1 Introduction

The last eruption of Vesuvius [latitude (dd): 40.82, longitude (dd): 14.43, elevation (m): 1281] in March 1944 closed a 300-year-long period of intense eruptive activity that began immediately after the 1631 AD sub-plinian event and was characterized by moderately explosive eruptions and effusive activity (Arrighi et al. 2001; Scandone et al. 2008). The 1944 eruption was described in detail by the director of Vesuvius Observatory Prof. Giuseppe Imbò (1949a, b) who observed and monitored the eruption from the observatory located on the western slope of the volcano at an altitude of 608 m. The eruption began with a modest effusive activity and then moved on to a lava-fountain phase that quickly culminated in a paroxysmal phase, during which the eruptive column reached its maximum altitude (more than 5 km) and the ash carried by the wind arrived at distances exceeding 470 km eastward from Vesuvius.

After 70 years, the eruption is still present in the memory of people living in the Vesuvian area, who were direct victims of the event, but also of those who lived in the area affected by the ash cloud not only in Campania, but also in the regions of Molise, Basilicata and Puglia.

In this paper we perform a magma evolution model based on the geochemical compositions (concentration of major and volatile elements) of products (minerals and volcanic glass) emitted throughout the stages of the eruption in order to reconstruct pre- and syn-eruptive magmatic processes. Moreover, through the use of a numerical model of volcanic ash dispersion in the atmosphere as well as the collection of new accounts from witnesses, we infer the eruptive mechanism and its impact on the environment and the population. This information has major implications for assessing volcanic risk. Indeed, the sudden growth of the eruptive column during the course of the eruption may cause an increase in the particle size and thickness of the pyroclastic deposits (lapilli) accumulating on the ground and on the roofs, with a major impact on roof collapse, rail networks, etc., as well as, particularly in distal areas, an increase in the area exposed to ash fall, with potentially severe repercussions on agricultural crops over wide areas and extensive loss of livestock as a consequence of fluorosis. Moreover, a large amount of ash dispersed in the atmosphere can seriously disrupt air traffic.

1.1 The Vesuvius eruption in March 1944

The eruption in March 1944 was studied and described by various researchers at the time (e.g., Imbò 1945; 1949a, b, 1952, 1956; Parascandola 1945, 1948; Lazzari 1948; Casertano 1949; Scherillo 1949, 1953) who represent the point of reference for modern studies of the different eruptive phases, the deposition of eruptive products and the evolution of the feeding system (Scandone et al. 1986, 2008; Santacroce 1987; Cubellis and Luongo 1997; Pesce and Rolandi 1994; Villemant et al. 1993; Trigila and De Benedetti 1993; Marianelli et al. 1999, 2005; Fulignati et al. 2000, 2004; Del Moro et al. 2001; Morgan et al. 2004; Pappalardo et al. 2004; Lima et al. 2006, 2007; Chester et al. 2007; Ventura and Vilardo 2008; Vilardo and Ventura 2008; Cole and Scarpati 2010; Cubellis et al. 2013, 2014; Pappalardo et al. 2014).

The most exhaustive description of the eruption was that of the Director of the Vesuvius Observatory Giuseppe Imbò (Imbò 1949a, b) who, despite the severe events that affected the historic edifice of the Observatory during World War II, decided to intensify his observations of the volcano. The allied forces had taken over a large part of the edifice despite Imbò’s concerns: “… […] But I tried to express not only my personal regret but indeed of all scientists, not excluding the American volcanologists, for the abrupt interruption of the observations just before a possible violent paroxysm. The Italo-American lieutenant responded laconically with a smile: Volcanology is not necessary for war” (Imbò 1949b).

He followed all the stages of this eruption without leaving the edifice: “All other US servicemen… during the night before the beginning of the most violent paroxysm… had suddenly gone away and with them almost all the residents of the area. […] The Captain told me that a car was available for me if I wanted to go away immediately. I did not hesitate to say that my place was in the Observatory and that I would not abandon my position […]” (Imbò 1949b).

Imbò distinguished four main phases (Fig. 1), briefly reported below.

Fig. 1
figure 1

Eruptive and seismic activity during the different phases of the Vesuvius 1944 eruption. Seismograms were registered by the Omori-Alfani-Malladra (O.A.M.) seismometer that in 1944 was operating at the Vesuvius Observatory; each line corresponds to 1 h (the represented segments correspond to 11 min for phases I and II–III and to 3 min for phase IV. Photographs come from the archive of Vesuvius Observatory

Full size image

Besides supplementing our knowledge of the literature, since the sixtieth anniversary from the eruption eye-witness accounts have been collected (Cubellis and Marturano 2010) and put to use in this study, some of which agree with what is already known while others have added new elements to the phenomenon.

1.1.1 Phase I: effusive

“At 16:30 local time on March 18th, after a few hours of conduit obstruction, there was an abrupt reopening with the start of an intense effusive and explosive activity” […] (from the communiqué on the eruptive event of Giuseppe Imbò and reported by the daily newspaper “Risorgimento” March 25, 1944, Naples).

The eruption began in the afternoon of March 18, 1944, with the overflow of magma from the crater and the development of two main lava flows, one heading SSE, while the second flowed first northward and later, after reaching the border of the Somma caldera, was channeled westward.

Imbò climbed to the summit of the volcano, up to the crater, where he observed a mild strombolian activity accompanying lava flows. On 19th March, the effusive and explosive activity became more intense, and the lava flows, with a speed ranging from 50 to 300 m/h (Imbò 1949b), threatened the towns of San Sebastiano and Massa (Fig. 2), on the western slope of the volcano, that were invaded in the early hours of the 21st. Shortly before, the Allied authorities had organized the evacuation of about 7000 people living in these towns. In the afternoon of the same day the flows slowed and finally stopped on March 22nd; the SSE lava flows had reached an altitude of 350 m, while those that invaded San Sebastiano had stopped at an altitude of 140 m.

Fig. 2
figure 2

The map shows the locations of the studied outcrops (red triangles: I Terzigno, II Palma Campania, III Vesuvius) and of the sites of ash accumulation (black dots) and/or locations of stories reported by eyewitness accounts. 1 Lavello (PZ), 2 Marsico Nuovo (PZ), 3 Viggiano (PZ), 4 Bari, 5. Conversano (Ba), 6 Noi Cattaro (Ba), 7 Terlizzi (Ba), 8 Sannicandro (Ba), 9 Binetto (Ba), 10 Foggia, 11 Brindisi, 12. Avellino, 13 Lapio (AV), 14 Montella (AV), 15 Bisaccia (AV), 16 Eboli (Sa), 17 Monte San Giacomo (Sa), 18 Campagna(Sa), 19 Castiglione dei Genovesi (Sa), 20 Cava dei Tirreni (Sa), 21 Devoli (Albania), 22 Portici, 23 Nocera, 24 Pagani, 25 San Sebastiano. 26 Massa di Somma, 27 Sala Consilina (Sa)

Full size image

The following are quotes from eyewitnesses newly compiled by the authors:

From Portici (NA) (N 93.4 W, Fig. 2).

… a loud rumble and a continuous shaking propagated in the air…; a column of smoke rose rapidly in the sky; smoke first white, then gray. After a few days, you could see impressive lava flows that moved sometimes slowly, sometimes quickly […] The continuous tremors and bubbling like water in the pot … were especially impressive (Pasquale Migliaccio, Naples 2011).

From Naples (N 75.3 W, Fig. 2).

“… in the afternoon of March 18th, the windows jingled and it felt like a storm, but instead from the road we heard ‘it is the Mountain.’ We lived in Piazza del Gesù. Then the situation became bad on the 19th and the following days […]. (Roberto De Simone, Naples 2013).

1.1.2 Phase II: lava fountains

This phase began approximately at 17:00 local time on March 21st with an intense explosive activity characterized by eight “lava fountains.” The volcanic fragments, falling on the summit of the volcano, formed a “pseudo-flow” (according Imbò 1949a) that appeared in the darkness of the night as “a black mass with red hot spots.” Following the description of Imbò (1949a), the first fountain lasted less than 1 h (fountain I: between 17.15 and 17.45), with peak activity in the first 15 min, followed by seven other fountains of variable duration (fountain II: 20:10–20:30, fountain III: 22:00–22:25, fountain IV: 01:40–02:10 (22 March), fountain V: 03:45–04:03, fountain VI: 05:35–06:15, fountain VII: 0630–07:05), while the last one lasted over 5 h (fountain VIII: 7:30–12:30 March 22). But this, rather than a single fountain, were a “succession of fountains separate by maximum and minimum activity.”

During the final phase of each fountain, an acoustic wave accompanied by a light arc was observed from the historical Vesuvian Observatory. This is likely the well-known phenomenon of a propagating shock wave, in which the temperature drop induced by rarefaction causes water vapor to condense into a visible arc.

Products associated with this explosive phase reached heights of up to approximately 4 km and, transported eastward by the wind, covered the towns around Vesuvius and the Agro Nocerino-Sarnese area, at distances of 15–20 km.

1.1.3 Phase III: mixed explosions

The end of the lava-fountain phase and the beginning of a new stage, called “phase of the mixed explosions,” occurred shortly after noon on 22nd March. The volcanic cloud reached a height of over 5000 m above the crater, with the emission of black ash associated with the expulsion of “explosive” bombs (according Imbò 1949a) and “electrical discharges.” Moreover, the partial collapse of the eruptive column generated small pyroclastic density currents sweeping at high speed (2–3 km/min) along the flanks of the Great Cone of Vesuvius down to its base. In this phase another important type of activity is the avalanching of the hot material accumulated on the cone as recognized by some authors (Hazlett et al. 1991; Chester et al. 2007). After a 3-h break in explosive and seismic activity, the second part of the mixed explosions began at 21:00 h of the same day with the emission of dark ash clouds from two contiguous vents and continued until the morning of 23rd March. This represented the climax of the eruption. Indeed, the column reached its maximum altitude above the crater and transported large quantities of ash that reached great distances from the vent. During the night of the 22nd and the morning of 23rd March at the oil field of Devoli (today Berat) in Albania (470 km from Vesuvius), a thin ash layer was deposited (Lazzari 1948).

During this phase of the eruption, 24 people died due to roof collapses caused by the accumulation of ash: twelve at Nocera, nine at Pagani and three at Terzigno (Fig. 2).

The following are quotes from eyewitnesses newly compiled by the authors:

From Naples (N 75.3 W, Fig. 2).

… I remember … it was about noon … we noticed a plume that was rising from Vesuvius and a dark cloud was expanding into the sky, a strange gray; it was very high. (Federico de Pandis, Riardo - CE, 2013).

On March 22nd, […] I remember the cloud of Mount Vesuvius was impressive: three times higher [about 4 km] than the volcano (Roberto De Simone, Naples 2013).

From what was then Yugoslavia (Fig. 2).

In March 1944 in Montenegro, Antivari (N 74.3 E), Porto Teodo and Budua (N 71.7 E), there was talk of ash falls from the Vesuvius eruption, and news about the event came from Radio Fante (Domenico Obrizzo, Naples, 2013).

From Devoli, Albania (N 91.6 E, Fig. 2).

On the night between 22nd and 23rd March, 1944, for an unspecified hour, at the AIPA oil shipyard in Devoli, Albania, a thin dark ash layer was deposited, I collected a sample thinking that it could be interesting to analyze (Lazzari 1948).

1.1.4 Phase IV: seismo-explosive

At about 12:00 noon on 23rd March, the explosions became discontinuous and the ash cloud reached altitudes not exceeding 2 km above the crater. Strong seismic activity was associated with the most intense explosive stages, while small pyroclastic density currents and landslides along the flanks of the Great Cone and electrical discharges in the atmosphere also characterized this phase of the eruption.

On the morning of 24th March, the eruptive activity had diminished and continued with less intensity until April 7th, when the crater was completely obstructed, giving rise to the current closed-conduit period of the volcano. This phase has been correlated to obstruction of the conduit and/or increase in magma viscosity.

1.2 Sampling and analytical methods

We examined representative samples of the deposits in order to study the chemical evolution of the eruption. The samples were collected from three stratigraphic sections located at different distances from the volcano: on the volcano’s slopes (lava samples, phase I), at Terzigno (6 km ESE from the Vesuvius) (scoria samples from phases II and III), and at Palma Campania (12 km ENE from the Vesuvius) (scoria samples from both phases III and IV), as well as one ash sample, representing an extreme distal facies, collected at the time of the eruption in Albania (at 470 km E of Vesuvius, Fig. 2) by the geologist Antonio Lazzari (2013, 2014) who was there for oil exploration.

At Terzigno the sampled stratigraphic section is composed of a basal fallout deposit, 20–25 cm thick formed essentially of juvenile well-vesiculated scoria fragments (Mdφ = −3, σϕ = 1) (phase II). It is overlain by a 50-cm-thick, finer lapilli deposit composed of both lithic clasts and dense scoria fragments (phase III) with a mean size of Mdφ = −2, σϕ = 1.

At Palma Campania the sampled stratigraphic section is composed of a thin scoria layer overlain by an ash-rich deposit (phase IV). The last is a massive, 20 cm thick, well-sorted σϕ = 1–1.15 (clasts size between 0 and 1ϕ) deposit.

1.2.1 Textural and geochemical analysis

Textural analyses were carried out on polished thin sections of epoxy-impregnated grains using a standard polarized light microscope. Analyses of major and volatile elements in groundmass glasses and mineral phases were performed with a JEOL-JXA-8200 electron microprobe (WD/ED combined micro analyzer) at the laboratories of the National Institute of Geophysics and Volcanology (Rome). Element concentrations were measured with a 10-μm beam at 15 keV, a beam current of 10 nA and a counting time of 10 s. For each analysis a defocused beam was used to minimize losses of alkalis and volatiles, which were counted first to avoid diffusion effects. We attempted to obtain at least ten analyses per sample. Data reduction was carried out using ZAF4/FLS software by Link Analytical. The analytical uncertainty was about 1 % for most elements. Results are reported in Table 1. The H2O content of all analyzed glass was estimated by using the “volatile by difference” method based on EMPA analyses. The last technique has been widely used in petrology when, as in our case, the small size of the volcanic glass not allow the application of other methods (for example FTIR).

Table 1 Representative electron microprobe analyses of phenocrysts and glass of rocks of the 1944 Vesuvius eruption
Full size table

2 Results

2.1 Textural analysis

Lava and scoria samples are porphyritic with euhedral phenocrysts consisting chiefly of olivine (<1 %), plagioclase (5 %), diopside (10 %) and leucite (25 %), while garnet (<1 %), and phlogopite (<1 %), are very common accessory phases. The groundmass is rich in microlites (up to 50 vol%) represented generally by abundant leucite, minor feldspar, pyroxene, amphibole, and mica. The distal ash sample is composed mainly of free minerals (70–80 %) and to a lesser extent of vesicular dark glass (20–30 %). The main mineral phases are clinopyroxene (30 %), leucite (20–30 %) and feldspar (20–30 %); less abundant olivine (5 %), magnetite (5 %), apatite (<1 %), titanite (<1 %) and garnet (<1 %), in accordance with what was reported also by Lazzari (1948). Similar composition has described by Scherillo (1953) for volcanic ash deposited at Brindisi during the same eruptive phase.

Ash is constituted by tabular or elongated clasts (aspect ratio = ~1.5 modal value; circularity = ~0.7 modal value), with moderate roundness (roundness = ~0.7 modal value) and size ranging between about 20 and 350 μm (Fig. 3).

Fig. 3
figure 3

Back-scattered electron (BSE) images of groundmass textural features in samples from the different stages of the Vesuvius 1944 eruption and of the volcanic ash collected at Devoli in Albania (470 km from Vesuvius) (lower-left) and histogram of the clast size (lower-right). The dimension of the clasts is between 20 and 350 μm, with a prevalence of smaller clasts (20 and 30 μm)

Full size image

In spite of the quite uniform phenocrysts content, the analyzed volcanic samples show different textural characteristics in microlites. The last feature has been deeply explored by Pappalardo et al. (2014) and briefly reported here. In particular, microlite crystals in lava (effusive phase I) and subordinately scoria (lava-fountains phase II) samples are composed of larger prismatic plagioclase (60–80 μm, 3.5–4ϕ) associated with smaller blocky clinopyroxene (30 μm, 5ϕ). While pumice and ash (sustained-column III phase and vulcanian IV phase), contain microlites represented by sparse smaller acicular clinopyroxene and in minor amount plagioclase (15–10 μm, 6ϕ), Fig. 3.

2.2 Mineral chemistry and glass composition

2.2.1 Feldspar

The feldspar compositions are reported in an albite–orthoclase–anorthite diagram in Fig. 4. Our data indicate that lavas, scoria and ash samples are characterized by plagioclase of bytownitic-labradoritic (An84–64) composition.

Fig. 4
figure 4

Chemical composition of the main mineral phases. a Ab–An–Or plot for feldspars; b En–Wo–Fs plot for clinopyroxenes; c K2O versus Na2O in leucite crystals and d Fo % versus Fa % in olivine. The figure shows that the chemical composition of phenocrysts is similar for all the studied products (lavas, scoria and ash) emitted during the different phases of the eruption. Gray symbols data from literature (Marianelli et al. 1999)

Full size image

2.2.2 Pyroxene

The composition of pyroxene phenocrysts are shown in the Enstatite–Ferrosalite–Wollastonite diagram (Fig. 4; Morimoto 1988). The analyzed pyroxenes fall in the diopside field for lava, scoria and ash samples and show a general trend toward Fe-rich composition (Fs4–Fs14).

2.2.3 Leucite and olivine

The phenocrysts of leucite show fairly homogeneous composition (K2O = 19.2–19.8 wt%). The olivine crystals have a content of forsterite of Fo71–90 (Fig. 4). In general, the chemical composition of phenocrysts is similar for all the studied products (lava, scoria and ash) emitted during the different phases of the eruption.

2.2.4 Volcanic glass

The groundmass glass has a composition ranging from tephrite to phonotephrite (Fig. 5) in the classification diagram of Le Bas et al. (1986). In Fig. 6 the most representative variation diagrams are shown for major elements (Harker diagrams) as a function of MgO, used as a magmatic differentiation index. In the Harker diagrams the total composition of the whole rocks (glass + crystals) is always less differentiated than that of the corresponding volcanic glass. With the decrease in MgO concentration, major elements describe regular patterns without compositional hiatus. In particular, we observe an increment of SiO2 (45–51 wt%), Al2O3 (13–22 wt%), Na2O (from 1 to 8 wt%), MnO (from 0.15 to 0.35), and a decrease in FeO * tot (from 11 to 7 %), CaO (14–6 %), while TiO2 (from 1 to 1.4 %) and P2O5 (from 0.6 to 1.4) slightly increase except for lava and scoria glass groundmass possibly due to their nearly complete microlite crystallization. By contrast, K2O increases in less differentiated rocks (MgO between 8 and 5 wt%), while it decreases in the samples with a higher degree of evolution (MgO between 5 and 2 wt%).

Fig. 5
figure 5

Total Alkali Silica classification diagram (Le Bas et al. 1986). The groundmass glass has a composition ranging from tephrite to phonotephrite

Full size image
Fig. 6
figure 6

Harker diagrams of major element compositions in whole rocks [open symbols data from Marianelli et al. (1999)] and volcanic glass (closed symbols data from this paper) emitted during the 1944 eruption. Symbols are as in Fig. 5. The regular patterns of major element concentrations versus the index of magmatic differentiation (MgO) indicate a gradual and continuous process of crystallization in the magmatic system. See text for further explanations

Full size image

These compositional variations can be attributed to the fractional crystallization of phenocryst phases observed in the rocks. In particular, the decrease in MgO, FeO * tot and CaO is consistent with the predominance of the fractionation of mafic minerals (rich in Mg and Fe, such as olivine and clinopyroxene) and calcic plagioclase, while the roughly constant concentrations of TiO2 and P2O5 indicate the crystallization of minerals rich in Ti and P (biotite, magnetite, apatite, spinel) as accessory phases. Furthermore, the reversal of the slope shown by K2O content (Fig. 6) which, from initial enrichment, decreases in samples less rich in MgO (<5 %) may indicate a late crystallization of leucite, in whose crystal lattice K is concentrated which is consequently subtracted in the more evolved liquid. The last result is in agreement with experimental data indicating that the crystallization of leucite in Vesuvius magmas occurred at the pressure <100 MPa and under rapid growth rate (Shea et al. 2009).

2.3 Magma evolution and conduit dynamics

To constrain the conditions of temperature and pressure at which the different mineral phases are crystallized, we used the chemical compositions of minerals and glass obtained by the electron microprobe in the thermodynamic calculations (Fig. 7) based on the minimization of Gibbs free energy (MELTS CODE, Ghiorso and Sack 1995) performed by Pappalardo and Mastrolorenzo (2010) on Vesuvius tephritic rocks. Furthermore, the pressure values obtained by using the concentrations of volatile phases (H2O, Cl, F) measured in volcanic glasses (which represent the fractions of residual degassed liquid in the conduit) were compared with those available in the literature (Marianelli et al. 1999; Fulignati et al. 2000) on melt inclusions (small drops of non-degassed magma) trapped in the crystals during their growth (Fig. 8), and on experimental studies (Pichavant et al. 2014).

Fig. 7
figure 7

Mineralogical assemblage (top) and magma (bottom) composition during the process of crystallization simulated using the program Melts following the procedure used by Pappalardo and Mastrolorenzo (2010). Symbols as in Fig. 6

Full size image
Fig. 8
figure 8

Composition of glass inclusions (magma chamber processes) trapped in the main mineral phases of the 1944 eruption (data: Marianelli et al. 1999; Fulignati et al. 2004) compared with the composition of volcanic glass (volcanic conduit processes). For volcanic glass analyzed in this study, the values of pressure are calculated on the basis of water content measured by “difference method.” See text for further details

Full size image

Our results indicate that crystallization is carried out in two distinct phases under different conditions of pressure and temperature. In the first phase the less evolved magma cools (T = 1150–950 °C) in a crustal magma chamber at depths between 16 and 8 km (P between >400 and 300 Mpa), confirming the data on melt inclusions (Marianelli et al. 1999; Fulignati et al. 2000). The same phase crystallizes mainly olivine and pyroxene, and to a lesser extent plagioclase, magnetite and apatite (Fig. 8). A second crystallization phase occurs with lower pressure (<200 Mpa) and temperature (<950 °C), probably during the ascent of the magma in the volcanic conduit, in which the leucite crystallizes in equilibrium with the evolving magmatic liquid (Fig. 8). The obtained data agree with experimental results showing that the primitive Vesuvius trachybasalts crystallize Fo-rich olivine as the liquidus phase between 1150 and 1200 °C and from 300 to <200 MPa (Pichavant et al. 2014).

Information on the ascent rate of the magma in the volcanic conduit during the different phases of the eruption can be deduced from the volatile content measured in volcanic glass (Fig. 9), whose composition corresponds to that of the residual magma during degassing in the volcanic conduit. In particular, the low water contents (H2O wt% ~0.5–1.6) which characterize the samples of lava emitted during the effusive phase are compatible with a slow decompression of the magma in the volcanic conduit that favors volatile exsolution, bubble diffusive growth and coalescence (that controls magma permeability) and then removal of gas (including H2O) from the liquid (near-equilibrium degassing path). In contrast, the highest water content (H2O 1.5–3 wt%) in scoria and ash of the explosive phases of the eruption is indicative of more rapid rising in an unlocked volcanic conduit that prevents the diffusion of the gas phase, which remains dissolved in the melt until explosive fragmentation occurs (disequilibrium degassing path). Actually, during decompression experiments both degassing paths were observed in Vesuvius magma (Marziano et al. 2007); particularly in the case of higher decompression rate, magma ascent is characterized by high water over- saturations under disequilibrium condition, while for slower ascent rates degassing of the melt by diffusive bubbles growth prevails, thus sensibly reducing the water over-saturation in the melt.

Fig. 9
figure 9

a Water content versus F, b CaO versus F and c CaO versus Cl in the volcanic glass of the different phases of the eruption. The lava emitted during the first effusive phase shows the lowest water content, compatible with a slow ascent of the magma in the conduit, which favors exsolution and the removal of volatiles including mainly H2O (equilibrium degassing); by contrast, glass samples of explosive phases show higher water content indicating faster rising, hindering the exsolution and escape of magmatic gas (disequilibrium degassing)

Full size image

Different degassing mechanism during rising in volcanic conduit is suggested also by the trend of Cl and F contents in function of CaO wt%, chosen as index of differentiation (Fig. 9b, c). The figures show that F and Cl contents increase in lavas with the degree of differentiation of rocks while remain constant in ash and scoria. Signorelli and Carroll (2000) have shown that Cl increases within melts, saturated with both a H2O vapor phase and a Cl-rich brine, as pressure decreases. The lavas of the phase I follow this Cl solubility model, corroborating the hypothesis of an initial slow ascending magma, while scoria and ash of the following phases show a constant Cl concentration, indicating a fast ascent rate which prevents fluid/melt Cl exchange.

2.4 Estimation of the maximum height of the eruptive column during the paroxysmal phase

Historical chronicles report that during the paroxysmal phase of the eruption, between March 22nd and 23rd, the volcanic ash reached great distances from the vent, as far as Devoli in Albania (470 km from Mount Vesuvius), and that the eruptive column reached 5000 m in altitude. However, this altitude estimate is not compatible with the long distance reached by the ash. Indeed, Chester et al. (2007) assumed that the cloudiness did not allow the witnesses to accurately determine the maximum height reached by the column. To test the hypothesis of Chester, we consulted the archives of the weather station of the Geophysics Institute of the University of Naples located at the city center. Data show that the sky on March 22nd 1944 had a cloud cover of 3/10 in the early afternoon (14:00), while in the evening (19:00) there was total cover (10/10), which lasted throughout the following day, thus not allowing observation of the eruptive column in its full development. Therefore, to constrain the maximum height reached by the eruptive column during the paroxysmal phase of the eruption, we adopted a numerical model of volcanic ash dispersion into the atmosphere (Mastrolorenzo et al. 2008), using both the dynamic parameters (initial velocity, erupted mass, ash density and size) of the eruption phase (Scandone et al. 1986; Pfeiffer et al. 2005; Macedonio et al. 2008) and the measured wind speed in March 1944 at different altitudes up to 12,000 m (Fig. 10 database of NOAA’s National Geophysical Data Center, http://www.ngdc.noaa.gov/).

Fig. 10
figure 10

Wind speed in the direction of the parallel (to the east), measured at different altitudes on 22 and 23 March 1944, during the paroxysmal phase of the eruption in which the arrival of volcanic ash was documented in Albania

Full size image

It is possible that the volcanic column has acted as a “weak” plume at least in the second phase of the eruption as suggested by photos taken at that time showing a few km high plume following bent-over trajectories strongly controlled by the wind strength and direction. While it possibly has switched to “strong” plume during the paroxysm of the third phase (for which unfortunately no photos are available) when it reached its maximum height and transported ash until Albania and Yugoslavia. In our model we have considered exclusively this second behavior to calculate the maximum height of the column necessary to justify the falling of ash, so far from the volcano during the third eruptive phase.

The results of the simulations, shown in Fig. 11, indicate that the eruptive column reached a height of around 10 km to justify the accumulation of a thin layer (about 1 mm) of ash in Albania approximately 470 km away from Mount Vesuvius. The data obtained are in agreement with those calculated by other authors from the analysis of lithic and scoria isopleths (Cole and Scarpati 2010). On the basis of the obtained height of the column and of the data on wind speed at different altitude (Fig. 10), it is possible to calculate that the ash, carried by high-altitude winds toward east with a mean speed between 50 and 70 km/h, has to have travelled for 7–9 h before reaching the Albania between the night of the 22nd and the morning of 23rd March as reported by Lazzari (1948).

Fig. 11
figure 11

Numerical simulation of dispersion of volcanic ash in the atmosphere carried out using the dynamic parameters of the 1944 eruption (initial velocity = 200 m/s, erupted mass = 0.4 × 1012 kg, average density = 1000 kg/m3 and average size of the ash ϕ = 1.5) and the wind speed measured in March 1944. The figure shows that, for a column height of 5 km, the ash reaches about 200 km away from the volcano; for a height of 8 km, ash arrives at 300 km away, while for an eruptive column higher than 10 km, the distance exceeds 400 km. Thickness of ash resulting from direct accounts is also reported. For distances below 100 km, the observed thicknesses are consistent with a cloud height of about 5 km, while the thicknesses accumulating at greater distances reflect a column height between 8 and 10 km

Full size image

Figure 11 also shows the thicknesses of the accumulated ash in the third eruptive stage along the dispersion axis, as testified by numerous witness reports collected in the last 10 years in a wide area (Cubellis and Marturano 2010, Cubellis et al. 2013), according to which the ash fall reached considerable distances from the volcano (Table 2).

Table 2 Thickness of volcanic ash emplaced during the phase III
Full size table

Data in Fig. 11 indicate that, for distances below 100 km from Vesuvius, ash thicknesses are consistent with a cloud height of about 5 km, while those accumulating at greater distances reflect a column height of between 8 and 10 km. This apparent bimodality could be explained by a rapid change in eruptive regime during column development. According to Pappalardo et al. (2014), during phase III the partial collapse of the deeper parts of the plumbing system caused a pressure drop at the top of the underlying chamber and then the rapid extrusion of a larger volume of magma in a short time. This process could explain the rapid increase in the height of the eruptive column during the paroxysmal phase and hence the bimodal distribution, observed in Fig. 11, relative to the thickness of ash accumulated during the phase III, due to the discontinuous growth of the eruptive column as well as in the proximal area (sections located <5 km from the crater) the well-developed reverse to normal grading, as described by Cole and Scarpati (2010).

2.5 Impact on the environment and people

The real impact of the last Vesuvius eruption on the environment and people is still now not well evaluated, possibly due to the extensive destruction of Neapolitan area and the death of tens of thousands of civilians related to the World War II, as recently speculated on the basis of new historical reconstructions (Cole and Scarpati 2010; Romano 2011; Cubellis et al. 2013, 2014; Cubellis and Marturano 2014) and confirmed by the eyewitness accounts of the eruption.

The following are quotes from eyewitnesses recently compiled by the authors:

From Naples:

Remembering after over 70 years an exceptional yet dramatic event like the 1944 eruption of Vesuvius arouses in me a strange emotion … […] We were living in difficult times, the war with all its horrors had overrun Naples, […] we were still suffering from hunger and the consequences of the bombings … without forgetting the outbreak of epidemic typhus.

[…] news came that there were ‘several dead’. But even this news, dramatic though it was, I don’t remember especially scaring me. … we were perhaps used to hearing about deaths due to bombings or in action on the various frontlines. (Ugo Foà, Rome, 2015)

Our new data on both the development of the eruptive column (previous paragraph) and the new witness reports allowed to add information on the effects of the eruption on the environment and human activities. Particularly, our calculation of the maximum height of the eruptive column and the consequent dispersion of ash over a wide area can justify the extensive loss of livestock documented by the new eyewitness accounts, also at great distances from the volcano (not less than 120 km). In fact the death of sheep, lamented by most of the survivors, was probably the consequence of fluoride poisoning (fluorosis) procured by the intake of grass from pastures covered with ash, in analogy with what has been observed for other historical and recent eruptions in other parts of the world. Particularly, in 1693, the first fluorine intoxication of animals due to a volcanic eruption was reported from Iceland after eruption of Hekla volcano (Thorarinsson 1979). A similar behavior was observed at Hekla again in the 1970 (Georgsson and Pétursson 1972). In 1783, ashfall and gases from Lakagigar volcano, once again on Iceland, caused the death of tens of thousands of sheep, cattle and horses (Cronin et al. 2003).

More recently, in 1988, about 10,000 farm animals were affected by fluoride-bearing ashes from the Lonquimay volcano eruption in Chile, killing more than 4000 head (goats, sheep, cattle and horses), while ashfall from Ruapehu volcano eruptions caused the death of thousands of sheep in 1995–1996 probably by fluorosis (Armienta et al. 2011).

To prove our hypothesis, we have measured the content of F in the volcanic products emitted during the course of the eruption, the obtained concentration actually is above the threshold of lethality for livestock. The amount of F in the lava, scoria and ash of the 1944 eruption is shown in Fig. 12. The amount considered lethal to grazing animals is at least 100 μg/g body weight (NRC, 1974) (4–5 g for animals weighing 40–50 kg), while values close to 50 μg/g seem to be tolerated by sheep and cattle. Fluoride concentrations of 2000 ppm (0.2 %) and Cl/F ratios similar to those reported in Fig. 12 of ash from explosive eruptions caused acute fluorosis and mortality in sheep during the eruption of the volcano Hekla in 1970 (Georgsson and Pétursson 1972). With these F concentrations, a layer of ash even one mm thickness is sufficient to cause the death of animals. In fact, following the reasoning of Dawson et al. (2010), we consider that the 1944 ash contains 2 g fluoride kg−1 ash (Fig. 12) with a density of 1 g/cm3, a 1-mm deposit covering 1 m2 would represent 1 kg of ash, equivalent to 10 tonnes ash ha−1. A sheep eating 1.6 kg dry mass herbage day−1 on a pasture at 4000 kg dry mass herbage ha−1 (Dawson et al. 2010) would be eating the equivalent of 4 m2 of herbage, and 4 kg volcanic ash (8 g fluoride), if all of the ash was deposited on the herbage. Under such circumstances, the sheep would receive fluoride at very high toxic levels.

Fig. 12
figure 12

F and Cl concentrations in lava, scoria and ash samples of the 1944 Vesuvius eruption

Full size image

The following are quotes from eyewitnesses recently compiled by the authors:

From Marsico Nuovo (PZ) (N 111.9 E, Fig. 2).

11 of 12 of them [goats] died after a few days of each other. The survivor, the only non-pregnant animal, lost considerable weight. Even the flocks of neighbors suffered more or less the same fate. (Raffaele Vignola, 2014).

From Sala Consilina (SA) (N 115.3 E) (Fig. 2).

[…] In the town (about 1700 inhabitants, 668 m a.s.l.) … it happened that grazing livestock died, mostly because they had eaten grass mixed with ‘charcoal’ (that was what the volcanic ash were called). With sheep, goats, lambs etc. a cow also died, again for the same reason. These animals all died with enormously bloated stomachs. (Italo Taccelli, Sala Consilina, SA, 2012).

Besides the loss of livestock also the number of casualties may be greater than known until now.

Reports of the Allied Military Government (Daily newspaper “Risorgimento” March 26th 1944 Naples) indicate that during the eruption 26 people died (12 at Nocera, 9 at Pagani, 3 at Terzigno) for the collapse of the roofs; also, two children from San Sebastiano and Massa di Somma died from the explosion of a tanker due to the passage of lava flows. The largest number of victims is due to the paroxysmal phase, as shown by numerous death certificates from the parishes of Nocera and Pagani. That said, the number of victims could have been even higher than officially reported at the time of the eruption. Indeed, because of the war some deaths went unreported.

The following are quotes from eyewitnesses newly compiled by the authors:

From Terzigno (NA) (N 104.7 E, Fig. 2).

[…] “The rain of this material [ash and lapilli] in the roads and on the roofs prevented normal traffic and threatened the stability of the dwellings.

This is why my brothers and my father started to clear the roof of our house on Via Roma. On the third day of the eruption, during the morning, my brothers were on the roof and my father, realising that it had begun to rain [ash and lapilli] once again, wanted to call them back. But while he was looking out of the front door, he was hit on the temple by a hot stone (a volcanic bomb) the size of a tennis ball and died on the spot.

The difficult moment and the poor state of the streets prevented the body being transported from our house to the cemetery”. […] (Maria Rosaria Fabbrocini, Ottaviano, 2014).

3 Discussion

Our results allowed us to reconstruct the eruptive dynamic during the last Vesuvius eruption in March 1944. In particular, the chemical compositions of volcanic rocks indicate that the magmas feeding the four phases of the eruption have similar geochemical and petrological imprints, such that the transition to the paroxysmal phase of the eruption is not related to changes in magma pre-eruptive conditions but instead to rising processes in the volcanic conduit.

Thermodynamic calculations based on the chemical equilibrium between crystals and liquid indicate that magmatic differentiation took place in two stages of crystallization under different temperatures and pressure conditions. In a first phase of higher pressure, the cooling of less evolved magma in a crustal magma chamber (between 16 and 8 km depth) caused the crystallization mainly of olivine and diopside and to a lesser extent plagioclase, magnetite and apatite; while in the later lower pressure stage (depth < 8 km), probably associated with the ascent of magma in the volcanic conduit, crystallized mainly leucite in equilibrium with the evolving residual liquids. This process of leucite syn-eruptive crystallization in the volcanic conduit, associated with a more rapid ascent of magma as suggested by our data on volatile concentrations in the melts, could explain the emission of magma richer in gas, and hence the sudden increase in the degree of explosiveness of the eruption in the intermediate stages (phases II and III), with the emission of a greater amount of ash. Other authors (Marianelli et al. 1999), on the basis of melt inclusions data, have recognized that the magmas feeding the eruption underwent differentiation at different pressures and suggested that the paroxysmal phase was fueled by the arrival of new, volatile rich, magma from the depth.

However, the increase of the ascent rate during the course of the eruption, has been recently documented by Pappalardo et al. (2014) trough the study of groundmass microlite textural variations (microlite number density MNDs and size distributions MSDs). Particularly, these authors have shown a progressively increase of the value of MNDs as well as of the slope of CSDs from lavas (MNDs = 11,673; slope = −30) and scorias (MNDs = 10,895–25,803; slope = −120), erupted, respectively, during the phase I and phase II, to pumices and ashes from the phase III and phase IV (MNDs = 40,000–35,000; slopes = from −175 to −608). These data allowed to demonstrate that decompression became gradually faster during the eruption, ranging from few days at the beginning of phase I up to few hours during the most violent paroxysm (phase III), with the development of a high eruptive column.

Historical chronicles report a height of the eruptive column during the paroxysm of about 5000 m. However, this value is hardly compatible with the high distance of ca. 500 km from the volcano that was reached by the volcanic ash. Possibly the presence of cloudiness on March 22–23rd 1944, did not allow the witnesses to accurately determine the maximum height reached by the column. In fact, our analysis of the data of the weather station of the Istituto di Geofisica of the University of Naples, in operation in 1944, showed that, from the afternoon of 22 March (14.00 coverage 03.10; 19.00 coverage 10/10) until the next day, the sky was completely covered with clouds from a height of about 1–2 km from the ground, thus making it difficult the observation of the eruptive column in all its development.

Adopting a numerical model of ash dispersion into the atmosphere, by using as input data both the dynamic parameters of the eruption and wind speed at various altitudes up to 12,000 m, we estimated that the eruptive column would reach a height of 10 km, higher than that previous supposed on the basis of the contemporary chronicles.

The last result justifies the dispersion of ash over wide areas as well as the extensive loss of livestock reported by the new witnesses accounts. The death of the sheep was probably the consequence of F (fluoride) poisoning (fluorosis) procured by the ingestion of grass, from pastures covered with ash, in analogy with what has been observed for other historical and contemporary eruptions in other parts of the world. The amount of F considered lethal to grazing animals is at least 100 μg/g body weight (NRC, 1974) (4–5 g for animals weighing 40–50 kg), while values close to 50 μg/g seem to be tolerated by sheep and cattle. We have measured F concentrations between 1500 and 3000 ppm in the 1944 Vesuvius volcanic ash. The obtained values indicate F content always higher than the lethality threshold; in fact, this F concentration in a layer of ash of even one mm thickness is sufficient to cause the death of the animals.

The 1944 deposit has been compared with those of other historical eruption as the Fuego 1974 and the Vesuvius 512 and 1906 AD deposits (e.g., Cole and Scarpati 2010); however for these events not much data are available on their environmental and human impacts. Thus to have an idea of the huge impact that a moderate eruption can have on the modern society, an interesting parallel emerges between this historic eruption and that of the Eyjafjallajökull volcano in Iceland in April 2010, the most recent in Europe (Gudmundsson et al. 2010; Dellino et al. 2012; Cioni et al. 2014) that dispersed a moderate amount of ash over much of Europe.

Due to the significant impact on international air traffic and the implications on the vulnerability of modern society, the Eyjafjallajökull eruption has become a landmark event. The succession of phases, with the first fire-fountains and fissure-fed lava flows followed by explosive activity, the height of the eruption column (5–10 km), the duration of the paroxysmal phase with winds that pushed the ash to the SE, the volume of emitted products into the atmosphere (5 * 106 m3), the average size of the clasts (40 μm) and the thickness of the deposit (~1 mm) to the Faroe Islands (a ~300 km away from the crater) are all comparable values with those related to the distal ash erupted during the paroxysm of Vesuvius in 1944. Of particular interest is the satellite view of the phenomenon in which it is possible to assess the extent of the ash cloud (>2000 km) that goes far beyond the limits of accumulation of products detected on the ground. However, the evolution of the eruption toward more explosive dynamics triggered by a sudden change in the composition of the magma and amplified by magma–water interaction processes/ash recycling (Iceland) and/or syn-eruptive crystallization (Vesuvius) characterizes both eruptions.

4 Conclusions

Our merged results on the 1944 eruption of Vesuvius, document the sudden increase in the degree of explosiveness during the middle phase (III mixed explosions) of the eruption, with the emission of a large amount of ash that was transported up to several hundred kilometers away from the volcano with strong impact on environment and people. In few hours, the eruption column reached a maximum height of at least 10 km on the crater, much greater than 5 km reported by the witnesses of the eruption due to the bad weather conditions preventing the direct observation of the eruptive cloud. Such sudden increase in the intensity of the volcanic scenario during ongoing eruptions, as shown by our study, should be taken into account as a precautionary measure for planning purposes. In fact, such a case of a moderate-intensity eruption of Vesuvius, similar to that which occurred in 1944, would be extremely disruptive, especially in our complex modern society.