The evaluation of volcanic risk in the Vesuvian area

Volcanic Risk has been defined as the product: R = Value × Vulnerability × Hazard, where value is the total amount of lives or properties at risk for a volcani eruption, the vulnerability is the percentage of value at risk for a given volcanic event, and the hazard is the probability that a given area may be affected by a certain volcanic phenomenon.We used this definition to evaluate the Risk of loss of human lives for volcanic eruptions of Vesuvius. Value has been determined based on the total number of inhabitants living in areas that could be affected by an eruption. Vulnerability is based on the relative probability of deaths as a result of different volcanic phenomena (tephra fall, pyroclastic flows, etc.). Hazard is evaluated based on the absolute probability of a given phenomenon in a certain area. This last parameter is the most difficult to evaluate.We subdivided the activity of Vesuvius, that produces risk of loss of human lives, into three classes of eruptions, based on the Volcanic Explosivity Index. We assume that the events of each class are distributed according to a poissonian distribution (this is demonstrated for VEI = 3, and inferred for the other classes), so that we can evaluate the absolute probability of an eruption for each class within a given time span. We use a time window of 10 years and evaluate the probabilities of occurrence of at least one eruption for VEI = 3, 4, 5; the probability is respectively: P₃ = 0.09896, P₄ = 0.01748, P₅ = 0.00298We have made a hazard evaluation for the entire Vesuvian area as well as an evaluation of Volcanic Risk.The obtained map shows that the areas with higher risk are on the southern side of Vesuvius, in the coastal region where each town is characterized by an average Risk of 1000 inhabitants/10 years. The risk regularly decreases with increasing distance from the volcano. The risk is mostly due to the events with VEI = 3 and 4, as the most destructive effects of VEI = 5 are counterbalanced by the lower probability of such events.The map of volcanic risk is not intended to predict the loss of human lives of an eruption, but to give a methodology that permits to identify those areas where it is necessary to operate in order to reduce the risk of an eruption before it occurs.     

The incorporation of stochasticity in risk analysis and management: a case study

 The selection of optimal management strategies for environmental contaminants requires detailed information on the risks imposed on populations. These risks are characterized by both inter-subject variability (different individuals having different levels of risk) and by uncertainty (there is uncertainty about the risk associated with the Yth percentile of the variability distribution). In addition, there is uncertainty introduced by the inability to agree fully on the appropriate decision criteria. This paper presents a methodology for incorporating uncertainty and variability into a multi-medium, multi-pathway, multi-contaminant risk assessment, and for placing this assessment into an optimization framework to identify optimal management strategies. The framework is applied to a case study of a sludge management system proposed for North Carolina and the impact of stochasticity on selection of an optimal strategy considered. Different sets of decision criteria reflecting different ways of treating stochasticity are shown to lead to different selections of optimal management strategies.     

Role of carbon dioxide in the dynamics of magma ascent in explosive eruptions

 The role of carbon dioxide in the dynamics of magma ascent in explosive eruptions is investigated by means of numerical modeling. The model is steady, one-dimensional, and isothermal; it calculates the separated flow of gas and a homogeneous mixture of liquid magma and crystals. The magma properties are calculated on the basis of magma composition and crystal content and are allowed to change along the conduit due to pressure decrease and gas exsolution. The effect of the presence of a two-component (water + carbon dioxide) exsolving gas phase is investigated by performing a parametric study on the CO₂/(H₂O+CO₂) ratio, which is allowed to vary from 0 to 0.5 at either constant total volatile or constant water content. The relatively insoluble carbon dioxide component plays an important role in the location of the volatile-saturation and magma-fragmentation levels and in the distribution of the flow variables in the volcanic conduit. In detail, the results show that an increase of the proportion of carbon dioxide produces a decrease of the mass flow rate, pressure, and exit mixture density, and an increase of the exit gas volume fraction and depth of the fragmentation level. A relevant result is the different role played by water and carbon dioxide in the eruption dynamics; an increasing amount of water produces an increase of the mass flow rate, and an increasing amount of carbon dioxide produces a decrease. Even small amounts of carbon dioxide have major consequences on the eruption dynamics, implying that the multicomponent nature of the volcanic gas must be taken into account in the prediction of the eruption scenario and the forecasting of volcanic hazard. Received: 6 March 1998 / Accepted: 28 October 1998     

Petrology of lavas from the Puu Oo eruption of Kilauea Volcano: III. The Kupaianaha episode (1986–1992)

 The Puu Oo eruption has been remarkable in the historical record of Kilauea Volcano for its duration (over 13 years), volume (>1 km³) and compositional variation (5.7–10 wt.% MgO). During the summer of 1986, the main vent for lava production moved 3 km down the east rift zone and the eruption style changed from episodic geyser-like fountaining at Puu Oo to virtually continuous, relatively quiescent effusion at the Kupaianaha vent. This paper examines this next chapter in the Puu Oo eruption, episodes 48 and 49, and presents new ICP-MS trace element and Pb-, Sr-, and Nd-isotope data for the entire eruption (1983–1994). Nearly aphyric to weakly olivine-phyric lavas were erupted during episodes 48 and 49. The variation in MgO content of Kupaianaha lavas erupted before 1990 correlates with changes in tilt at the summit of Kilauea, both of which probably were controlled by variations in Kilauea's magma supply rate. These lavas contain euhedral olivines which generally are in equilibrium with whole-rock compositions, although some of the more mafic lavas which erupted during 1990, a period of frequent pauses in the eruption, accumulated 2–4 vol.% olivine. The highest forsterite content of olivines (∼85%) in Kupaianaha lavas indicates that the parental magmas for these lavas had MgO contents of ∼10 wt.%, which equals the highest observed value for lavas during this eruption. The composition of the Puu Oo lavas has progressively changed during the eruption. Since early 1985 (episode 30), when mixing between an evolved rift zone magma and a more mafic summit reservoir-derived magma ended, the normalized (to 10 wt.% MgO) abundances of highly incompatible elements and CaO have systematically decreased with time, whereas ratios of these trace elements and Pb, Sr, and Nd isotopes, and the abundances of Y and Yb, have remained relatively unchanged. These results indicate that the Hawaiian plume source for Puu Oo magmas must be relatively homogeneous on a scale of 10–20 km³ (assuming 5–10% partial melting), and that localized melting within the plume has apparently progressively depleted its incompatible elements and clinopyroxene component as the eruption continued. The rate of variation of highly incompatible elements in Puu Oo lavas is much greater than that observed for Kilauea historical summit lavas (e.g., Ba/Y 0.09 a^–1 vs ∼0.03 a^–1). This rapid change indicates that Puu Oo magmas did not mix thoroughly with magma in the summit reservoir. Thus, except for variable amounts of olivine fractionation, the geochemical variation in these lavas is predominantly controlled by mantle processes. Received: 8 March 1996 / Accepted: 30 April 1996     

A Miocene coarse volcaniclastic mass-flow deposit in the Shimane Peninsula,SW Japan: product of a deep submarine eruption?

 A subaqueous volcaniclastic mass-flow deposit in the Miocene Josoji Formation, Shimane Peninsula, is 15–16 m thick, and comprises mainly blocks and lapilli of rhyolite and andesite pumices and non- to poorly vesiculated rhyolite. It can be divided into four layers in ascending order. Layer 1 is an inversely to normally graded and poorly sorted lithic breccia 0.3–6 m thick. Layer 2 is an inversely to normally graded tuff breccia to lapilli tuff 6–11 m thick. This layer bifurcates laterally into minor depositional units individually composed of a massive, lithic-rich lower part and a diffusely stratified, pumice-rich upper part with inverse to normal grading of both lithic and pumice clasts. Layer 3 is 2.5–3 m thick, and consists of interbedded fines-depleted pumice-rich and pumice-poor layers a few centimeters thick. Layer 4 is a well-stratified and well-sorted coarse ash bed 1.5–2 m thick. The volcaniclastic deposit shows internal features of high-density turbidites and contains no evidence for emplacement at a high temperature. The mass-flow deposit is extremely coarse-grained, dominated by traction structures, and is interpreted as the product of a deep submarine, explosive eruption of vesicular magma or explosive collapse of lava. Received: 10 January 1996 / Accepted: 23 February 1996     

Shallow and deep reservoirs involved in magma supply of the 1944 eruption of Vesuvius

 During the 1944 eruption of Vesuvius a sudden change occurred in the dynamics of the eruptive events, linked to variations in magma composition. K-phonotephritic magmas were erupted during the effusive phase and the first lava fountain, whereas the emission of strongly porphyritic K-tephrites took place during the more intense fountain. Melt inclusion compositions (major and volatile elements) highlight that the magmas feeding the eruption underwent differentiation at different pressures. The K-tephritic volatile-rich melts (up to 3 wt.% H₂O, 3000 ppm CO₂, and 0.55 wt.% Cl) evolved to reach K-phonotephritic compositions by crystallization of diopside and forsteritic olivine at total fluid pressure higher than 300 MPa. These magmas fed a very shallow reservoir. The low-pressure differentiation of the volatile-poor K-phonotephritic magmas (H₂O<1 wt.%) involved mixing, open-system degassing, and crystallization of leucite, salite, and plagioclase. The eruption was triggered by intrusion of a volatile-rich magma batch that rose from a depth of 11–22 km into the shallow magma chamber. The first phase of the eruption represents the partial emptying of the shallow reservoir, the top of which is within the volcanic edifice. The newly arrived magma mixed with that resident in the shallow reservoir and forced the transition from the effusive to the lava fountain phase of the eruption. Received: 14 September 1998 / Accepted: 10 January 1999     

Modeling of the ascent of magma during the plinian eruption of Vesuvius in A.D. 79

The ascent of magma during the A.D. 79 eruption of Vesuvius was studied by a steady-state, one-dimensional, and nonequilibrium two-phase flow model. The gas exsolution process was modeled by assuming a chemical equilibrium between the exsolved and dissolved gas, whereas the magma density and viscosity were modeled by accounting for the crystal content in magma. The exsolution, density, and viscosity models consider the effect of different compositions of the white and gray magmas. By specifying the conduit geometry and magma composition, and employing the model to search for the maximum discharge rate of magma which is consistent with the specified geometry and magma composition, the model was then used to establish the two-phase flow parameters along the conduit. It was found that for all considered conditions the magma pressure in the conduit decreases below the lithostatic pressure near the magma fragmentation level, and that in the deep regions of the conduit the white magma pressure is larger and the gray magma pressure is lower than the lithostatic one. The exsolution and fragmentation levels were found to be deeper for the white than for the gray magma, and the changing composition during the eruption causes an increase of the exit pressure and decrease of the exit gas volumetric fraction. The model also predicted a minimum conduit diameter which is consistent with the white and gray magma compositions and mass flow-rates. The predictions of the model were shown to be consistent with column collapses during the gray eruption phase, large presence of carbonate lithics in the gray pumice fall deposit, and magma-water interaction during a late stage of the eruption.     

Continuous monitoring of volcanic eruption dynamics: a review of various techniques and new results from a frequency-modulated radar Doppler system

 In situ measurement of volcanic eruption velocities is one of the great challenges left in geophysical volcanology. In this paper we report on a new radar Doppler technique for monitoring volcanic eruption velocities. In comparison with techniques employed previously (e.g., photographic methods or acoustic Doppler measurements), this method allows continuous recordings of volcanic eruptions even during poor visibility. Also, radar Doppler instruments are usually light weight and energy efficient, which makes them superior to other Doppler techniques based on laser light or sound. The proposed new technique was successfully tested at Stromboli Volcano in late 1996 during a period of low activity. The recorded data allow a clear distinction between particles rising from the vent and particles falling back towards the vent. The mean eruption velocity was approximately 10 m/s. Most of the eruptions recorded by radar were correlated to seismic recordings. The correlation between the magnitude of the volcanic shocks and the eruption force index defined in the paper may provide new insights into magma transport in the conduit. Received: 15 May 1998 / Accepted: 15 December 1998     

Tsunami generation by pyroclastic flow during the 3500-year B.P. caldera-forming eruption of Aniakchak Volcano, Alaska

 A discontinuous pumiceous sand, a few centimeters to tens of centimeters thick, is located up to 15 m above mean high tide within Holocene peat along the northern Bristol Bay coastline of Alaska. The bed consists of fine-to-coarse, poorly to moderately well-sorted, pumice-bearing sand near the top of a 2-m-thick peat sequence. The sand bed contains rip-up clasts of peat and tephra and is unique in the peat sequence. Major element compositions of juvenile glass from the deposit and radiocarbon dating of enclosing peat support correlation of the pumiceous sand with the caldera-forming eruption of Aniakchak Volcano. The distribution of the sand and its sedimentary characteristics are consistent with emplacement by tsunami. The pumiceous sand most likely represents redeposition by tsunami of climactic fallout tephra and beach sand during the approximately 3.5 ka Aniakchak caldera-forming eruption on the Alaska Peninsula. We propose that a tsunami was generated by the sudden entrance of a rapidly moving, voluminous pyroclastic flow from Aniakchak into Bristol Bay. A seismic trigger for the tsunami is unlikely, because tectonic structures suitable for tsunami generation are present only south of the Alaska Peninsula. The pumiceous sand in coastal peat of northern Bristol Bay is the first documented geologic evidence of a tsunami initiated by a volcanic eruption in Alaska. Received: 3 December 1997 / Accepted: 11 April 1998     

The 1984 to 1996 cyclic activity of Lascar Volcano, northern Chile: cycles of dome growth, dome subsidence, degassing and explosive eruptions

 Lascar Volcano (5592 m; 23°22'S, 67°44'W) entered a new period of vigorous activity in 1984, culminating in a major explosive eruption in April 1993. Activity since 1984 has been characterised by cyclic behaviour with recognition of four cycles up to the end of 1993. In each cycle a lava dome is extruded in the active crater, accompanied by vigorous degassing through high-temperature, high-velocity fumaroles distributed on and around the dome. The fumaroles are the source of a sustained steam plume above the volcano. The dome then subsides back into the conduit. During the subsidence phase the velocity and gas output of the fumaroles decrease, and the cycle is completed by violent explosive activity. Subsidence of both the dome and the crater floor is accommodated by movement on concentric, cylindrical or inward-dipping conical fractures. The observations are consistent with a model in which gas loss from the dome is progressively inhibited during a cycle and gas pressure increases within and below the lava dome, triggering a large explosive eruption. Factors that can lead to a decrease in gas loss include a decrease in magma permeability by foam collapse, reduction in permeability due to precipitation of hydrothermal minerals in the pores and fractures within the dome and in country rock surrounding the conduit, and closure of open fractures during subsidence of the dome and crater floor. Dome subsidence may be a consequence of reduction in magma porosity (foam collapse) as degassing occurs and pressurisation develops as the permeability of the dome and conduit system decreases. Superimposed upon this activity are small explosive events of shallow origin. These we interpret as subsidence events on the concentric fractures leading to short-term pressure increases just below the crater floor. Received: 12 December 1996 / Accepted: 6 May 1997     

Degassing and microlite crystallization during pre-climactic events of the 1991 eruption of Mt. Pinatubo, Philippines

 Dacite tephras produced by the 1991 pre-climactic eruptive sequence at Mt. Pinatubo display extreme heterogeneity in vesicularity, ranging in clast density from 700 to 2580 kg m^–3. Observations of the 13 surge-producing blasts that preceded the climactic plinian event include radar-defined estimates of column heights and seismically defined eruptive and intra-eruptive durations. A comparison of the characteristics of erupted material, including microlite textures, chemical compositions, and H₂O contents, with eruptive parameters suggests that devolatilization-induced crystallization of the magma occurred to a varying extent prior to at least nine of the explosive events. Although volatile loss progressed to the same approximate level in all of the clasts analyzed (weight percent H₂O=1.26-1.73), microlite crystallization was extremely variable (0–22%). We infer that syn-eruptive volatile exsolution from magma in the conduit and intra-eruptive separation of the gas phase was facilitated by the development of permeability within magma residing in the conduit. Correlation of maximum microlite crystallinity with repose interval duration (28–262 min) suggests that crystallization occurred primarily intra-eruptively, in response to the reduction in dissolved H₂O content that occurred during the preceding event. Detailed textural characterization, including determination of three-dimensional shapes and crystal size distributions (CSD), was conducted on a subset of clasts in order to determine rates of crystal nucleation and growth using repose interval as the time available for crystallization. Shape and size analysis suggests that crystallization proceeded in response to lessening degrees of feldspar supersaturation as repose interval durations increased. We thus propose that during repose intervals, a plug of highly viscous magma formed due to the collapse of vesicular magma that had exsolved volatiles during the previous explosive event. If plug thickness grew proportionally to the square root of time, and if magma pressurization increased during the eruptive sequence, the frequency of eruptive pulses may have been modulated by degassing of magma within the conduit. Dense clasts in surge deposits probably represent plug material entrained by each subsequent explosive event. Received: 4 December 1997 / Accepted: 13 September 1998     

The caldera-forming eruption of Ksudach volcano about cal. A.D. 240: the greatest explosive event of our era in Kamchatka, Russia

The largest Plinian eruption of our era and the latest caldera-forming eruption in the Kuril-Kamchatka region occurred about cal. A.D. 240 from the Ksudach volcano. This catastrophic explosive eruption was similar in type and characteristics to the 1883 Krakatau event. The volume of material ejected was 18–19 km³ (8 km³ DRE), including 15 km³ of tephra fall and 3–4 km³ of pyroclastic flows. The estimated height of eruptive column is 22–30 km. A collapse caldera resulting from this eruption was 4 × 6.5 km in size with a cavity volume of 6.5–7 km³. Tephra fall was deposited to the north of the volcano and reached more than 1000 km. Pyroclastic flows accompanied by ash-cloud pyroclastic surges extended out to 20 km. The eruption was initially phreatomagmatic and then became rhythmic, with each pulse evolving from pumice falls to pyroclastic flows. Erupted products were dominantly rhyodacite throughout the eruption. During the post-caldera stage, when the Shtyubel cone started to form within the caldera, basaltic-andesite and andesite magma began to effuse. The trigger for the eruption may have been an intrusion of mafic magma into the rhyodacite reservoir. The eruption had substantial environmental impact and may have produced a large acidity peak in the Greenland ice sheet.     

The 1669 eruption at Mount Etna: chronology,petrology and geochemistry,with inferences on the magma sources and ascent mechanisms

 Analysis of the petrochemical characters of the 1669 Etnean lavas shows that they can be grouped into two sets: SET1 lavas were erupted from 11 to 20 March and are more primitive in composition than SET2, erupted later until the end of activity. Both sets may be interpreted as the result of crystallization under different conditions of two primary magmas which are compositionally slightly distinct and which fractionate different volumetric proportions of minerals. To explain why more mafic lavas (SET1) were erupted earlier than more acid ones (SET2), we argue that new deeper magma rose up into a reservoir where residing magma was fractionating. Density calculations demonstrate that new magma is less dense and may originate a plume, rapidly rising through the residing magma which is cooler and more volatile-depleted than the new magma. Calculations of uprise velocity assuming laminar flow are consistent with this hypothesis. Received: 20 November 1995 / Accepted: 2 August 1996     

水氡动态图强震危险区预测的新方法

水氡动态图强震危险区预测法是一种地下流体的地震分析预报新方法。 对该方法做了系统的介绍,其内容有: 水氡基值变化率动态图像的生成方法,水氡变化率高值异常区图形演化与地震震中位置的关系,强震危险区预测的标志与方法。检验性预测结果表明,该方法具有预测效能,可在实际地震预报中使用。     

On the fractal dimension of the fallout deposits: A case study of the 79 A.D. Plinian eruption at Mt. Vesuvius

We investigated the existence of a fractal law (power law) distribution of size pyroclastic fragments erupted during the fallout phase of the 79 A.D. Plinian eruption at Mt. Vesuvius. In particular, we performed a particle size distribution analysis on 18 white and grey pumice samples collected in six sites distributed in the SW sector of Mt. Vesuvius. Our measurements show that the fragmentation of samples in the investigated range (from 32 mm to 850 μm) follows a power law, guaranteeing the scale invariance of the process. The relationship frequency-size distribution of the fragments is verified independently from the nature (i.e., pumices and lithics) and stratigraphic height of the considered samples in the pyroclastic deposit. Therefore, the fractal fragmentation theory can be indicated for evaluating the relationship between the intensity of fragmentation (fractal dimension D) and eruption energy. In this way the apparent chaotic distribution of the particles in the fallout deposits hides a self-organized complexity revealed by the retrieved power law distribution. We further remark that a key aspect of our analysis is the founded evidence that the fractal dimension of the lithics is systematically greater than that of the pumices.     

Tephra-fall deposits from the 1992 eruption of Crater Peak, Alaska: implications of clast textures for eruptive processes

 The 1992 eruption of Crater Peak, Mount Spurr, Alaska, involved three subplinian tephra-producing events of similar volume and duration. The tephra consists of two dense juvenile clast types that are identified by color, one tan and one gray, of similar chemistry, mineral assemblage, and glass composition. In two of the eruptive events, the clast types are strongly stratified with tan clasts dominating the basal two thirds of the deposits and gray clasts the upper one third. Tan clasts have average densities between 1.5 and 1.7 g/cc and vesicularities (phenocryst free) of approximately 42%. Gray clasts have average densities between 2.1 and 2.3 g/cc, and vesicularities of approximately 20%; both contain abundant microlites. Average maximum plagioclase microlite lengths (13–15 μm) in gray clasts in the upper layer are similar regardless of eruptive event (and therefore the repose time between them) and are larger than average maximum plagioclase microlite lengths (9–11 μm) in the tan clasts in the lower layer. This suggests that microlite growth is a response to eruptive processes and not to magma reservoir heterogeneity or dynamics. Furthermore, we suggest that the low vesicularities of the clasts are due to syneruptive magmatic degassing resulting in microlitic growth prior to fragmentation and not to quenching of clasts by external groundwater. Received: 5 September 1997 / Accepted: 1 February 1998     

Temporal variations in gravity at Mt. Etna (Italy) associated with the 1989 and 1991 eruptions

 Results are presented from 11 microgravity surveys on Mt. Etna between 1987 and 1993, a period including the major 1989 and 1991–1993 flank eruptions and subordinate 1990 activity. Measurements were made with LaCoste and Romberg D-62 and D-157 gravity meters along a network around the volcano between 1000 and 1900 m a.s.l. and, since 1992, a N–S summit profile. Gravity changes of as much as 200 μGal were observed at scales from the size of the summit region to that of the volcano. None was associated with significant changes in ground elevation. The data show an increase in gravity for 2 years before the 1989 eruption. The increase is attributed to the accumulation of magma (0.25–1.7×10⁹m³) in an elongate zone, oriented NNW–SSE, between 2.5 and 6 km below sea level. Part of this magma was injected into the volcanic pile to supply the 1989 and 1990 eruptions. It also probably fed the start of the 1991–1993 eruption, since this event was not preceded by significant gravity changes. A large gravity increase (up to 140 μGal) detected across the volcano between June and September 1992 is consistent with the arrival in the accumulation zone of 0.32–2.2×10⁹m³ of new magma, thus favoring continued flank effusion until 1993. A large gravity decrease (200 μGal) in the summit region marked the closing stages of the 1991–1993 event and is associated with magma drainage from the upper levels of Etna's central feeding system. Received: 15 July 1995 / Accepted: 27 October 1997     

The Bag Tephra, a widespread tephrochronological marker in Middle Europe: chemical and mineralogical investigations

 The Bag Tephra is a widespread tephra layer interbedded in Quaternary loess deposits along the Danubian valley of Hungary and Slovakia. Its age is poorly defined between 788 and 380 ka B.P. The glass and mineral composition – micropumice clasts of phono-tephrite and blocky shards of tephri-phonolite associated with two kinds of clinopyroxene, fassaitic diopside, and salite – is very distinctive. This tephra could be used as a chronological marker, as soon as its age is refined. The probable origin is the middle Italian volcanic area. Received: 3 November 1998 / Accepted: 18 January 1999     

Pyroclastic deposits as a guide for reconstructing the multi-stage evolution of the Somma-Vesuvius Caldera

 The evolution of the Somma-Vesuvius caldera has been reconstructed based on geomorphic observations, detailed stratigraphic studies, and the distribution and facies variations of pyroclastic and epiclastic deposits produced by the past 20,000 years of volcanic activity. The present caldera is a multicyclic, nested structure related to the emptying of large, shallow reservoirs during Plinian eruptions. The caldera cuts a stratovolcano whose original summit was at 1600–1900 m elevation, approximately 500 m north of the present crater. Four caldera-forming events have been recognized, each occurring during major Plinian eruptions (18,300 BP "Pomici di Base", 8000 BP "Mercato Pumice", 3400 BP "Avellino Pumice" and AD 79 "Pompeii Pumice"). The timing of each caldera collapse is defined by peculiar "collapse-marking" deposits, characterized by large amounts of lithic clasts from the outer margins of the magma chamber and its apophysis as well as from the shallow volcanic and sedimentary units. In proximal sites the deposits consist of coarse breccias resulting from emplacement of either dense pyroclastic flows (Pomici di Base and Pompeii eruptions) or fall layers (Avellino eruption). During each caldera collapse, the destabilization of the shallow magmatic system induced decompression of hydrothermal–magmatic and hydrothermal fluids hosted in the wall rocks. This process, and the magma–ground water interaction triggered by the fracturing of the thick Mesozoic carbonate basement hosting the aquifer system, strongly enhanced the explosivity of the eruptions. Received: 24 November 1997 / Accepted: 23 March 1999