The November 2002 eruption of Piton de la Fournaise,Réunion: tracking the pre-eruptive thermal evolution of magma using melt inclusions

The November 2002 eruption of Piton de la Fournaise in the Indian Ocean was typical of the activity of the volcano from 1999 to 2006 in terms of duration and volume of magma ejected. The first magma erupted was a basaltic liquid with a small proportion of olivine phenocrysts (Fo₈₁) that contain small numbers of melt inclusions. In subsequent flows, olivine crystals were more abundant and richer in Mg (Fo_83–84). These crystals contain numerous melt and fluid inclusions, healed fractures, and dislocation features such as kink bands. The major element composition of melt inclusions in this later olivine (Fo_83–84) is out of equilibrium with that of its host as a result of extensive post-entrapment crystallization and Fe²⁺ loss by diffusion during cooling. Melt inclusions in Fo₈₁ olivine are also chemically out of equilibrium with their hosts but to a lesser degree. Using olivine–melt geothermometry, we determined that melt inclusions in Fo₈₁ olivine were trapped at lower temperature (1,182 ± 1°C) than inclusions in Fo_83–84 olivine (1,199–1,227°C). This methodology was also used to estimate eruption temperatures. The November 2002 melt inclusion compositions suggest that they were at temperatures between 1,070°C and 1,133°C immediately before eruption and quenching. This relatively wide temperature range may reflect the fact that most of the melt inclusions were from olivine in lava samples and therefore likely underwent minor but variable amounts of post-eruptive crystallization and Fe²⁺ loss by diffusion due to their relatively slow cooling on the surface. In contrast, melt inclusions in tephra samples from past major eruptions yielded a narrower range of higher eruption temperatures (1,163–1,181°C). The melt inclusion data presented here and in earlier publications are consistent with a model of magma recharge from depth during major eruptions, followed by storage, cooling, and crystallization at shallow levels prior to expulsion during events similar in magnitude to the relatively small November 2002 eruption.     

Sulfur and chlorine degassing from primitive arc magmas: temporal changes during the 1982–1983 eruptions of Galunggung (West Java, Indonesia)

The 1982–1983 eruptions of Galunggung represent a nine-month period of intermittent volcanic activity with significant changes in explosivity and emission of volatiles. Eruptions started with Vulcanian explosions but changed gradually to Strombolian activity. Compositions of juvenile material changed from basaltic andesite to high-Mg basalt, which are among the most primitive rock types known in the Indonesian arc system. Although bulk compositions suggest a single evolution trend, we infer from the compositions of melt inclusions in olivine phenocrysts that the magmas represent derivatives of a complex spectrum of primary melts. Primitive inclusions in olivine phenocrysts from magma erupted during the Strombolian phase contain up to 2000 ppm sulfur, but concentrations decrease rapidly with increasing SiO₂ down to matrix glass values (50–100 ppm). ‘Vulcanian’ inclusions appear to be degassed before eruption (200 ppm S). Chlorine concentrations increase from 750 to 2200 ppm in Strombolian, and from 800 to 1500 in Vulcanian magmas, whereas matrix glass contains about 1000 ppm in both cases. Ash leachates show two cycles of decreasing S/Cl ratios: from 9.7 to 5.6 at the start of the activity, and from 12.2 to 2.0 after four months. As the second cycle follows upon increased seismic activity at shallow depth, it probably reflects degassing of fresh sulfur-rich magma arriving in the shallow Galunggung reservoir. In contrast to the degassed state of Vulcanian magma, the significant amounts of adsorbed sulfur on the ashes point to an excess source of sulfur, which was most likely derived from intruding Strombolian magma. Hence, the observed sulfur flux of 2 Mt is not in accordance with a petrologic estimate of 0.09 Mt. Using a published value of 550 Mt of erupted material about 0.34 km³ fresh undegassed magma is needed to account for the observed sulfur flux. This is close to the erupted volume of Vulcanian magma (0.26 km³), which presumably was replaced completely by Strombolian magma during the eruption. Using the petrologic method, we calculate a total release of 0.3 Mt chlorine, which agrees well with an output of 0.47 Mt estimated independently from S/Cl ratios of the ash leachates and TOMS sulfur yields. Ash leachates show that about 35% of the sulfur and 30% of the chlorine was scavenged from the eruption plumes. Our results suggest that sulfur and chlorine were largely decoupled during degassing, which resulted in considerable variations in S/Cl ratios during the Galunggung eruptions. We infer that sulfur degassing reflects the arrival of fresh magma at shallow depth, whereas chlorine is largely derived from simultaneously erupted material. As a consequence, the petrologic estimates are more consistent with observed emissions for chlorine than for sulfur.     

Submarine flank eruption preceding caldera subsidence during the 2000 eruption of Miyakejima Volcano, Japan

During the early part of a seismic swarm preceding eruption and caldera formation at Miyakejima Volcano, discoloured sea surfaces were observed 1.5 km off the western coast of Miyakejima on 27 June 2000. A later survey of the area using a multi-beam side scan sonar and a remotely operated small submarine revealed four craters of 20–30 m diameter aligned east-west in a 100×10–30 m area on the seafloor, with hot water at 140°C being released from one of the centres. Each crater consists of submarine spatter overlain in part by scoria lapilli. Dredged spatter from the craters was fresh, and there was no evidence of activity of marine organisms on the spatter surface, indicating that the discoloured sea surface resulted from magmatic eruption on the seafloor. This eruption occurred when a westward-propagating seismic swarm, initiated beneath Miyakejimas summit, passed through the area. Finding new magma on the seafloor demonstrates that this seismic swarm was associated with intruding magma, moving outward from beneath Miyakejima. Submarine spatter shows flattened shapes with a brittle crust formed by cooling in water, and its composition is aphyric andesite of 54 wt% SiO₂. The spatter is similar in whole rock and mineral composition to spatter erupted in 1983. However, the wide range of Cl in melt inclusions in plagioclase of the 27 June submarine spatter shows that it is not simply a remnant of the 1983 magma, which has only high Cl melt inclusions in plagioclase. The mixed character of melt inclusions suggests involvement of a magma with low Cl melt inclusions. The magma erupted explosively on 18 August from Miyakejimas summit, considered as the second juvenile magma in this eruption, contains low Cl melt inclusions in plagioclase. Based on these observations and the eruption sequence, we present the following model: (1) A shallow magma chamber was filled with a remnant of 1983 magma that had evolved to a composition of 54–55 wt% SiO₂. (2) Injection of the 18 August magma into this chamber generated a mixed magma having a wide range of Cl in melt inclusions contained plagioclase. The magma mixing might have occurred shortly before the submarine eruption and could have been a trigger for the initiation of the removal of magma from the chamber as an extensive dyke, which eventually led to caldera subsidence.Editorial responsibility: S Nakada, T Druitt     

Multiple levels of magma storage during the 1980 summer eruptions of Mount St. Helens, WA

Transitions in eruptive style—explosive to effusive, sustained to pulsatory—are a common aspect of volcanic activity and present a major challenge to volcano monitoring efforts. A classic example of such transitions is provided by the activity of Mount St. Helens, WA, during 1980, where a climactic Plinian event on May 18 was followed by subplinian and vulcanian eruptions that became increasing pulsatory with time throughout the summer, finally progressing to episodic growth of a lava dome. Here we use variations in the textures, glass compositions and volatile contents of melt inclusions preserved in pyroclasts produced by the summer 1980 eruptions to determine conditions of magma ascent and storage that may have led to observed changes in eruptive activity. Five different pyroclast types identified in pyroclastic flow and fall deposits produced by eruptions in June 12, July 22 and August 7, 1980, provide evidence for multiple levels of magma storage prior to each event. Highly vesicular clasts have H₂O-rich (4.5–5.5 wt%) melt inclusions and lack groundmass microlites or hornblende reaction rims, characteristics that require magma storage at P≥160 MPa until shortly prior to eruption. All other clast types have groundmass microlites; P_H20 estimated from both H₂O-bearing melt inclusions and textural constraints provided by decompression experiments suggest pre-eruptive storage pressures of ∼75, 40, and 10 MPa. The distribution of pyroclast types within and between eruptive deposits can be used to place important constraints on eruption mechanisms. Fall and flow deposits from June 12, 1980, lack highly vesicular, microlite-free pyroclasts. This eruption was also preceded by a shallow intrusion on June 3, as evidenced by a seismic crisis and enhanced SO₂ emissions. Our constraints suggest that magma intruded to a depth of ≤4 km beneath the crater floor fed the June eruption. In contrast, eruptions of July and August, although shorter in duration and smaller in volume, erupted deep volatile-rich magma. If modeled as a simple cylinder, these data require a step-wise decrease in effective conduit diameter from 40–50 m in May and June to 8–12 m in July and August. The abundance of vesicular (intermediate to deep) clast types in July and August further suggests that this change was effected by narrowing the shallower part of the conduit, perhaps in response to solidification of intruded magma remaining in the shallow system after the June eruption. Eruptions from July to October were distinctly pulsatory, transitioning between subplinian and vulcanian in character. As originally suggested by Scandone and Malone (1985), a growing mismatch between the rate of magma ascent and magma disruption explains the increasingly pulsatory nature of the eruptions through time. Recent fragmentation experiments Spieler et al. (2004) suggest this mismatch may have been aided by the multiple levels at which magma was stored (and degassed) prior to these events.Editorial responsibility: J Stix     

Fundamental changes in the activity of the natrocarbonatite volcano Oldoinyo Lengai, Tanzania

On September 4, 2007, after 25 years of effusive natrocarbonatite eruptions, the eruptive activity of Oldoinyo Lengai (OL), N Tanzania, changed abruptly to episodic explosive eruptions. This transition was preceded by a voluminous lava eruption in March 2006, a year of quiescence, resumption of natrocarbonatite eruptions in June 2007, and a volcano-tectonic earthquake swarm in July 2007. Despite the lack of ground-based monitoring, the evolution in OL eruption dynamics is documented based on the available field observations, ASTER and MODIS satellite images, and almost-daily photos provided by local pilots. Satellite data enabled identification of a phase of voluminous lava effusion in the 2 weeks prior to the onset of explosive eruptions. After the onset, the activity varied from 100 m high ash jets to 2–15 km high violent, steady or unsteady, eruption columns dispersing ash to 100 km distance. The explosive eruptions built up a ∼400 m wide, ∼75 m high intra-crater pyroclastic cone. Time series data for eruption column height show distinct peaks at the end of September 2007 and February 2008, the latter being associated with the first pyroclastic flows to be documented at OL. Chemical analyses of the erupted products, presented in a companion paper (Keller et al. 2010), show that the 2007–2008 explosive eruptions are associated with an undersaturated carbonated silicate melt. This new phase of explosive eruptions provides constraints on the factors causing the transition from natrocarbonatite effusive eruptions to explosive eruptions of carbonated nephelinite magma, observed repetitively in the last 100 years at OL.     

Effusive eruptions from a large silicic magma chamber: the Bearhead Rhyolite,Jemez volcanic field,NM

Large continental silicic magma systems commonly produce voluminous ignimbrites and associated caldera collapse events. Less conspicuous and relatively poorly documented are cases in which silicic magma chambers of similar size to those associated with caldera-forming events produce dominantly effusive eruptions of small-volume rhyolite domes and flows. The Bearhead Rhyolite and associated Peralta Tuff Member in the Jemez volcanic field, New Mexico, represent small-volume eruptions from a large silicic magma system in which no caldera-forming event occurred, and thus may have implications for the genesis and eruption of large volumes of silicic magma and the long-term evolution of continental silicic magma systems.⁴⁰Ar/³⁹Ar dating reveals that most units mapped as Bearhead Rhyolite and Peralta Tuff (the Main Group) were erupted during an ∼540 ka interval between 7.06 and 6.52 Ma. These rocks define a chemically coherent group of high-silica rhyolites that can be related by simple fractional crystallization models. Preceding the Main Group, minor amounts of unrelated trachydacite and low silica rhyolite were erupted at ∼11–9 and ∼8 Ma, respectively, whereas subsequent to the Main Group minor amounts of unrelated rhyolites were erupted at ∼6.1 and ∼1.5 Ma.The chemical coherency, apparent fractional crystallization-derived geochemical trends, large areal distribution of rhyolite domes (∼200 km²), and presence of a major hydrothermal system support the hypothesis that Main Group magmas were derived from a single, large, shallow magma chamber. The ∼540 ka eruptive interval demands input of heat into the system by replenishment with silicic melts, or basaltic underplating to maintain the Bearhead Rhyolite magma chamber.Although the volatile content of Main Group magmas was within the range of rhyolites from major caldera-forming eruptions such as the Bandelier and Bishop Tuffs, eruptions were smaller volume and dominantly effusive. Bearhead Rhyolite domes occur at the intersection of faults, and are cut by faults, suggesting that the magma chamber was structurally vented preventing volatiles from accumulating to levels high enough to trigger a caldera-forming eruption.     

Effects of eruption and lava drainback on the H2O contents of basaltic magmas at Kilauea Volcano

 Lava drainback has been observed during many eruptions at Kilauea Volcano: magma erupts, degasses in lava fountains, collects in surface ponds, and then drains back beneath the surface. Time series data for melt inclusions from the 1959 Kilauea Iki picrite provide important evidence concerning the effects of drainback on the H₂O contents of basaltic magmas at Kilauea. Melt inclusions in olivine from the first eruptive episode, before any drainback occurred, have an average H₂O content of 0.7±0.2 wt.%. In contrast, many inclusions from the later episodes, erupted after substantial amounts of surface degassed lava had drained back down the vent, have H₂O contents that are much lower (≥0.24 wt.% H₂O). Water contents in melt inclusions from magmas erupted at Pu'u 'O'o on the east rift zone vary from 0.39–0.51 wt.% H₂O in tephra from high fountains to 0.10–0.28 wt.% H₂O in spatter from low fountains. The low H₂O contents of many melt inclusions from Pu'u 'O'o and post-drainback episodes of Kilauea Iki reveal that prior to crystallization of the enclosing olivine host, the melts must have exsolved H₂O at pressures substantially less than those in Kilauea's summit magma reservoir. Such low-pressure H₂O exsolution probably occurred as surface degassed magma was recycled by drainback and mixing with less degassed magma at depth. Recognition of the effects of low-pressure degassing and drainback leads to an estimate of 0.7 wt.% H₂O for differentiated tholeiitic magma in Kilauea's summit magma storage reservoir. Data for MgO-rich submarine glasses (Clague et al. 1995) and melt inclusions from Kilauea Iki demonstrate that primary Kilauean tholeiitic magma has an H₂O/K₂O mass ratio of ∼1.3. At transition zone and upper mantle depths in the Hawaiian plume source, H₂O probably resides partly in a small amount of hydrous silicate melt. Received: 31 March 1997 / Accepted: 17 November 1997     

Role of precursory less-viscous mixed magma in the eruption of phenocryst-rich magma: evidence from the Hokkaido-Komagatake 1929 eruption

During the 1929 activity of Hokkaido-Komagatake volcano, the Plinian eruption of a phenocryst-rich andesite was preceded by a small eruption of more mafic magma formed by magma mixing. A similar eruption sequence has been reported for some other eruptions (Pallister et al. 1996; Venezky and Rutherford 1997), suggesting that eruption of a mixed magma is a precursor of phenocryst-rich magmas. For the purpose of understanding the tapping processes of the phenocryst-rich magma chamber, we investigated the temporal variation in the erupted magma and estimated the viscosity and density of the end-member and mixed magmas with constraints drawn from petrography. For the precursory mixed magma we estimate 33dž vol.% phenocrysts, andesitic-dacitic melt composition, 3 wt.% H₂O content, and temperature of 1040°C. In comparison, for the climactic, silicic end-member magma we estimate 48Dž vol.% phenocryst, high-silica rhyolitic melt, 3 wt.% H₂O, and temperature of 950°C, respectively. The mafic end-member magma, which was not erupted, is thought to be an almost aphyric basaltic-andesitic magma, based on mass balance calculation of the phenocryst content. The proportion of the mafic end-member magma component in the mixed magma was calculated to be 20-40 wt.%. On the basis of these data, we estimate magma viscosities of 10^3.9, 10^6.9, and 10^2.0 Pa s for the mixed, silicic end-member, and mafic end-member magmas, respectively. The calculated density differences among these magmas are inconsequential when possible errors are considered. We calculate the minimum excess pressure required for dike propagation to be 31 MPa for the silicic end-member magma and 8 MPa for the mixed magma, using the estimated viscosity and dike propagation model of Rubin (1995). If we assume that excess pressure is limited by the wall rock strength of the magma chamber, excess pressure retainable in the magma chamber is less than ca. 20 MPa. This suggests that the mixed magma was able to ascend to the surface without freezing, whereas the viscous silicic end-member magma could not. The formation and precursory eruption of the mixed magma are, therefore, effective and necessary initiation processes for the phenocryst-rich, viscous magma eruption.     

Deep pooling of low degree melts and volatile fluxes at the 85°E segment of the Gakkel Ridge: Evidence from olivine-hosted melt inclusions and glasses

We present new analyses of volatile, major, and trace elements for a suite of glasses and melt inclusions from the 85°E segment of the ultra-slow spreading Gakkel Ridge. Samples from this segment include limu o pele and glass shards, proposed to result from CO₂-driven explosive activity. The major element and volatile compositions of the melt inclusions are more variable and consistently more primitive than the glass data. CO₂ contents in the melt inclusions extend to higher values (167–1596 ppm) than in the co-existing glasses (187–227 ppm), indicating that the melt inclusions were trapped at greater depths. These melt inclusions record the highest CO₂ melt concentrations observed for a ridge environment. Based on a vapor saturation model, we estimate that the melt inclusions were trapped between seafloor depths (～ 4 km) and ～ 9 km below the seafloor. However, the glasses are all in equilibrium with their eruption depths, which is inconsistent with the rapid magma ascent rates expected for explosive activity. Melting conditions inferred from thermobarometry suggest relatively deep (25–40 km) and cold (1240°–1325 °C) melting conditions, consistent with a thermal structure calculated for the Gakkel Ridge. The water contents and trace element compositions of the melt inclusions and glasses are remarkably homogeneous; this is an unexpected result for ultra-slow spreading ridges, where magma mixing is generally thought to be less efficient based on the assumption that steady-state crustal magma chambers are absent in these environments. All melts can be described by a single liquid line of descent originating from a pooled melt composition that is consistent with the aggregate melt calculated from a geodynamic model for the Gakkel Ridge. These data suggest a model in which deep, low degree melts are efficiently pooled in the upper mantle (9–20 km depth), after which crystallization commences and continues during ascent and eruption. Based on our melting model and the assumption that CO₂ is perfectly incompatible, we show that the highest CO₂ concentrations of the melt inclusions (～ 1600 ppm) are consistent with the calculated CO₂ concentrations of primary undegassed melts. The highest measured CO₂/Nb ratio (443) of Gakkel Ridge melt inclusions predicts a mantle CO₂ content of 134 ppm and would result in a global ridge flux of 2.0 × 10¹² mol CO₂/yr.     

Behavioral patterns of Fuego volcano, Guatemala

The times of activity at Fuego (one of the most active volcanoes in the world) since 1800 correlate with the activity of other Central American volcanoes. Approximately 0.7 km³ of olivine-bearing, high-Al₂O₃ basalt has been erupted since 1932, and about 1.7 km³ has been produced during 450 years of historic records. A minimum of 13,000 years and a maximum of 100,000 years were required to build Fuego's cone of 50 km³. Within the recent cluster of activity since 1932, rates of magma production have increased to 0.5 m³/s and the trend has been toward more eruptions (shorter reposes) of progressively more mafic basalt. 47% of the eruptions occurred within 2 days of the fortnightly tidal maximum and 56% occurred within 2 hours of the semi-diurnal minimum of the vertical tidal gravity acceleration. Thus the maximum compressional component of the tidal cycles can trigger an eruption at Fuego. Eruptions with higher effusion rates produce larger volumes of materials, although they only last a few hours. The 20–70 year clusters of activity beginning at 80–170-year intervals are interpreted as reflecting the ascent of primary batches of magma. A deeper (8–16 km), larger (> 1 km³) primary chamber and a shallower (2–5 km), smaller (0.1 km³), dike-like secondary chamber best explain Fuego's behavioral pattern.     

Magma degassing and basaltic eruption styles: a case study of 2000 year BP Xitle volcano in central Mexico

To investigate the relationship between volatile abundances and eruption style, we have analyzed major element and volatile (H₂O, CO₂, S) concentrations in olivine-hosted melt inclusions in tephra from the 2000 yr BP eruption of Xitle volcano in the central Trans-Mexican Volcanic Belt. The Xitle eruption was dominantly effusive, with fluid lava flows accounting for 95% of the total dense rock erupted material (1.1 km³). However, in addition to the initial, Strombolian, cinder cone-building phase, there was a later explosive phase that interrupted effusive activity and deposited three widespread ash fall layers. Major element compositions of olivine-hosted melt inclusions from these ash layers range from 52 to 58 wt.% SiO₂, and olivine host compositions are Fo_84–86. Water concentrations in the melt inclusions are variable (0.2–1.3 wt.% H₂O), with an average of 0.45±0.3 (1σ) wt.% H₂O. Sulfur concentrations vary from below detection (50 ppm) to 1000 ppm but are mostly ≤200 ppm and show little correlation with H₂O. Only the two inclusions with the highest H₂O have detectable CO₂ (310–340 ppm), indicating inclusion entrapment at higher pressures (700–900 bars) than for the other inclusions (≤80 bars). The low and variable H₂O and S contents of melt inclusions combined with the absence of less soluble CO₂ indicates shallow-level degassing before olivine crystallization and melt inclusion formation. Olivine morphologies are consistent with the interpretation that most crystallization occurred rapidly during near-surface H₂O loss. During cinder cone eruptions, the switch from initial explosive activity to effusive eruption probably occurs when the ascent velocity of magma becomes slow enough to allow near-complete degassing of magma at shallow depths within the cone as a result of buoyantly rising gas bubbles. This allows degassed lavas to flow laterally and exit near the base of the cone while gas escapes through bubbly magma in the uppermost part of the conduit just below the crater. The major element compositions of melt inclusions at Xitle show that the short-lived phase of renewed explosive activity was triggered by a magma recharge event, which could have increased overpressure in the storage reservoir beneath Xitle, leading to increased ascent velocities and decreased time available for degassing during ascent.     

Explosive volcanism (VEI 6) without caldera formation: insight from Huaynaputina volcano, southern Peru

Through examination of the vent region of Volcán Huaynaputina, Peru, we address why some major explosive eruptions do not produce an equivalent caldera at the eruption site. Here, in 1600, more than 11 km³ DRE (VEI 6) were erupted in three stages without developing a volumetrically equivalent caldera. Fieldwork and analysis of aerial photographs reveal evidence for cryptic collapse in the form of two small subsidence structures. The first is a small non-coherent collapse that is superimposed on a cored-out vent. This structure is delimited by a partial ring of steep faults estimated at 0.85 by 0.95 km. Collapse was non-coherent with an inwardly tilted terrace in the north and a southern sector broken up along a pre-existing local fault. Displacement was variable along this fault, but subsidence of approximately 70 m was found and caused the formation of restricted extensional gashes in the periphery. The second subsidence structure developed at the margin of a dome; the structure has a diameter of 0.56 km and crosscuts the non-coherent collapse structure. Subsidence of the dome occurred along a series of up to seven concentric listric faults that together accommodate approximately 14 m of subsidence. Both subsidence structures total 0.043 km³ in volume, and are much smaller than the 11 km³ of erupted magma. Crosscutting relationships show that subsidence occurred during stages II and III when ∼2 km³ was erupted and not during the main plinian eruption of stage I (8.8 km³). The mismatch in erupted volume vs. subsidence volume is the result of a complex plumbing system. The stage I magma that constitutes the bulk of the erupted volume is thought to originate from a ∼20-km-deep regional reservoir based on petrological constraints supported by seismic data. The underpressure resulting from the extraction of a relatively small fraction of magma from the deep reservoir was not sufficient enough to trigger collapse at the surface, but the eruption left a 0.56-km diameter cored-out vent in which a dome was emplaced at the end of stage II. Petrologic evidence suggests that the stage I magma interacted with and remobilized a shallow crystal mush (∼4–6 km) that erupted during stage II and III. As the crystal mush erupted from the shallow reservoir, depressurization led to incremental subsidence of the non-coherent collapse structure. As the stage III eruption waned, local pressure release caused subsidence of the dome. Our findings highlight the importance of a connected magma reservoir, the complexity of the plumbing system, and the pattern of underpressure in controlling the nature of collapse during explosive eruptions. Huaynaputina shows that some major explosive eruptions are not always associated with caldera collapse. Editorial responsibility: J Stix     

Phenocryst-hosted melt inclusions record stalling of magma during ascent in the conduit and upper magma reservoir prior to vulcanian explosions, Soufrière Hills volcano, Montserrat, West Indies

The mechanics of explosive eruptions influence magma ascent pathways. Vulcanian explosions involve a stop–start mechanism that recurs on various timescales, evacuating the uppermost portions of the conduit. During the repose time between explosions, magma rises from depth and refills the conduit and stalls until the overpressure is sufficient to generate another explosion. We have analyzed major elements, Cl, S, H₂O, and CO₂ in plagioclase-hosted melt inclusions, sampled from pumice erupted during four vulcanian events at Soufrière Hills volcano, Montserrat, to determine melt compositions prior to eruption. Using Fourier transform infrared spectroscopy, we measured values up to 6.7 wt.% H₂O and 80 ppm CO₂. Of 42 melt inclusions, 81 % cluster between 2.8 and 5.4 wt.% H₂O (57 to 173 MPa or 2–7 km), suggesting lower conduit to upper magma reservoir conditions. We propose two models to explain the magmatic conditions prior to eruption. In Model 1, melt inclusions were trapped during crystal growth in magma that was stalled in the lower conduit to upper magma reservoir, and during trapping, the magma was undergoing closed-system degassing with up to 1 wt.% free vapor. This model can explain the melt inclusions with higher H₂O contents since these have sampled the upper parts of the magma reservoir. However, the model cannot explain the melt inclusions with lower H₂O because the timescale for plagioclase crystallization and melt inclusion entrapment is longer than the magma residence time in the conduit. In Model 2, melt inclusions were originally trapped at deeper levels of the magma chamber, but then lost hydrogen by diffusion through the plagioclase host during periodic stalling of the magma in the lower conduit system. In this second scenario, which we favor, the melt inclusions record re-equilibration depths within the lower conduit to upper magma reservoir.     

Petrology and geochemistry of lava and ash erupted from Volcán Colima, Mexico, during 1998–2005

Lava (n = 8) and bulk ash samples (n = 6) erupted between July 1999 and June 2005 were investigated to extend time-series compositional and textural studies of the products erupted from Volcán Colima since 1869. In particular, we seek to evaluate the possibility that the current activity will culminate in major explosive Plinian-style event similar to that in 1913. Lava samples continue to show relatively heterogeneous whole-rock compositions with some significant mafic spikes (1999, 2001) as have prevailed since 1976. Groundmass SiO₂ contents continue trends to lower levels that have prevailed since 1961, in the direction of the still lower groundmass SiO₂ contents found in 1913 scoriae. Importantly, ash samples from investigated Vulcanian-style explosive eruptions in 2005 are devoid of particles with micro-vesiculated groundmass textures; such textures characterized the 1913 scoriae, signifying expansion of in-situ magmatic gas as the propellant of the 1913 eruption. All magmas erupted since 1913 appear to have arrived in the upper volcanic conduit system in a degassed state. The small to moderate Vulcanian-style explosive eruptions, which have been common since 1999 (> 16,000 events), have blasted ash clouds as high as 11 km a.s.l. and sent pyroclastic flows out to distances of 5 km. These eruptions do not appear to be powered by expansion of in-situ magmatic gas. New small lava domes have been observed in the crater prior to many explosive eruptions. These plugs of degassed lava may temporarily seal the conduit and allow the build-up of magmatic gases streaming upward from below ahead of rising and degassing magma. In this interpretation, when gas pressure exceeds the strength of the plug seal in the upper conduit, an explosive Vulcanian-style eruption occurs. Alternatively these explosive eruptions may represent interactions of hot rock and groundwater (phreato-magmatic).     

Breadcrust bombs as indicators of Vulcanian eruption dynamics at Guagua Pichincha volcano,Ecuador

Vulcanian eruptions are common at many volcanoes around the world. Vulcanian activity occurs as either isolated sequences of eruptions or as precursors to sustained explosive events and is interpreted as clearing of shallow plugs from volcanic conduits. Breadcrust bombs characteristic of Vulcanian eruptions represent samples of different parts of these plugs and preserve information that can be used to infer parameters of pre-eruption magma ascent. The morphology and preserved volatile contents of breadcrust bombs erupted in 1999 from Guagua Pichincha volcano, Ecuador, thus allow us to constrain the physical processes responsible for Vulcanian eruption sequences of this volcano. Morphologically, breadcrust bombs differ in the thickness of glassy surface rinds and in the orientation and density of crack networks. Thick rinds fracture to create deep, widely spaced cracks that form large rectangular domains of surface crust. In contrast, thin rinds form polygonal networks of closely spaced shallow cracks. Rind thickness, in turn, is inversely correlated with matrix glass water content in the rind. Assuming that all rinds cooled at the same rate, this correlation suggests increasing bubble nucleation delay times with decreasing pre-fragmentation water content of the melt. A critical bubble nucleation threshold of 0.4–0.9 wt% water exists, below which bubble nucleation does not occur and resultant bombs are dense. At pre-fragmentation melt H₂O contents of >∼0.9 wt%, only glassy rinds are dense and bomb interiors vesiculate after fragmentation. For matrix glass H₂O contents of ≥1.4 wt%, rinds are thin and vesicular instead of thick and non-vesicular. A maximum measured H₂O content of 3.1 wt% establishes the maximum pressure (63 MPa) and depth (2.5 km) of magma that may have been tapped during a single eruptive event. More common H₂O contents of ≤1.5 wt% suggest that most eruptions involved evacuation of ≤1.5 km of the conduit. As we expect that substantial overpressures existed in the conduit prior to eruption, these depth estimates based on magmastatic pressure are maxima. Moreover, the presence of measurable CO₂ (≤17 ppm) in quenched glass of highly degassed magma is inconsistent with simple models of either open- or closed-system degassing, and leads us instead to suggest re-equilibration of the melt with gas derived from a deeper magmatic source. Together, these observations suggest a model for the repeated Vulcanian eruptions that includes (1) evacuation of the shallow conduit during an individual eruption, (2) depressurization of magma remaining in the conduit accompanied by open-system degassing through permeable bubble networks, (3) rapid conduit re-filling, and (4) dome formation prior to the subsequent explosion. An important part of this process is densification of upper conduit magma to allow repressurization between explosions. At a critical overpressure, trapped pressurized gas fragments the nascent impermeable cap to repeat the process.     

Influence of pre-eruptive storage conditions and volatile contents on explosive Plinian style eruptions of basic magma

Sub-Plinian to Plinian eruptions of basic magma present a challenge to modeling volcanic behavior because many models rely on magma becoming viscous enough during ascent to behave brittlely and cause fragmentation. Such models are unable, however, to strain low viscosity magma fast enough for it to behave brittlely. That assumes that such magmas actually have low viscosities, but the rare Plinian eruptions of basic magma may in fact result from them being anomalously viscous. Here, we examine two such eruptions, the 122 B.C. eruption of hawaiitic basalt from Mt. Etna and the late Pleistocene eruption of basaltic andesite from Masaya Caldera, to test whether they were anomalously viscous. We carried out hydrothermal experiments on both magmas and analyzed glass inclusions in plagioclase phenocrysts from each to determine their most likely pre-eruptive temperatures and water contents. We find that the hawaiite was last stored at 1,000–1,020°C, whereas the basaltic andesite was last stored at 1,010–1,060°C, and that both were water saturated with ∼3.0 wt.% water dissolved in them. Such water contents are not high enough to trigger Plinian explosive behavior, as much more hydrous basic magmas erupt less violently. In addition, despite being relatively cool, the viscosities of both magmas would range from ∼10^2.2–2.5 Pa s before erupting to ∼10⁴ Pa s when essentially degassed, all of which are too fluid to cause brittle disruption. Without invoking special external forces to explain all such eruptions, one of the more plausible explanations is that when the bubble content reaches some critical value the fragile foam-like magma disrupts. The rarity of Plinian eruptions of basic magma may be because such magmas must ascend fast enough to retain their bubbles.     

Determination of pre-eruptive H2O,F and Cl contents of silicic magmas using melt inclusions: Examples from Taupo volcanic center,New Zealand

Water, F, and Cl contents of melt inclusions in phenocrysts from the 2-ka-old Taupo and Hatepe plinian tephras, and the 22-ka-old Okaia tephra from the Taupo volcanic center, New Zealand, were measured by electron and ion microprobe. Major and trace element chemistry of the inclusions is similar to that of bulk rock, supporting our assumption that volatile contents of inclusions are representative of the magma in which the crystals grew. Inclusions in the 2-ka Taupo plinian tephra contain a mean of 4.3 wt% H₂O, 450 ppm F, and 1700 ppm Cl; from the Hatepe plinian tephra 4.3 wt% H₂O, 430 ppm F, and 1700 ppm Cl; and from the Okaia tephra 5.9 wt% H₂O, 470 ppm F, and 2100 ppm Cl. Sulfur was below the detection limit of 200 ppm. The constant H₂O, F and Cl from a number of stratigraphic horizons in the tephra deposits suggest that the Taupo and Hatepe plinian tephras (>8.2 km³ magma volume) were derived from a magma body that did not contain a strong volatile gradient. By inference, there is no pre-eruptive volatile difference between these plinian eruptions and a phrea-toplinian eruption which occurred between the two. Virtually no major element zonation is seen in this eruptive sequence. Although the Okaia tephra was also erupted from the Taupo volcanic center, probably from a similar vent area, its higher volatile contents and distinct composition as compared to the Taupo tephras show that it was derived from a different, and possibly deeper, magma body.     

Erratum to Compositional variations and magma mixing in the 1991 eruptions of Hudson volcano,Chile

The August 1991 eruptions of Hudson volcano produced ~2.7 km³ (dense rock equivalent, DRE) of basaltic to trachyandesitic pyroclastic deposits, making it one of the largest historical eruptions in South America. Phase 1 of the eruption (P1, April 8) involved both lava flows and a phreatomagmatic eruption from a fissure located in the NW corner of the caldera. The paroxysmal phase (P2) began several days later (April 12) with a Plinian-style eruption from a different vent 4 km to the south-southeast. Tephra from the 1991 eruption ranges in composition from basalt (phase 1) to trachyandesite (phase 2), with a distinct gap between the two erupted phases from 54–60 wt% SiO₂. A trend of decreasing SiO₂ is evident from the earliest part of the phase 2 eruption (unit A, 63–65 wt% SiO₂) to the end (unit D, 60–63 wt% SiO₂). Melt inclusion data and textures suggest that mixing occurred in magmas from both eruptive phases. The basaltic and trachyandesitic magmas can be genetically related through both magma mixing and fractional crystallization processes. A combination of observed phase assemblages, inferred water content, crystallinity, and geothermometry estimates suggest pre-eruptive storage of the phase 2 trachyandesite at pressures between ~50–100 megapascal (MPa) at 972 ± 26°C under water-saturated conditions (log fO₂ –10.33 (±0.2)). It is proposed that rising P1 basaltic magma intersected the lower part of the P2 magma storage region between 2 and 3 km depth. Subsequent mixing between the two magmas preferentially hybridized the lower part of the chamber. Basaltic magma continued advancing towards the surface as a dyke to eventually be erupted in the northwestern part of the Hudson caldera. The presence of tachylite in the P1 products suggests that some of the magma was stalled close to the surface (<0.5 km) prior to eruption. Seismicity related to magma movement and the P1 eruption, combined with chamber overpressure associated with basalt injection, may have created a pathway to the surface for the trachyandesite magma and subsequent P2 eruption at a different vent 4 km to the south-southeast. Electronic supplementary material  The online version of this article (doi:) contains supplementary material, which is available to authorized users.     

Petrology and sulfur and chlorine emissions of the 1963 eruption of Gunung Agung,Bali, Indonesia

 The 1963 eruption of Gunung Agung produced 0.95 km³ dense rock equivalent (DRE) of olivine±hornblende-bearing, weakly phyric, basaltic andesite tephra and lava. Evidence for magma mixing in the eruptive products includes whole-rock compatible and incompatible trace element trends, reverse and complex compositional zoning of mineral phases, disequilibrium mineral assemblages, sieve-textured plagioclase phenocrysts, and augite rims on reversely zoned orthopyroxene. Basalt magma mixed with pre-existing andesite magma shortly before eruption to yield basaltic andesite with a temperature of 1040–1100  °C at an assumed pressure of 2 kb, f O₂>NNO, and an average melt volatile content (H₂O±CO₂) of 4.3 wt.%. Magma-mixing end members may have provided some of the S and Cl emitted in the eruption. Glass inclusions in phenocrysts contain an average of 650 ppm S and 3130 ppm Cl as compared with 70 ppm and 2220 ppm, respectively, in the matrix glass. Maximum S and Cl contents of glass inclusions approach 1800 and 5000 ppm, respectively. Application of the petrologic method to products of the 1963 eruption for estimating volatile release yields of 2.5×10¹²g (Mt) of SO₂ and 3.4 Mt of Cl released from the 0.65 km³ of juvenile tephra which contributed to stratospheric injection of H₂SO₄ aerosols on 17 March and 16 May, when eruption column heights exceeded 20 km above sea level. An independent estimate of SO₂ release from atmospheric aerosol loading (11–12 Mt) suggests that approximately 7 Mt of SO₂ was injected into the stratosphere. The difference between the two estimates can be most readily accounted for by the partitioning of S, as well as some Cl, from the magma into a water-rich vapor phase which was released upon eruption. For other recent high-S-release eruptions of more evolved and oxidized magmas (El Chichón, Pinatubo), the petrologic method gives values two orders of magnitude less than independent estimates of SO₂ emissions. Results from this study of the Agung 1963 magma and its volatile emissions, and from related studies on eruptions of more mafic magmas, suggest that SO₂ emissions from eruptions of higher-S-solubility magma may be more reliably estimated by the petrologic method than may those from more-evolved magma eruptions. Received: 29 June 1994 / Accepted: 25 April 1996     

Variation of volatile concentration in a magma system of Satsuma-Iwojima volcano deduced from melt inclusion analyses

Chemical analyses of 30 melt inclusions from Satsuma-Iwojima volcano, Japan, were carried out to investigate volatile evolution in a magma chamber beneath the volcano from about 6300 yr BP to the present. Large variations in volatile concentrations of melts were observed. (1) Water concentration of rhyolitic melts decreases with time; 3–4.6 wt.% at the time of latest caldera-forming eruption of Takeshima pyroclastic flow deposit (ca. 6300 yr BP), 3 wt.% for small pyroclastic flow (ca. 1300 yr BP) of Iwodake, post-caldera rhyolitic dome, and 0.7–1.4 wt.% for submarine lava eruption (Showa-Iwojima) in 1934. (2) Rhyolitic melts of the Takeshima and Iwodake eruptions contained CO₂ of less than 40 ppm, while the Showa-Iwojima melt has higher CO₂ concentration of up to 140 ppm. (3) Water and CO₂ concentrations of basaltic to andesitic melt of Inamuradake, a post-caldera basaltic scoria cone, are 1.2–2.8 wt.% and ≤290 ppm, respectively.Volatile evolution in the magma chamber is interpreted as follows: (1) the rhyolitic magma at the time of the latest caldera-forming eruption (ca. 6300 yr BP) was gas-saturated due to pressure variation in the magma chamber because the large variation in water concentration of the melt was attributed to exsolution of volatile in the magma prior to the eruption. Iwodake eruption (ca. 1300 yr BP) was caused by a remnant of the caldera-forming rhyolitic magma, suggested from the similarity of major element composition between these magmas. (2) Volatile composition of the Showa-Iwojima rhyolitic melt agrees with that of magmatic gases presently discharging from a summit of Iwodake, indicating the low pressure degassing condition. (3) The degassing of the magma chamber by magma convection in a conduit of Iwodake during non-eruptive but active degassing period for longer than 800 years decreased water concentration of the rhyolitic magma. (4) Geological and petrological observations indicate that a stratified magma chamber, which consists of a lower basaltic layer and an upper rhyolitic layer, might have existed during the post-caldera stage. Addition of CO₂ from the underlying basaltic magma to the upper gas-undersaturated (degassed) rhyolitic magma increased CO₂ concentration of the rhyolitic magma.