Deep Magma Reservoir Research Articles

Sill intrusion as a source mechanism of unrest at volcanic calderas

AbstractThe interpretation of dynamic processes that occur in volcanic calderas is not simple. The ground deformations and the local seismicity, which in other volcanic contexts are usually regarded as precursors to eruption, in caldera environment in many cases are not followed by any eruption. We formulate a general hypothesis that can explain these behaviors. Our hypothesis is that the intrusion of a sill can be responsible for the dynamics observed during unrest at calderas. In order to investigate the reliability of this hypothesis, we developed a dynamic model of sill intrusion in a shallow volcanic environment. In our model, the sill, fed by a deeper magma reservoir, intrudes below a horizontal elastic plate, representing the overlying rocks, and expands with axisymmetric geometry. The model is based on the numerical solution of the equation for the elastic plate, coupled with a Navier‐Stokes equation for simulating the dynamics of the sill intrusion. We performed a number of simulations, with the objective of showing the main features of the model. In the experiments, when the feeding process stops, the vertical movement reverses its trend and the area of maximum uplift undergoes subsidence. Under certain conditions the subsidence can occur even during the intrusion of the sill. The stress field produced by the intrusion is mainly concentrated in a circular zone that follows the sill intrusion front. The features predicted by the model are consistent with many observations carried out on different calderas as reported in the scientific literature.

Self-similar clustering distribution of structural features on Ascraeus Mons (Mars): implications for magma chamber depth

Abstract The occurrence and distribution of monogenic eruptive features in volcanic areas testify to the presence of deep-crustal or subcrustal magma reservoirs hydraulically connected to the surface via a fracture network. The spatial distribution of vents can be studied in terms of self-similar (fractal) clustering, described by a fractal exponent D and defined over a range of lengths ( l ) between a lower and upper cutoff, L co and U co , respectively. The computed U co values for several volcanic fields on Earth match the thickness of the crust between vents and magma reservoirs at depth. This analysis can thus be extended to other volcanic fields and volcanoes on rocky planets in the solar system where features such as vents and dykes occur, and for where complementary geophysical data are currently lacking. We applied this method to the Ascraeus Mons volcano on Mars, which presents hundreds of collapse pits similar to those observed on Earth volcanoes that are most likely related to feeder dykes. Based on structural mapping with High Resolution Stereo Camera data at 12 m/px and Context Camera data at 6 m/px mosaics, more than 2300 collapse pits and dyke traces were analysed, revealing two distinct fractal clustered populations. The obtained U co values reveal the presence and likely depth of both a deep magma reservoir ( c. 60 km deep) and a small shallower chamber ( c. 11 km deep). This analysis can help to better constrain the depth and time evolution of volcanic processes on Tharsis, and on terrestrial planets’ volcanoes in general.

Anatomy of Piton de la Fournaise volcano (La Réunion, Indian Ocean)

The aim of this work is to propose a general model of Piton de la Fournaise volcano using information from geological and geophysical studies. Firstly, we make a graphical compilation of all available geophysical information along a W–E profile. Secondly, we construct a geological section that integrates both the geophysical information and the geological information. The lithosphere beneath Piton de la Fournaise is not significantly flexed, and the crust is underlain by an underplating body, which might represent the deep magma reservoir for La Reunion volcanism. Piton de la Fournaise is a relatively thin volcano lying on a huge volcanic construction attributed mostly to Les Alizes volcano. Indeed, if the differentiated rocks observed at the bottom of the Riviere des Remparts are the top of Les Alizes volcano, the interface with Piton de La Fournaise may be located at about sea level beneath the summit area. The endogenous constructions (intrusive complexes) related to Les Alizes and Piton de la Fournaise volcanoes represent a large volume. The huge intrusive complex of Les Alizes volcano probably rests on the top of the oceanic crust and appears to have a buttressing effect for the present eastern volcano-tectonic activity of Piton de la Fournaise. The early Piton de la Fournaise edifice was built around a focus located beneath the Plaine des Sables area. The center subsequently moved 5–6 km eastward to its current location. The dense, high-velocity body beneath the Plaines des Sables and the western part of the Enclos probably corresponds to the hypovolcanic intrusive complex that developed before the volcanic center shifted to its present-day position. Magma reservoirs may have existed, and may still exist, as illustrated by the March 1998 crisis, at the mechanical and density interface between the oceanic crust and the Les Alizes edifice. Strong evidence also exists for the presence of a shallower magma reservoir located near sea level beneath the summit. The March 1998 pre-eruptive seismic pattern (location and upward migration) seems to be evidence for a transfer of magma between the two reservoirs. The dominant structural feature of the central zone is a collapse structure beneath the summit craters, above the inferred magma reservoir near sea level. The collapsed column constitutes a major mechanical heterogeneity and concentrates most of the seismic, intrusive, and hydrothermal activity because of its higher permeability and weaker mechanical strength.

Helium isotopes in the Izu Peninsula, Japan: Relation of magma and crustal activity

The relation of magma and crustal activity has been studied from spatial distribution of 3He/ 4He ratios of gas and/or water samples over the Izu Peninsula, where significant crustal deformation associated with seismic swarm activities has been observed since 1970s. The air-corrected values of 3He/ 4He ratios ranged from 3.5 to 8.2 R A, where R A is the atmospheric 3He/ 4He ratio = 1.4 × 10 − 6 , indicating that helium is mostly of magmatic origin. Among the three pressure sources proposed to explain the crustal deformation, two inflation sources beneath the inland of northeast and the mid east coast of the Izu Peninsula locate in the broad distribution of high 3He/ 4He ratios, which supports relation of magma to the crustal uplift. In contrast, the distribution of 3He/ 4He ratios around the tensile fault assumed in the area of seismic swarms appears not to indicate existence of significant amount of magma below the tensile fault. Alternatively, the results suggest magma below a point several kilometers south of the tensile fault. The seismic swarms are explained either by fluid pressurization of thermal water heated by this magma or by intrusion of magma to the tensile fault moved obliquely from the deep magma reservoir.

Transient rift opening in response to multiple dike injections in the Manda Hararo rift (Afar, Ethiopia) imaged by time‐dependent elastic inversion of interferometric synthetic aperture radar data

Interferometric synthetic aperture radar (InSAR) data spanning the time intervals separating thirteen dike intrusions in the Manda Hararo–Dabbahu rift (Afar, Ethiopia) from 2005 to 2009 show that transient deformation occurs in the inter‐diking period. This deformation can be explained by the presence of seven inflating or deflating pressure sources. By combining the data acquired on four different InSAR tracks, through time‐dependent elastic models, we are able to track these deformation modes with a time resolution smaller than 1 month. Sustained deflation of a deep magma reservoir at Dabbahu in the 6 months following the main rifting event of 2005, and slow decelerating post‐eruptive re‐inflation of two shallow magma reservoirs below Dabbahu and Gabho volcanoes, are monitored. A deflation signal of deep origin on the neighboring rift system is also detected, possibly caused by outflow of material from a preexisting reservoir into the deep plate boundary. In contrast, rapidly evolving deformation is observed at the center of the Manda Hararo rift segment. Transient deformation events are monitored in the weeks/months following the diking events, with pulses of localized rift opening after the dike intrusions, followed by an exponential‐like decay of opening rate. This signal may be associated with the replenishment of the central magma reservoir involved in feeding the 2005–2009 dikes. Alternatively, the predominantly horizontal mode of deformation suggests an interaction between the response of the lithosphere to tectonic strain accumulation, and the process of hydraulic connectivity within the central magma plumbing system.

A new petrological and geophysical investigation of the present‐day plumbing system of Mount Vesuvius

A model of the electrical resistivity of Mt. Vesuvius has been elaborated to investigate the present structure of the volcanic edifice. The model is based on electrical conductivity measurements in the laboratory, on geophysical information, in particular, magnetotelluric (MT) data, and on petrological and geochemical constraints. Both 1‐D and 3‐D simulations explored the effect of depth, volume and resistivity of either one or two reservoirs in the structure. For each configuration tested, modeled MT transfer functions were compared to field transfer functions from field magnetotelluric studies. The field electrical data are reproduced with a shallow and very conductive layer (∼0.5 km depth, 1.2 km thick, 5 ohm.m resistive) that most likely corresponds to a saline brine present beneath the volcano. Our results are also compatible with the presence of cooling magma batches at shallow depths (<3–4 km depth). The presence of a deeper body at ∼8 km depth, as suggested by seismic studies, is consistent with the observed field transfer functions if such a body has an electrical resistivity > ∼100 ohm.m. According to a petro‐physical conductivity model, such a resistivity value is in agreement either with a low‐temperature, crystal‐rich magma chamber or with a small quantity of hotter magma interconnected in the resistive surrounding carbonates. However, the low quality of MT field data at long periods prevent from placing strong constraints on a potential deep magma reservoir. A comparison with seismic velocity values tends to support the second hypothesis. Our findings would be consistent with a deep structure (8–10 km depth) made of a tephriphonolitic magma at 1000°C, containing 3.5 wt%H2O, 30 vol.% crystals, and interconnected in carbonates in proportions ∼45% melt −55% carbonates.

Magma storage and ascent at Lipari Island (Aeolian archipelago, Southern Italy) at 223–81 ka: the role of crustal processes and tectonic influence

Fluid inclusion studies together with volcanological and petrochemical data allow reconstruction of the magma feeding system of basaltic-andesitic to andesitic activity during the oldest and intermediate stages of development of Lipari Island (223–81 ka). A major magma storage zone is active during the overall investigated time span at depths of 22 km, close to the crust-mantle Moho transition, at which mantle-derived mafic magmas tend to accumulate due to neutral buoyancy conditions. Beneath central-type volcanoes (M. Mazzacaruso, M. S.Angelo, M. Chirica-Costa d’Agosto), a shallower magma reservoir is located within the upper crust at 5.5–3.5 km, associated with a major lithological discontinuity. For fissural-type volcanoes (Timpone Ospedale, Monterosa, M. Chirica), tectonic structures are suggested to influence further magma ascent and storage at mid-crustal depths (∼14 km), with no ponding at shallower levels. Partial crustal melting processes at the roofs of the deep magma reservoirs (∼17 km) are invoked to explain the origin of cordierite-bearing lavas beneath M. S.Angelo and M. Chirica-Costa d’Agosto volcanoes, which were active during the intermediate stages of development of Lipari (105–81 ka). The generation of felsic anatectic melts in the lower crust could have created density and rheologic barriers to impede the passage of mafic melts and promote their ponding, with influence on the subsequent evolution of Lipari volcano.

SAR- and gravity change-based characterization of the distribution pattern of pyroclastic flow deposits at Mt. Merapi during the past 10 years

Mt. Merapi, Indonesia, is one of the most active and dangerous volcanoes in the Torrid Zone. This volcano has erupted frequently and has produced pyroclastic flows following the collapse of the summit lava dome. We used Synthetic Aperture Radar (SAR) data acquired by JERS-1 and RADARSAT-1 satellites from April 1996 to July 2006 to clarify the distribution patterns of the pyroclastic flow deposits. The extent of the deposits, termed P-zones, was accurately extracted by ratio operation and low-level feature extraction from SAR intensity images. These images highlighted temporal changes of the distribution area, perimeter, flow distance, included angle, and collapse direction. To validate the image-processing results, reflectance spectra of the rock samples collected after the eruption in June 2006 were measured in a laboratory. The reflectance spectra of all samples showed similar characteristics to the reference spectra, which were derived from atmospheric correction of Hyperion sensor image data covering the lava dome at the summit. Therefore, P-zones were confirmed to be the pyroclastic flow deposits originating from destruction of the lava dome at the summit. The image-processing results clarified that the extent of the distribution areas, perimeter, flow distances, and included angle of the P-zones were variable among the eruptions, while the collapse direction had a constant pattern. The collapse pattern followed a clockwise change from the south toward the west. By comparing the ratio maps of Bouguer gravity anomaly data in two periods, the change was interpreted to originate from the inclination of the conduit and the formation of shallow and deep magma reservoirs.

A Temporal Record of Magma Accumulation and Evolution beneath Nevado de Toluca, Mexico, Preserved in Plagioclase Phenocrysts

Plagioclase crystals from the 8 km 10 5 ka Upper Toluca Pumice (UTP) eruption from Nevado deToluca provide a detailed temporal record of pre-eruptive magmatic processes. The crystals display a range of textures and major and trace element concentrations. A distinct feature of the crystals is the presence of several sharp increases in MgO, FeO andTiO2, which occur at some of the numerous resorption horizons and coincide with increases in XAn and in some cases Ce and La. These abrupt compositional peaks, and the associated textural variations, reflect recharge events of more mafic melt.Three distinct recharge events can be recognized.The general compositional trends suggest that crystallization occurred within a common chamber. However, the crystal cores display a range of textures and a wide range of compositions suggesting that they are relics of earlier crystallization episodes within the upper crust. The temporal record of Sr and Ba melt contents recorded by the crystals, calculated using partition coefficients, fluctuates significantly within single crystals. Overall, two trends of Sr^Ba variation in the melt are apparent. The first trend involves decreasing Sr and slightly increasing Ba, consistent with plagioclase crystallization.The second trend involves an increase in both Sr and Ba that is not consistent with crystallization of plagioclase alone. This trend is ascribed to chemical variations within the deeper magma reservoirs from which the various magma batches entering the UTP shallow reservoir were ultimately sourced. The magmatic system under Nevado de Toluca was open and received intermittent, but relatively small, pulses of magma from a deeper source. Dacite was the predominant recharge composition, although some pulses were distinctly more mafic and hotter. These processes led to the accumulation of a large volume of dacitic magma in the upper crust. One of the more mafic pulses appears to have triggered the eruption.

Petrology of 1977 to 1998 eruptions of Piton de la Fournaise, La Réunion Island

Geochemical and petrological variations in 35 historical eruptions from the April 1977 eruption to the March–September 1998 eruption at Piton de la Fournaise have been analyzed for their major elements abundances, mineralogical compositions and oxygen isotopes. During these 21 years, samplings were repeated several times during eruptions as well as temperature measurements. Since the end of 1977 to 1998, all erupted lavas, except 1998-Hudson basalt, belong to the group of the “steady state basalts”. However, evidences for small degrees of differentiation can be observed. During a single eruption, the earliest lava emitted contains less phenocrysts and has almost always a lower MgO⁎ (= MgO/[MgO + FeOt]) than the lavas emitted later, suggesting that it is more differentiated. From one eruption to the next, between 1977 and 1998, bulk compositions are more and more differentiated towards the first lava of 1998. A process of low-pressure fractional crystallisation, involving small proportions of olivine, clinopyroxene and plagioclase, allows to explain this evolution. A major result of our study is to show that this evolution does not correspond to a unique fractionation trend. Over the considered period, different types of magmas have been distinguished. The first magma type, erupted from October 1977 to November 1987, was richer in potassium (0.82 wt.% in average) than the second one, and apparently derived from the liquid of the 1977 oceanite. The second magma type, erupted from February 1988 to September 1998, was poorer in K 2O (0.69 wt.% in average). In March 1998, a third magma type (Hudson basalt) erupted simultaneously with the second one. Hudson basalt has undergone a wehrlitic fractionation and is supposed to come from a deeper magma reservoir than other erupted magmas for this period. From 1977 to 1998, the ∂ 18O of lavas progressively decreases with time, together with the increasing chemical differentiation of the magmas by low-pressure fractional crystallisation. This variation agrees with the hypothesis of magma contamination at shallow level in the altered volcanic edifice (Vlastélic, I., Deniel, C., Bosq, C., Télouk, P., Boivin, P., Bachèlery, P., Famin, V., Staudacher, T., 2009-this issue. Pb isotope geochemistry of Piton de la Fournaise historical lavas. Journal of Volcanology and Geothermal Research, 184, 63−78 (this issue)). Overall, petrological variations are in good agreement with a model of a shallow reservoir discontinuously supplied between 1977 and 1998. The speed of the variations, the coexistence of different liquids, the chemical evolution of the lavas during the same eruption, lead us to favour a model of chemically stratified reservoir, constituted by a complex of dykes, sills and laccoliths close to the surface and located above larger magma bodies.

Characteristics, extent and origin of hydrothermal alteration at Mount Rainier Volcano, Cascades Arc, USA: Implications for debris-flow hazards and mineral deposits

Hydrothermal alteration at Mount Rainier waxed and waned over the 500,000-year episodic growth of the edifice. Hydrothermal minerals and their stable-isotope compositions in samples collected from outcrop and as clasts from Holocene debris-flow deposits identify three distinct hypogene argillic/advanced argillic hydrothermal environments: magmatic-hydrothermal, steam-heated, and magmatic steam (fumarolic), with minor superimposed supergene alteration. The 3.8 km 3 Osceola Mudflow (5600 y BP) and coeval phreatomagmatic F tephra contain the highest temperature and most deeply formed hydrothermal minerals. Relatively deeply formed magmatic-hydrothermal alteration minerals and associations in clasts include quartz (residual silica), quartz–alunite, quartz–topaz, quartz–pyrophyllite, quartz–dickite/kaolinite, and quartz–illite (all with pyrite). Clasts of smectite–pyrite and steam-heated opal–alunite–kaolinite are also common in the Osceola Mudflow. In contrast, the Paradise lahar, formed by collapse of the summit or near-summit of the edifice at about the same time, contains only smectite–pyrite and near-surface steam-heated and fumarolic alteration minerals. Younger debris-flow deposits on the west side of the volcano (Round Pass and distal Electron Mudflows) contain only low-temperature smectite–pyrite assemblages, whereas the proximal Electron Mudflow and a < 100 y BP rock avalanche on Tahoma Glacier also contain magmatic-hydrothermal alteration minerals that are exposed in the avalanche headwall of Sunset Amphitheater, reflecting progressive incision into deeper near-conduit alteration products that formed at higher temperatures. The pre-Osceola Mudflow alteration geometry is inferred to have consisted of a narrow feeder zone of intense magmatic-hydrothermal alteration limited to near the conduit of the volcano, which graded outward to more widely distributed, but weak, smectite–pyrite alteration within 1 km of the edifice axis, developed chiefly in porous breccias. The edifice was capped by a steam-heated alteration zone, most of which resulted from condensation of fumarolic vapor and oxidation of H 2S in the unsaturated zone above the water table. Weakly developed smectite–pyrite alteration extended into the west and east flanks of the edifice, spatially associated with dikes that are localized in those sectors; other edifice flanks lack dikes and associated alteration. The Osceola collapse removed most of the altered core and upper east flank of the volcano, but intensely altered rocks remain on the uppermost west flank. Major conclusions of this study are that: (1) Hydrothermal–mineral assemblages and distributions at Mount Rainier can be understood in the framework of hydrothermal processes and environments developed from studies of ore deposits formed in analogous settings. (2) Frequent eruptions supplied sufficient hot magmatic fluid to alter the upper interior of the volcano hydrothermally, despite the consistently deep (≥ 8 km) magma reservoir which may have precluded formation of economic mineral deposits within or at shallow depths beneath Mount Rainier. The absence of indicator equilibrium alteration-mineral assemblages in the debris flows that effectively expose the volcano to a depth of 1–1.5 km also suggests a low potential for significant high-sulfidation epithermal or porphyry-type mineral deposits at depth. (3) Despite the long and complex history of the volcano, intensely altered collapse-prone rocks were spatially restricted to near the volcano's conduit system and summit, and short distances onto the upper east and west flanks, due to the necessary supply of reactive components carried by ascending magmatic fluids. (4) Intensely altered rocks were removed from the summit, east flank, and edifice interior by the Osceola collapse, but remain on the upper west flank in the Sunset Amphitheater area and present a continuing collapse hazard. (5) Visually conspicuous rocks on the lower east and mid-to-lower west flanks are not intensely altered and probably have not significantly weakened the rock, and thus do not present significant collapse hazards. (6) Alteration developed most intensely within breccia units, because of their high permeability and porosity. Volcanoes with abundant near-conduit upper-edifice breccias are prone to alteration increasing the possibility of collapse, whereas those that are breccia-poor (e.g., massive domes) are less prone to alteration.

Crustal structure of the Trans‐Atlantic Geotraverse (TAG) segment (Mid‐Atlantic Ridge, 26°10′N): Implications for the nature of hydrothermal circulation and detachment faulting at slow spreading ridges

New seismic refraction data reveal that hydrothermal circulation at the Trans‐Atlantic Geotraverse (TAG) hydrothermal field on the Mid‐Atlantic Ridge at 26°10′N is not driven by energy extracted from shallow or mid‐crustal magmatic intrusions. Our results show that the TAG hydrothermal field is underlain by rocks with high seismic velocities typical of lower crustal gabbros and partially serpentinized peridotites at depth as shallow as 1 km, and we find no evidence for low seismic velocities associated with mid‐crustal magma chambers. Our tomographic images support the hypothesis of Tivey et al. (2003) that the TAG field is located on the hanging wall of a detachment fault, and constrain the complex, dome‐shaped subsurface geometry of the fault system. Modeling of our seismic velocity profiles indicates that the porosity of the detachment footwall increases after rotation during exhumation, which may enhance footwall cooling. However, heat extracted from the footwall is insufficient for sustaining long‐term, high‐temperature, hydrothermal circulation at TAG. These constraints indicate that the primary heat source for the TAG hydrothermal system must be a deep magma reservoir at or below the base of the crust.

Aftershock decay, productivity, and stress rates in Hawaii: Indicators of temperature and stress from magma sources

We examined dozens of aftershock sequences in Hawaii in terms of Gutenberg‐Richter and modified Omori law parameters. We studied p, the rate of aftershock decay; Ap, the aftershock productivity, defined as the observed divided by the expected number of aftershocks; and c, the time delay when aftershock rates begin to fall. We found that for earthquakes shallower than 20 km, p values &gt;1.2 are near active magma centers. We associate this high decay rate with higher temperatures and faster stress relaxation near magma reservoirs. Deep earthquakes near Kilauea's inferred magma transport path show a range of p values, suggesting the absence of a large, deep magma reservoir. Aftershock productivity is &gt;4.0 for flank earthquakes known to be triggered by intrusions but is normal (0.25 to 4.0) for isolated main shocks. We infer that continuing, post‐main shock stress from the intrusion adds to the main shock's stress step and causes higher Ap. High Ap in other zones suggests less obvious intrusions and pulsing magma pressure near Kilauea's feeding conduit. We calculate stress rates and stress rate changes from pre‐main shock and aftershock rates. Stress rate increased after many intrusions but decreased after large M7–8 earthquakes. Stress rates are highest in the seismically active volcano flanks and lowest in areas far from volcanic centers. We found sequences triggered by intrusions tend to have high Ap, high (&gt;0.10 day) c values, a stress rate increase, and sometimes a peak in aftershock rate hours after the main shock. We interpret these values as indicating continuing intrusive stress after the main shock.

A quantitative approach to the dike intrusion process inferred from a joint analysis of geodetic and seismological data for the 1998 earthquake swarm off the east coast of Izu Peninsula, central Japan

A series of earthquake swarms off the east coast of Izu Peninsula, central Japan, is typically associated with dike intrusion around active volcanoes. Numerous studies in this region have documented the seismic activity and ground deformations without fully resolving the process of dike intrusion. This study involves a joint analysis of geodetic and seismic data from the most recent major activity in this region, occurring in 1998, to reveal the process of dike intrusion quantitatively. We combine data from a dense GPS array, including 10 temporary stations, with seismic analysis published previously. We estimate the location of the dikes from seismic data and the volume of intruded magma from geodetic data. Combined analysis of both data sets enables us to reveal the kinematic process of dike intrusion associated with the 1998 earthquake swarm. The process of dike intrusion is quantitatively modeled using the time evolution of recorded dike volume and the use of seismic hypocenters as an indicator of the site of instantaneous dike intrusion. On the basis of these observations, we propose the following process of the dike intrusion: The magma rose upward through a pathway aligned vertically and created a dike when it lost buoyancy. We also evaluate the dike intrusion process quantitatively using a model of a circular dike that grows under a continuous magma supply from a deep magma reservoir. From the observational data, we infer the time evolution in the dike shape, pressure distribution, and magma viscosity in the dike. The acquired viscosity increases over time and it may be reasonable because magma cools gradually. However, its value is too large compared to typical basaltic magma expected in this region. It may represent magma within the dike that is nearly solidified. Otherwise, the magma is supplied intermittently and the dike growth is mainly controlled by the magma supply rate from a deep reservoir. The quantitative description of the dike intrusion process in nature described in this study is very rare, and our results provide constraints on several parameters that have not been solved in many previous theoretical studies on the dike intrusion process.

The geochemical evolution of the Mt. Somma-Vesuvius volcano

This work integrates new geochemical data with the numerous published analyses on rocks from the Mt. Somma-Vesuvius volcano. New quantitative models for the evolution of magma source regions and magma at different depths are proposed. The origin of the Somma-Vesuvius parental magma is modeled as 0.05–0.1 melt fractions of a MORB-type source composed of 54% olivine, 30% orthopyroxene, 10% clinopyroxene, 1% garnet, and 4% amphibole, and 1–5% sediment introduced through the adjacent arc system. The excess concentrations of Rb, Ba, K, and Sr are attributed to a subduction-related fluid phase. Major and trace element concentrations, coupled with Sr–Nd–Pb isotope signatures suggest that the bulk composition of sediments being subducted below southern Italy is similar to that of the carbonate rich sediment columns described by Plank and Langmuir (1998) and Vroon et al. (1995). Furthermore, it appears that the sediment contribution was introduced as a partial melt, which would account for some geochemical patterns, such as 143Nd/144Nd versus Th/Ce. The EC–AFC model (Spera and Bohrson, 2001) is then used to track the evolution of Somma-Vesuvius magmas. The results are consistent with the melting of crustal Hercynian basement at depths of 12 and >20 km (De Natale et al., 2001). Such a model is also consistent with the thermal model of Annen and Sparks (2002) for the evolution of magmatic provinces. Here, magmas from the upper mantle form a melt intrusion and storage zone at 12 to >20 km allowing for crustal melting to take place. At Vesuvius, Plinian eruptions involve the first magma withdrawn from a deep magma reservoir. Interplinian eruptions involve reduced volumes of magma stored over a larger depth range until the volcanic activity stops. This suggests that little magma is left in the melt intrusion and storage zone. A new cycle is started by a Plinian event when new magma rises from the upper mantle and is emplaced in the lower crust.

Unprecedented pressure increase in deep magma reservoir triggered by lava‐dome collapse

The collapse of the Soufrière Hills Volcano lava dome on Montserrat in July 2003 is the largest such event worldwide in the historical record. Here we report on borehole dilatometer data recording a remarkable and unprecedented rapid (∼600s) pressurisation of a magma chamber, triggered by this surface collapse. The chamber expansion is indicated by an expansive offset at the near dilatometer sites coupled with contraction at the far site. By analyzing the strain data and using added constraints from experimental petrology and long‐term edifice deformation from GPS geodesy, we prefer a source centered at approximately 6 km depth below the crater for an oblate spheroid with overpressure increase of order 1 MPa and average radius ∼1 km. Pressurisation is attributed to growth of 1–3% of gas bubbles in supersaturated magma, triggered by the dynamics of surface unloading. Recent simulations demonstrate that pressure recovery from bubble growth can exceed initial pressure drop by nearly an order of magnitude.

The Campi Flegrei caldera: unrest mechanisms and hazards

Abstract In the last four decades, Campi Flegrei caldera has been the world’s most active caldera characterized by intense unrest episodes involving huge ground deformation and seismicity, but, at the time of writing, has not culminated in an eruption. We present a careful review, with new analyses and interpretation, of all the data and recent research results. We deal with three main problems: the tentative reconstruction of the substructure; the modelling of unrest episodes to shed light on possible pre-eruptive scenarios; and the probabilistic estimation of the hazards from explosive pyroclastic products. The results show, for the first time at a volcano, that a very peculiar mechanism is generating episodes of unrest, involving mainly activation of the geothermal system from deeper magma reservoirs. The character and evolution of unrest episodes is strongly controlled by structural features, like the ring-fault system at the borders of the caldera collapse. The use of detailed volcanological, mathematical and statistical procedures also make it possible to obtain a detailed picture of eruptive hazards in the whole Neapolitan area. The complex behaviour of this caldera, involving interaction between magmatic and geothermal phenomena, sheds light on the dynamics of the most dangerous types of volcanoes in the world.

Quartz and feldspar zoning in the eastern Erzgebirge volcano-plutonic complex (Germany, Czech Republic): evidence of multiple magma mixing

Zoned quartz and feldspar phenocrysts of the Upper Carboniferous eastern Erzgebirge volcano-plutonic complex were studied by cathodoluminescence and minor and trace element profiling. The results verify the suitability of quartz and feldspar phenocrysts as recorders of differentiation trends, magma mixing and recharge events, and suggest that much heterogeneity in plutonic systems may be overlooked on a whole-rock scale. Multiple resorption surfaces and zones, element concentration steps in zoned quartz (Ti) and feldspar phenocrysts (anorthite content, Ba, Sr), and plagioclase-mantled K-feldspars etc. indicate mixing of silicic magma with a more mafic magma for several magmatic phases of the eastern Erzgebirge volcano-plutonic complex. Generally, feldspar appears to be sensitive to the physicochemical changes of the melt, whereas quartz phenocrysts are more stable and can survive a longer period of evolution and final effusion of silicic magmas. The regional distribution of mixing-compatible textures suggests that magma mingling and mixing was a major process in the evolution of these late-Variscan granites and associated volcanic rocks. Quartz phenocrysts from 14 magmatic phases of the eastern Erzgebirge volcano-plutonic complex provide information on the relative timing of different mixing processes, storage and recharge, allowing a model for the distribution of magma reservoirs in space and time. At least two levels of magma storage are envisioned: deep reservoirs between 24 and 17 km (the crystallisation level of quartz phenocrysts) and subvolcanic reservoirs between 13 and 6 km. Deflation of the shallow reservoirs during the extrusion of the Teplice rhyolites triggered the formation of the Altenberg-Teplice caldera above the eastern Erzgebirge volcano-plutonic complex. The deep magma reservoir of the Teplice rhyolite also has a genetic relationship to the younger mineralised A-type granites, as indicated by quartz phenocryst populations. The pre-caldera biotite granites and the rhyodacitic Schönfeld volcanic rocks represent temporally and spatially separate magma sources. However, the deep magma reservoir of both is assumed to have been at a depth of 24–17 km. The drastic chemical contrast between the pre-caldera Schönfeld (Westfalian B–C) and the syn-caldera Teplice (Westfalian C–D) volcanic rocks is related to the change from late-orogenic geotectonic environment to post-orogenic faulting, and is considered an important chronostratigraphic marker.

Magma changes at Mount Etna: the 2001 and 2002–2003 eruptions

The 2001 and 2002–2003 Etna eruptions were strongly explosive and their lavas (trachybasalts and alkali basalts) exhibit quite unusual mineralogy and geochemistry. In addition to plagioclase, clinopyroxene and olivine, some of these lavas contain amphibole megacrysts which are shown to be formed in a volatile-rich magma at the top of a shallow chamber, and relicts of hypersthene phenocrysts resulting from an earlier, high pressure stage of crystallization (1.1–1.4 GPa, i.e. 30–40 km depth). Variations observed in whole-rock chemistry, trace element contents, 87Sr/ 86Sr ratios and 238U– 230Th– 226Ra– 210Pb disequilibria are consistent with arrival of new basaltic magma from depth, which intruded two other batches of slightly differentiated magmas (trachybasalts) occupying the upper parts of the volcanic conduits. The 2001 basaltic magma presents geochemical signatures slightly different from those of the 1974 lava, which heralded a drastic change in chemical composition and eruptive activity of the past 30 years. 210Pb– 226Ra disequilibria strongly suggest that Ra, K, Rb, Cs and 87Sr enrichments which characterize the new magma(s) took place at least several decades before eruption, and are thus deep-seated phenomena occurring in the mantle source during partial melting. After a beginning of crystallization at depth (hypersthene), this magma then ascended within a few months in upper crustal levels, triggering dike emplacement, degassing and eruption. The presently exceptional Etna's activity seems to result, therefore, from regional mantle processes that control both tectonics and magma ascent. After about three centuries of nearly steady state behaviour of the deep magma reservoir, the present period of activity since 1971 corresponds to an increased influx from the mantle of (a) distinct basaltic magma(s), which escape(s) mixing in the deep reservoir. Its geochemical signatures and higher production rate could be explained by melting of a fluid-enriched, metasomatized portion of the mantle.

Dynamics of magma degassing

Abstract Gas exsolution and segregation are fundamental controls on eruption dynamics and magma genesis. Basaltic magma loses gas relatively easily because of its low viscosity. However, bubbles grown by decompression and diffusion during magma ascent are too small to segregate. Coalescence, however, can create bubbles big enough for gas to escape from the rising basalt magma. In evolved magmas, such as andesite and rhyolite, high viscosity prevents bubbles rising independently through the magma. The original gas content of magma erupted as lava is commonly the same as that erupted explosively, so that a gas separation mechanism is required. A permeable magma foam can form to allow gas escape once bubbles become interconnected. Magma permeabilities can be much higher than wall-rock permeabilities, and so vertical gas loss can be an important escape path, in addition to gas loss through the conduit walls. This inference is consistent with observations from the Soufrière Hills Volcano, Montserrat, where gas escapes directly from the dome, and particularly along shear zones (faults) related to the conduit wall. Dynamical models of magma ascent have been developed which incorporate gas escape. The magma ascent rate is sensitive to gas escape, as the volume proportion of gas affects density, magma compressibility and rheology, resulting in both horizontal and vertical pressure gradients in the magma column to allow gas escape. Slight changes in gas loss can make the difference between explosive and effusive eruption, and multiple steady-state flow states can exist. In certain circumstances, there can be abrupt jumps between effusive and explosive activity. Overpressures develop in the ascending magma, caused primarily by the rheological stiffening of magma as gas exsolves and crystals grow. A maximum overpressure develops in the upper parts of volcanic conduits. The overpressure is typically several MPa and increases as permeability decreases. Thus, the possibility of reaching conditions for explosions increases as permeability decreases, both due to overpressure increase and the retention of more gas. Models of magma ascent from an elastic magma chamber, combined with concepts of permeability and overpressure linked to degassing, provide an explanation for the periodic patterns of dome growth with short-lived explosive activity, as in the 1980–1986 activity of Mount St Helens. Degassing of magma in conduits can also cause strong convective circulation between deep magma reservoirs and the Earth’s surface. Such circulation not only allows degassing to occur from deep reservoirs, but may also be a significant driving force for crystal differentiation.