Peter Cervelli, and an anonymous reviewer greatly con-tributed to the improvement of the paper with their suggestions.

ReferencesAprile, F., and F. Ortolani (1979), Sulla struttura profonda della PianaCampana, Boll. Soc. Nat. Napoli, 88, 243–261.

Arnold, R., and J. Townend (2007), A Bayesian approach to estimatingtectonic stress from seismological data, Geophys. J. Int., 170, 1336–1356, doi:10.1111/j.1365-246X.2007.03485.x.

Auger, E., P. Gasparini, J. Virieux, and A. Zollo (2001), Seismic evidenceof an extended magmatic sill under Mt. Vesuvius, Science, 294, 1510–1512, doi:10.1126/science.1064893.

Avino, R., P. Celico, L. Esposito, M. R. Ghiara, P. La Gioia, R. Pece,V. Piscopo, and D. Stanzione (1994), Caratteristiche idrogeologiche eidrogeochimiche del sistema Vesuviano, paper presented at the AnnualMeeting, Cons. Naz. della Ric.-Gruppo Naz. per la Vulcanologia, Rome,19–21 Dec.

Berrino, G., G. Corrado, and U. Riccardi (1998), Sea gravity data in thegulf of Naples: A contribution to delineating the structural pattern of theVesuvian are, J. Volcanol. Geotherm. Res., 82, 139–150, doi:10.1016/S0377-0273(97)00061-9.

Bianco, F., M. Castellano, G. Milano, G. Ventura, and G. Vilardo (1998),The Somma-Vesuvius stress field induced by regional tectonics: Evi-dences from seismological and mesostructural data, J. Volcanol.Geotherm. Res., 82, 199–218, doi:10.1016/S0377-0273(97)00065-6.

Brune, J. N. (1970), Tectonic stress and seismic shear waves fromearthquakes, J. Geophys. Res., 75, 4997 – 5009, doi:10.1029/JB075i026p04997.

Brusca, L., A. Aiuppa, W. D’Alessandro, F. Parello, P. Allard, andA. Michel (2001), Geochemical mapping of magmatic gas-water-rockinteractions in the aquifer of Mount Etna volcano, J. Volcanol. Geotherm.Res., 108, 199–218, doi:10.1016/S0377-0273(00)00286-9.

Caliro, S., C. Panichi, and D. Stanzione (1998), Baseline study of theisotopic and chemical composition of waters associated with the Som-ma-Vesuvius volcanic system, Acta Vulcanol., 10, 19–25.

Capasso, G., W. D’Alessandro, R. Favara, S. Inguaggiato, and F. Parello(2001), Interaction between the deep fluids and the shallow groundwaterson Vulcano island (Italy), J. Volcanol. Geotherm. Res., 108, 187–198,doi:10.1016/S0377-0273(00)00285-7.

Castellano, M., C. Buonocunto, M. Capello, and M. La Rocca (2001),Seismic surveillance of active volcanoes: The Osservatorio VesuvianoSeismic Network (OVSN-southern Italy), Seismol. Res. Lett., 73, 168–175.

Celico, P., D. Stanzione, L. Esposito, M. R. Ghiara, V. Piscopo, S. Caliro,and P. La Gioia (1998), Caratterizzazione idrogeologica e idrogeochimicadell’area Vesuviana, Boll. Soc. Geol. Ital., CXVII, 3–20.

Chiodini, G., L. Marini, and M. Russo (2001), Geochemical evidences ofhigh temperature hydrothermal brines at Vesuvio volcano (Italy), Geo-chim. Cosmochim. Acta, 65(13), 2129 – 2147, doi:10.1016/S0016-7037(01)00583-X.

Chouet, B. (1996), New methods and future trends in seismological volcanomonitoring, in Monitoring and Mitigation of Volcano Hazards, edited byR. Scarpa and R. I. Tilling, pp. 23–97, Springer, New York.

Cioni, R., P. Marianelli, and R. Santacroce (1998), Thermal and composi-tional evolution of the shallow magma chambers of Vesuvius: Evidencefrom pyroxene phenocrysts and melt inclusions, J. Geophys. Res., 103,18,277–18,294, doi:10.1029/98JB01124.

Claesson, L., A. Skelton, C. Grahmn, C. Dietl, M. Morth, P. Torssander,and I. Kockum (2004), Hydrogeochemical changes before and after amajor earthquake, Geology, 32(8), 641–644, doi:10.1130/G20542.1.

Coombs, M. L., S. M. White, and D. W. Scholl (2007), Massive edificefailure at Aleutian Arc volcanoes, Earth Planet. Sci. Lett., 256, 403–418,doi:10.1016/jepsl.2007.01.030.

Cutillo, P. A., S. Ge, and E. J. Screaton (2006), Hydrodynamic response ofsubduction zones to seismic activity: A case study for the Costa Ricamargin, Tectonophysics, 426(1 – 2), 167 – 188, doi:10.1016/j.tecto.2006.02.017.

Del Pezzo, E., F. Bianco, and G. Saccorotti (2004), Seismic source dy-namics at Vesuvius volcano, Italy, J. Volcanol. Geotherm. Res., 133,23–39, doi:10.1016/S0377-0273(03)00389-5.

Dreger, D. S., H. Tkaleiæ, and M. Johnston (2000), Dilational processesaccompanying earthquakes in the Long Valley caldera, Science, 288,122–125, doi:10.1126/science.288.5463.122.

Federico, C. (1999), Interaction between magmatic gases and the hydro-logical system at Vesuvius (southern Italy) during current dormancy:Evidence from water and gas geochemistry, Ph.D thesis, Univ. of Paler-mo, Palermo, Italy.

Federico, C., A. Aiuppa, P. Allard, S. Bellomo, P. Jean-Baptiste, F. Parello,and M. Valenza (2002), Magma-derived gas influx and water-rock inter-actions in the volcanic aquifer of Mt. Vesuvius, Italy, Geochim. Cosmo-chim. Acta, 66, 963–981, doi:10.1016/S0016-7037(01)00813-4.

Federico, C., A. Aiuppa, R. Favara, S. Gurrieri, and M. Valenza (2004),Geochemical monitoring of groundwaters (1998–2001) at Vesuvius vol-cano (Italy), J. Volcanol. Geotherm. Res., 133, 81–104, doi:10.1016/S0377-0273(03)00392-5.

Fischer, T. P., N. C. Sturchio, J. Stix, G. B. Arehart, D. Counce, and S. N.Williams (1997), The chemical and isotopic composition of fumarolicgases and spring discharges from Galeras Volcano, Colombia,J. Volcanol. Geotherm. Res., 77, 229 – 253, doi:10.1016/S0377-0273(96)00096-0.

Frondini, F., G. Chiodini, S. Caliro, C. Cardellini, D. Granieri, andG. Ventura (2004), Diffuse CO2 degassing at Vesuvio, Italy, Bull. Volca-nol., 66, 642–651, doi:10.1007/s00445-004-0346-x.

Fyfe, W. S., N. J. Price, and A. B. Thompson (1978), Fluids in the Earth’sCrust, Elsevier, New York.

Galluzzo, D., E. Del Pezzo, M. La Rocca, and S. Petrosino (2004), Peakground acceleration produced by local earthquakes in volcanic areas ofCampi Flegrei and Mt. Vesuvius, Ann. Geophys., 47(4), 1377–1390.

Giggenbach, W. F., N. Garcia, A. Londono, L. Rodriguez, N. Rojas, andM. L. Calvache (1990), The geochemistry of fumarolic vapor and ther-mal-spring-discharges from the Nevado del Ruiz volcanic-magmatic-hy-drothermal system, Colombia, J. Volcanol. Geotherm. Res., 42, 13–39,doi:10.1016/0377-0273(90)90067-P.

Herrmann, R. B., and C. J. Ammon (2002), Computer programs in seis-mology – Source inversion: Users manual, report, St. Louis Univ., St.Louis, Mo.

Hudson, J. A., R. G. Pearce, and R. M. Rogers (1989), Source type plot forinversion of the moment tensor, J. Geophys. Res., 94, 765 – 774,doi:10.1029/JB094iB01p00765.

Ippolito, F., F. Ortolani, and M. Russo (1973), Struttura marginale tirrenicadell’Appennino Campano: Reinterpretazione dei dati di antiche ricerchedi idrocarburi, Mem. Soc. Geol. Ital., 12, 227–250.

Johnson, H. P., M. Hutnak, R. P. Dziak, C. G. Fox, I. Urcuyo, J. P. Cowen,J. Nabelek, and C. Fisher (2000), Earthquake-induced changes in a hy-drothermal system on the Juan de Fuca mid-ocean ridge, Nature, 407,174–177, doi:10.1038/35025040.

Lay, T., and T. Wallace (1995), Modern Global Seismology, Elsevier, SanDiego, Calif.

Lewicki, J. L., T. Fischer, and S. N. Williams (2000), Chemical and isotopiccompositions of fluids at Cumbal Volcano, Colombia: Evidence for mag-matic contribution, Bull. Volcanol., 62, 347 – 361, doi:10.1007/s004450000100.

Lomax, A., J. Virieux, P. Volant, and C. Berge (2000), Probabilistic earth-quake location in 3D and layered models: Introduction of a Metropolis-Gibbs method and comparison with linear locations, in Advances inSeismic Event Location Thurber, edited by C. H. Rabinowitz andN. Rabinowitz, pp. 101–134, Springer, Amsterdam.

Lomax, A., A. Zollo, P. Capuano, and J. Virieux (2001), Precise, absoluteearthquake location under Somma-Vesuvius volcano using a new 3Dvelocity model, Geophys. J. Int., 146, 313–321, doi:10.1046/j.0956-540x.2001.01444.x.

Marzocchi, W., R. Scandone, and F. Mulargia (1993), The tectonic settingof Mount Vesuvius and the correlation between its eruptions and the

B05206 MADONIA ET AL.: SEISMICITY AND FLUID FLOW AT VESUVIUS

11 of 12

B05206

earthquakes of the southern Apennines, J. Volcanol. Geotherm. Res., 58,27–41, doi:10.1016/0377-0273(93)90100-6.

McKenzie, D. P. (1969), The relation between fault plane solutions forearthquakes and the directions of the principal stresses, Bull. Seismol.Soc. Am., 59(2), 591–601.

Menke, W. (1984), Geophysical Data Analysis: Discrete Inverse Theory,pp. 96–97, Elsevier, New York.

Miller, S. A., and A. Nur (2000), Permeability as a toggle-switch in fluid-controlled crustal processes, Earth Planet. Sci. Lett., 183, 133–146,doi:10.1016/S0012-821X(00)00263-6.

Miller, S. A., A. Nur, and D. A. Olgaard (1996), Earthquakes as a coupledshear stress, high pore pressure dynamical system, Geophys. Res. Lett.,23(2), 197–200, doi:10.1029/95GL03178.

Mulargia, F., and S. Tinti (1985), Seismic sample areas defined from in-complete catalogues: An application to the Italian territory, Phys. EarthPlanet. Inter., 40, 273–300, doi:10.1016/0031-9201(85)90038-X.

Pilipenko, G. F. (1989), Thermal springs of the Karymskii volcanic centerin Kamchatka, Volcanol. Seismol., 6, 85–96.

Poitrasson, F., S. Dundas, J. P. Toutain, M. Munoz, and A. Rigo (1999),Earthquake-related elemental and isotopic lead anomaly in a springwater,Earth Planet. Sci. Lett., 169 , 269 – 276, doi:10.1016/S0012-821X(99)00085-0.

Press, W. H., S. A. Teukolsky, W. T. Vetterling, and B. P. Flannery (1992),Numerical Recipes in C: The Art of Scientific Computing, CambridgeUniv. Press, New York.

Roeloffs, E., M. Sneed, D. L. Galloway, M. L. Sorey, C. D. Farrar, J. F.Howle, and J. Hughes (2003), Water-level changes induced by local anddistant earthquakes at Long Valley caldera, California, J. Volcanol.Geotherm. Res., 127, 269–303, doi:10.1016/S0377-0273(03)00173-2.

Rojstaczer, S., and S. Wolf (1992), Permeability changes associated withlarge earthquakes: An example from Loma Prieta, California, Geology,20(3), 211–214, doi:10.1130/0091-7613(1992)020<0211:PCAWLE>2.3.CO;2.

Rojstaczer, S., S. Wolf, and R. Michel (1995), Permeability enhancement inthe shallow crust as a cause of earthquake-induced hydrological changes,Nature, 373, 237–239, doi:10.1038/373237a0.

Saccorotti, G., G. Ventura, and G. Vilardo (2002), Seismic swarms relatedto diffusive processes: The case of Somma-Vesuvius volcano, Italy, Geo-physics, 67(1), 199–203, doi:10.1190/1.1451551.

Santacroce, R., R. Cioni, L. Civetta, P. Marianelli, and N. Metrich (1994),How Vesuvius works, Atti Conveg. Lincei, 112, 185–196.

Scandone, R., F. Bellucci, L. Lirer, and G. Rolandi (1991), The structure ofthe Campanian Plain and the activity of the Neapolitan volcanoes (Italy),J. Volcanol. Geotherm. Res. , 48 , 1 – 31, doi:10.1016/0377-0273(91)90030-4.

Scarpa, R., F. Tronca, F. Bianco, and E. Del Pezzo (2002), High resolutionvelocity structure beneath Mount Vesuvius from seismic array data, Geo-phys. Res. Lett., 29(21), 2040, doi:10.1029/2002GL015576.

Toutain, J. P., M. Munoz, F. Poitrasson, and A. C. Lienard (1997), Spring-water chloride ion anomaly prior to a ML = 5.2 Pyrenean earthquake,Earth Planet. Sci. Lett., 149 , 113 – 119, doi:10.1016/S0012-821X(97)00066-6.

United States Geological Survey (2003), Notable changes in thermal activ-ity at Norris Geyser Basin provide opportunity to study hydrothermalsystem, report, Reston, Va.

Ventura, G., and G. Vilardo (1999), Slip tendency analysis of the Vesuviusfaults: Implications for the seismotectonic and volcanic hazard assess-ment, Geophys. Res. Lett., 26(21), 3229 – 3232, doi:10.1029/1999GL005393.

Zollo, A., W. Marzocchi, P. Capuano, A. Lomax, and G. Iannaccone(2002), Space and time behavior of seismic activity at Mt. Vesuviusvolcano, southern Italy, Bull. Seismol. Soc. Am., 92(2), 625 – 640,doi:10.1785/0120000287.

�����������������������A. Aiuppa, Dipartimento di Chimica e Fisica della Terra ed Applicazioni

alle Georisorse e ai Rischi Naturali, Universita degli studi di Palermo, I-90123 Palermo, Italy.P. Cusano and S. Petrosino, Osservatorio Vesuviano, Sezione di Napoli,

Istituto Nazionale di Geofisica e Vulcanologia, Via Diocleziano 328, I-80124 Naples, Italy. ([email protected])C. Federico, S. Gurrieri, and P. Madonia, Sezione di Palermo, Istituto

Nazionale di Geofisica e Vulcanologia, Via Ugo La Malfa 153, I-90146Palermo, Italy.

B05206 MADONIA ET AL.: SEISMICITY AND FLUID FLOW AT VESUVIUS

12 of 12

B05206