Vesicularity of various types of pyroclastic deposits of Campi Flegrei volcanic field: evidence of analogies in magma rise and vesiculation mechanisms

doi:10.1016/S0377-0273(00)00303-6

Journal of Volcanology and Geothermal Research

Volume 109, Issues 1–3, 30 August 2001, Pages 41-53

https://doi.org/10.1016/S0377-0273(00)00303-6 Get rights and content

Abstract

In order to investigate the factors controlling the style of explosive eruption in Campi Flegrei volcanic field, image analyses of ash collected in all types of magmatic and hydromagmatic deposits have been carried out. Our results indicate that bubble shape, size distribution and number density are not correlated with the eruptive styles. Moderate ash vesicularity and relatively high bubble number density suggest that in most cases magma rose with moderate gas oversaturation and in critical conditions, near the transition between viscosity-controlled and diffusivity-controlled regimes. In these conditions, melt viscosity limited the bubble growth causing fragmentation to occur under the typical limits of gas volume fraction (75–85%) for foam rupture. The scarce deviation in vesicularity between pure magmatic and phreatomagmatic ashes suggests that the bubble growth was always arrested, although not completed, before the ultimate occurrence of magma–water interaction and eruption. Thus, magma flow rate and upper conduit events rather than conspicuous differences in magma vesiculation and ascent rate background likely controlled the eruptive style.

Introduction

The vesicularity of juvenile clasts in pyroclastic deposits is a primary source of information on the gas exolution, and hence on the magma-rising processes from the chamber to the surface, as well as on fragmentation and magma–water interaction. Various authors have highlighted different processes by which exolution, expansion and migration of bubbles may trigger an eruption (Pyle and Pyle, 1995). A widely accepted general model of the magma vesiculation implies three stages: nucleation, bubble growth by diffusion and bubble expansion by decompression. Theory and experiments indicate that in ascending magma, near the totality of nuclei are formed in a short time lapse at the gas saturation level. In ideal conditions, the total number of bubbles remains constant after the nucleation event, while the bubble number density (number of bubble for unit volume) decreases due to the further bubble expansion (Toramaru, 1989). Bubble growth by gas diffusion occurs in different conditions of gas saturation and chemical equilibrium, while bubble expansion for pure decompression occurs, in nearly isothermal conditions, dominates in the upper conduit and culminates in an explosive burst.

In the last few years, the investigation of the factors controlling the eruptive processes has benefited by numerical modelling, which provides insight on the relations between rising dynamic, eruption intensity and magma vesiculation (Gardner et al., 1996, Kaminski and Jaupart, 1997, Proussevitch et al., 1993, Proussevitch and Sahagian, 1996).

However, data from volcanic products are the only real constraints to the models. Direct density measurements of pyroclastic products indicate that typical values of pumice vescicularity lie around 75%, coinciding with the maximum packing of bubble in a liquid and also with the onset of magma fragmentation (Sparks, 1978). Nevertheless, the foam permeability and coalescence, by connecting the dispersed gas phase, allow the exolved gas to escape (Klug and Cashman, 1996) causing the upper limit of pumice vesiculation to shift up to an extremely high value of 99%. These knowledge support the importance of the vesicularity studies in investigating magma properties in pre-eruptive and eruptive conditions. Even if a general interpretation of distinct sub-populations in bubble size sub-distributions is questionable, bubble growth theory and rock analysis indicate that the small bubble fraction (<100 μ) likely derives from syn-eruptive vesiculation in the conduit (Sparks, 1978, Klug and Cashman, 1996). Coarser bubble sub-population have been considered, in turn, as having formed in pre-eruptive conditions, within the magma chamber or resulting from syn-eruptive coalescence (Klug and Cashman, 1994). Parametric modelling (Proussevitch and Sahagian, 1996) indicates that only when the bubble size approaches 0.5 mm, significant degassing occurs. Evidence from pyroclasts indicate that, post-fragmentation events, ‘in the late stage of the magma ascent’, including wall-thinning, relaxation and clast degassing, play an important role in determining the final bubble size distribution (Gardner et al., 1996).

In general, decompression rate, gas diffusivity, initial saturation pressure, viscosity, and initial magma ascent rate control the vesiculation in the ascending magma. Numerical solutions of constitutive equations show two distinct conditions in bubble nucleation and growth: a diffusivity-controlled and a viscosity-controlled regime (Toramaru, 1995). The diffusivity-controlled regime, with relatively low bubble number, dominates in the case of high magma diffusivity, slow magma rise rate, low decompression rate, low magma viscosity or a coupling of these conditions. The viscosity-controlled regime, with higher bubble number density, dominates in opposite conditions. Since numerical simulations evidence a transition between the viscosity-controlled and diffusivity-controlled regimes in the interval of bubble number density (number of bubble/m³), this parameter is a primary source of information on the magma rise and bubble growth processes.

The stability of magma foam and the eruptive conditions at vent depend, ultimately, on magma ascent and vesiculation history. For example, according to Toramaru (1995), in the viscosity-controlled regime, the abundance of tiny bubbles with high internal pressure cause the magma to easily fragment into fine ash under the effect of external shock or other disturbances.

Houghton and Wilson (1989) empirically showed that the pyroclasts formed in magmatic eruptions are closely grouped in a narrow range of vesicularity, regardless of magma viscosity. In contrast, clasts formed in phreatomagmatic events span over a wider range of values. This difference, being likely due to complex variations in the relative timing of vesiculation and water-induced fragmentation suggests that, at least in the examined cases, interaction occurred before complete vesiculation.

On the contrary, Proussevitch and Sahagian, (1996) evidences that the bubble growth in acid magma depends primarily on the ascent condition in the conduit. At high initial ascent rate, bubble growth is inhibited and gas oversaturation increases upward in the conduit. This causes the bubble internal pressure and magma explosivity to increase at shallow depth. In such a condition, shallow magma–water interaction always involves oversaturated foam.

However, although theoretical approaches and experiments explain many features of pumice and provide information on magma-rising mechanisms, the importance of bubble growth regime in controlling the eruptive style remains poorly understood.

In this paper, in order to investigate the relationship between vesiculation process and eruptive style, we carry on comparative analyses of the bubble shape, size and number density on a set of pyroclastic products erupted in Campi Flegrei. These products derive from eruptions occurred in the same geological context and possibly from a unique magma chamber with near-constant composition (Rosi and Sbrana, 1987, Rosi et al., 1983), but through drastically different eruptive mechanisms. Our approach is based on previous evidences from individual deposits in Campi Flegrei, which suggest that magma flow rate and shallow processes rather than the conditions at depth controlled the eruptive style (Mastrolorenzo, 1994).

Section snippets

Units selections, sampling and analytical procedures

The pyroclastic deposits of Campi Flegrei include a near-complete assortment of types (Fig. 1) that can be grouped into two end members with transitions between them:

(a) primary magmatic deposits (including plinian, vulcanian, strombolian types and magmatic pyroclastic flows);
(b) hydromagmatic deposits (including all types of wet and dry surges forming cinder cones, tuff rings, tuff cones and surge bedsets not associated with pyroclastic cones as well as hydromagmatic pyroclastic flows).

The

Results

Fig. 3 shows representative BSE source images of polished thin-sections, relative computer elaborations and log₂ of bubble size vs. log₂ of frequency histograms for ranked data. The selected examples show, as well as most of the analysed specimens, a regular, steep decreasing of the number of bubbles with increasing bubble size, which approaches to zero for bubble size approaching to the specimen dimension.

Table 2 summarises statistical data and computed parameters of the analysed specimens

Discussion and conclusions

This study implies the following basic assumptions, commonly accepted in recent volcanological literature:

(i) The small fraction in bubble size distribution is formed during the ascent of the magma in the conduit.
(ii) The vesicularity of a specific eruptive unit is well represented in a few grain specimen with a total surface from a few square millimeter to a few tens of square millimeter.
(iii) The vesicularity of an individual eruptive unit is homogeneous thus representing the specific

References (18)

M. De’ Gennaro et al.
Depositional mechanisms and alteration processes in different types of pyroclastic deposits in Campi Flegrei volcanic field (Southern Italy)
J. Volcanol. Geotherm. Res.
(1999)
R.A. Herd et al.
Bubble coalescence in basaltic lava: its implication on the evolution of bubble population
J. Volcanol. Geotherm. Res.
(1997)
D.M. Pyle et al.
Bubble migration and iniziation of volcanic eruptions
J. Volcanol. Geotherm. Res.
(1995)
A.A. Proussevitch et al.
Stability of foams in silicate melts
J. Volcanol. Geothermal. Res.
(1993)
M. Rosi et al.
The Phlegrean fields: structural evolution, volcanic history and eruptive machanism
J. Volcanol. Geotherm. Res.
(1983)
M.F. Sheridan et al.
Hydrovolcanism: basic considerations and review
Explosive volcanism
J. Volcanol. Geotherm. Res.
(1983)
R.S.J. Sparks
The dynamics of bubble formation and growth in magmas: A rewiew and analysis
J. Volcanol. Geotherm. Res.
(1978)
K. Wohletz et al.
Eruptive mechanisms of the Neapolitan Yellow Tuff interpreted from stratigraphic, chemical, and granulometric data
J. Volcanol. Geotherm. Res.
(1995)
J.E. Gardner et al.
Fragmentation of magma during Plinian volcanic eruptions
Bull. Volcanol.
(1996)

There are more references available in the full text version of this article.

Cited by (20)

Monogenetic volcanoes with initial phreatomagmatic phases in the Ceboruco graben, western Mexico: The cases of Potrerillo I, Potrerillo II, and San Juanito
2021, Journal of Volcanology and Geothermal Research
Citation Excerpt :
Under such oversaturated conditions, magmatic foam is susceptible to burst under the effect of external shocks caused by magma-water interaction or flow/conduit instability (Sheridan and Wohletz 1983; Toramaru 1989). Accordingly, in many examples from Campi Flegrei, explosive interaction with external water (groundwater and/or surface water) has been inferred to involve an already vesiculated and partially fragmented high-silica magma (Mastrolorenzo et al. 2001). Potrerillo II pumiceous clasts may then be interpreted to have resulted from the premature cessation of vesiculation and fragmentation of a magma foam (Mastrolenzo et al. 2001) that interacted explosively with water.
Potrerillo I, Potrerillo II, and San Juanito volcanoes are part of the 28 Late Quaternary monogenetic eruptive vents concentrated along a ~2-km-wide and ~30-km-long stripe in the vicinity of Ceboruco stratovolcano (western Mexican Volcanic Belt). The studied volcanoes are closely spaced and located within a ~12 km² sub-drainage basin, northwest of Ceboruco's edifice. Eruptive dynamics were inferred from sedimentary, stratigraphic, and pyroclast characteristics. Unlike Potrerillo I, which was formed by an entirely dry explosive magmatic eruption, evidence for the prevalence of initial short-lived phreatomagmatic phases was found for the Potrerillo II and the San Juanito eruptions that ended effusively.
In this study we examine the nature and proportion of different types of pyroclasts within the exposed deposits of these three volcanoes, as well as their sedimentary features to better understand the initiation and progression of their eruptions. Our results indicate that both Potrerillo II and San Juanito tephra ring deposits show dominantly atypical, crudely bedded, coarse-grained deposits where the evidence for energetic, diluted pyroclastic density currents is scarce or subordinated. Instead, low energy, shallow seated, phreatomagmatic explosions generated medium-to-coarse grain-sized pyroclasts that were transported within dense tephra jets that were ballistically ejected, which on deposition formed non-turbulent currents with non-extensive lateral movement. The nature of such eruptive style was controlled by the substrate hydrogeology conditions that prevailed at the time of the eruption. So, magmas of different composition (dacitic for Potrerillo II, and basaltic andesite for San Juanito) followed similar dynamics when interacting with water. We propose that fractured aquifers with shallow water tables (or even surface water) existed at the time of both eruptions and interacted explosively with the magmas.
In contrast, the construction of Potrerillo I scoria cone, which is related exclusively to explosive magmatic activity, does not show evidence of phreatomagmatic activity. This eruption may have occurred when the water table was deep, or surface water was unavailable. The variations in the depth of phreatomagmatic explosions, the availability of the water involved in the phreatomagmatic phases, and the absence or presence of these phases in the studied volcanoes were probably strongly influenced by the water-bearing conditions of the aquifers at the time of the eruption, which in turn, depended on the amount and distribution of annual rainfall.
This study improves our understanding of phreatomagmatic eruptions of diverse magmatic compositions in similar settings. The recognition of associated hazards will allow to further define eruptive scenarios of monogenetic eruptions and help update the volcanic hazard map of the Ceboruco volcano area in the future.
The holocene marine record of unrest, volcanism, and hydrothermal activity of campi flegrei and somma-vesuvius
2020, Vesuvius, Campi Flegrei, and Campanian Volcanism
This chapter illustrates the marine record of a spectrum of volcanic, hydrothermal, and sedimentary features that characterize the Latest Pleistocene–Holocene evolution of the Naples Bay offshore Campi Flegrei and Somma–Vesuvius. The work is based on review of previous literature integrated with interpretation of new high-resolution marine Digital Terrain Models (DTMs) derived from swath bathymetry surveys and high-resolution reflection seismic profiles calibrated with marine gravity core data.
Seismic profiles from Pozzuoli Bay provide detailed images of the ring fault system and resurgent dome associated with the evolution of the Neapolitan Yellow Tuff (NYT) collapse caldera and document a series of uplift episodes associated with volcanic unrest as documented on land throughout the Holocene, with a notable subsidence phase occurring between ∼ 2.5 ka BP and 1538 CE (Monte Nuovo eruption). Offshore seismic images also revealed the occurrence of ascending hydrothermal fluids and volcanic/subvolcanic intrusions along the ring fault zone of the NYT caldera.
Seismic data acquired along the SW submerged slope of Somma–Vesuvius display evidence of gravitational instability, associated with slump folding and faulting, of sand waves originated by pyroclastic flows that entered the seawater after destroying the Roman city of Herculaneum during the 79 CE eruption of Vesuvius.
Between the Somma–Vesuvius and Pozzuoli Bay, seismic profiles calibrated with gravity core data revealed the occurrence of a hummocky seafloor region, known as Banco della Montagna (i.e., the Montagna bank). This volcanic bank was shaped by the dragging and rising up of volcaniclastic diapirs (mostly unconsolidated pumice) because of pore fluid overpressure at depth and associated active fluid venting at the seafloor.
Message in a bottle: Spontaneous phase separation of hydrous Vesuvius melt even at low decompression rates
2018, Earth and Planetary Science Letters
Violent explosive volcanic eruptions are destructive and threaten millions of people and infrastructure. Ejected ash, pumice and gases are end-products of volcanic factories sourced deep within magma chambers and conduits. The different production stages are not directly observable. However, experimental simulation reveals different stages of dynamic volcanic processes. The starting point of explosive volcanic eruptions is determined by the phase separation of an H₂O fluid from a supersaturated hydrous silicate melt. The number of formed H₂O fluid vesicles per unit volume of silicate melt (VND) is a basic property that controls the efficiency of fluid-melt separation, ascent velocity and finally explosive volcanism. We performed decompression experiments at superliquidus temperatures to simulate phase separation of a single phase hydrous silicate melt during ascent with AD79 Vesuvius white pumice composition using decompression rates of 0.024–1.7 MPa⋅s⁻¹. The white pumice buried Herculaneum and Pompeii and is representative of other catastrophic phonolitic and trachytic explosive eruptions like the violent 39 ka Campi Flegrei and the 1815 AD Tambora eruption. Here we report a high $\log VND$ of 5.2 (in mm⁻³) that is independent from decompression rate within the investigated range. Even at a decompression rate of 0.024 MPa⋅s⁻¹ the formation of a high VND inevitably causes rapid degassing due to short H₂O diffusion distances from the melt into fluid vesicles. A decompression rate meter based on nucleation theory, which is commonly used to estimate magma ascent velocity during volcanic eruptions using VND of volcanic ejecta, cannot be adapted to explain our experimentally determined decompression rate independent VND. Alternatively, decompression induced H₂O–silicate melt phase separation may be described by diffusion controlled spinodal decomposition where maximum supersaturation is reached. This process occurs spontaneously and free of activation energy if hydrous melt is driven into thermodynamic instability where the second derivative of free energy of mixing to the H₂O content is ≤0. However, the decompression rate independent VND has profound consequences for the dynamics of natural polyphase hydrous magma phase separation. Even at low ascent rates spontaneous hydrous melt phase separation facilitates rapid density decrease accompanied by sudden increase of magma buoyancy triggering explosive eruptions.
Resuspension of ash after the 2014 phreatic eruption at Ontake volcano, Japan
2018, Journal of Volcanology and Geothermal Research
Citation Excerpt :
The threshold velocity depends on particle size and density, and the degree of moisture and consolidation of the deposit (Bagnold, 1941; Shao and Lu, 2000). These properties of ash deposits differ between eruption styles (Walker, 1971; Walker, 1973; Mastrolorenzo et al., 2001; Mueller et al., 2011; Maeno et al., 2016). In contrast to magmatic eruptions, resuspension after a phreatic eruption has not yet been reported.
We determined the resuspension process of an ash deposit after the phreatic eruption of September 27th, 2014 at Ontake volcano, Japan, by analyzing the time series data of particle concentrations obtained using an optical particle counter and the characteristics of an ash sample. The time series of particle concentration was obtained by an optical particle counter installed 11 km from the volcano from September 21st to October 19th, 2014. The time series contains counts of dust particles (ash and soil), pollen, and water drops, and was corrected to calculate the concentration of dust particles based on a polarization factor reflecting the optical anisotropy of particles. The dust concentration was compared with the time series of wind velocity. The dust concentration was high and the correlation coefficient with wind velocity was positive from September 28th to October 2nd. Grain-size analysis of an ash sample confirmed that the ash deposit contains abundant very fine particles (< 30 μm). Simple theoretical calculations revealed that the daily peaks of the moderate wind (a few m/s at 10 m above the ground surface) were comparable with the threshold wind velocity for resuspension of an unconsolidated deposit with a wide range of particle densities. These results demonstrate that moderate wind drove the resuspension of an ash deposit containing abundant fine particles produced by the phreatic eruption.
Magma degassing and crystallization processes during eruptions of high-risk Neapolitan-volcanoes: Evidence of common equilibrium rising processes in alkaline magmas
2006, Earth and Planetary Science Letters
Compositional, textural and experimental data on products from explosive and effusive eruptions of Neapolitan volcanoes (Campi Flegrei and Somma-Vesuvio) allow us to constrain degassing and fragmentation conditions during eruptions of alkaline magmas. Significant differences in compositional and textural features have been recognized between lavas, scoria and pumice resulting respectively from effusive, moderately and extremely explosive eruptions. Pumice samples have highly-vesicular glassy matrix, low microlite number density and moderate to high water content. Crystal Size Distributions (CSD) are steep with high intercept values; the narrow microlite size range indicates single nucleation event. Scoria products are characterized by moderate vesicularity and water content. They have high number density of microlites which are bimodal in size. CSD show distinct inflections that are explained as two crystal populations growing in distinct time. Lava samples generally have low vesicularities, moderate to high microcrystalline groundmass and low glass water content. The comparison between textural and compositional features of natural rocks with samples obtained by decompression experiments allows us to conclude that degassing processes during magma ascent occurs in near-equilibrium conditions even at high decompression rate. Moderate to long magma rise times, calculated in the order of a few days, produce open-degassing responsible for moderately explosive to effusive activity. Short magma rise times, calculated in the order of a few hours, result in closed-system degassing that allow explosive fragmentation when the volume of growing bubble reaches a fixed threshold. Vesicularity and water content measured on matrix glass of pumice indicate that this process occurs at pressure of 10–30 MPa. In these conditions, degassing, fragmentation and in turn the eruptive style is strongly influenced by initial conditions in the magma chamber (volatile content, temperature, pressure) instead of decompression rate, in contrast with that observed for rhyolitic melts. These differences have important consequences in terms of volcanic hazards and risk. The low-viscosity alkaline magma is able to maintain efficient degassing even during the final stage of magma ascent, favoring, in the case of closed-system, fragmentation and explosive activity.
The Somma-Vesuvius volcano (Southern Italy): Structure, dynamics and hazard evaluation
2006, Earth-Science Reviews
We review the main results, with several new analyses, obtained in recent times about the structure, present dynamics and hazard evaluation at Somma–Vesuvius volcanic complex. We present a global review and interpretation of structural features, both at local and regional scale, constrained both by seismic and petrological data. The local structure of Somma–Vesuvius is reviewed in three depth ranges, shallow, intermediate and deep. The shallow velocity structure is inferred by the joint inversion of shot and local earthquake arrival time data. The main feature pointed out at shallow depth is a high velocity anomaly at the crater axis, extending down to about 5 km of depth. Such an anomaly, first observed at Vesuvius, seems to be common to many other volcanoes. It can be interpreted in terms of the presence of solidified residual magma in the shallow conduits, accumulated in last eruptive cycles. The local seismicity is strongly clustered around this anomaly, due to the focusing effect of the rigidity contrast. The seismic occurrence appears as a result of the superposition of a background level, mainly due to gravitational instability of the Vesuvius cone, and of intense activity episodes, which possibly reflect episodic internal activity. Two main zones of magma accumulation in the upper crust are evidenced by the joint interpretation of seismic and petrological data. The first one, located in the depth range 4–6 km, is mainly constrained by the crystallisation depth of phonolitic magmas which fed Plinian and sub-Plinian eruptions; the second one, around 11–15 km of depth, is mainly constrained by reflected–converted seismic waves, and in agreement with crystallization depths inferred for the moderate eruptions. The study of the deep structure, performed by regional tomography with teleseisms, further points out magma roots at higher depths (15–30 km). An additional result for the deep structure, studied at regional scale and very important for geodynamic interpretations of the Tyrrhenian volcanisms, has been the evidence for a subducting slab under the Apennines, in an area where previous models hypothesised a slab window.
New original studies of crystal growth (phenocrystals and microlites) on the eruptive products allow to infer typical times of magma rising from such reservoirs, which appear very low, on the order of minutes to tens of minutes. Static deformation at this volcano, in the last 30 yr, has been detected by the joint use of levelling, GPS and DIFSAR techniques. It indicates subsidence, very concentrated in the crater area and in a narrow strip all around the volcanic edifice, with maximum rates less than 0.01 m/yr. Static deformation in the crater area appears in agreement with the mechanism of gravitational instability generating local volcano-tectonic seismicity, while the peculiar pattern around the volcanic edifice is probably due to the combination of extensional stress and volcanic loading, generating a ring normal fault-like structure. While the key results about structure and dynamics help to define pre-eruptive scenarios, a new probabilistic procedure to combine volcanological data and computer simulations has been used, in this paper, to build hazard maps giving the probability, at each location in the area, to be hit by a pyroclastic flow or to experience a destructive fall-out deposit. The review and new results of this work give then the first complete picture of the state of the art in our knowledge about Somma–Vesuvius volcano.

View all citing articles on Scopus

View full text

Vesicularity of various types of pyroclastic deposits of Campi Flegrei volcanic field: evidence of analogies in magma rise and vesiculation mechanisms

Abstract

Introduction

Section snippets

Units selections, sampling and analytical procedures

Results

Discussion and conclusions

J. Volcanol. Geotherm. Res.

J. Volcanol. Geotherm. Res.

J. Volcanol. Geotherm. Res.

J. Volcanol. Geothermal. Res.

J. Volcanol. Geotherm. Res.

J. Volcanol. Geotherm. Res.

J. Volcanol. Geotherm. Res.

J. Volcanol. Geotherm. Res.

Fragmentation of magma during Plinian volcanic eruptions

Bull. Volcanol.