Laboratory investigations of viscous effects in replenished magma chambers

Some laboratory experiments are described which investigate the dynamical effects of replenishment of a magma chamber containing high viscosity magma by hotter, denser and much more fluid magma. In the experiments a layer of hot KNO₃ solution is emplaced beneath cold glycerine, which has a viscosity 3000 times greater. Less dense fluid is released immediately and continuously from the interface as a result of crystallization in the lower layer and rises as plumes through the overlying glycerine. Further crystallization occurs in the plumes, and the crystals fall out; but there is little mixing between the two fluids and a layer of depleted KNO₃ solution forms at the top. The experiments demonstrate that interfacial processes begin to dominate where there are large viscosity differences between adjacent fluid layers as would be the case in a rhyolitic magma chamber replenished by basaltic magma.     

Simultaneous crystallization and melting at both the roof and floor of crustal magma chambers: Experimental study using NH4Cl–H2O binary eutectic system

Thermal and compositional evolution of magmas after emplacement of basalt into continental crust has been investigated by means of fluid dynamic experiments using a cold solid mixture with eutectic composition and a hot liquid with higher salinity in the NH₄Cl–H₂O binary eutectic system. The experiments were designed to simulate cases where crystallization of a basalt magma is accompanied by melting at both the roof and floor of a crustal magma chamber. The results show that thermal and compositional convection occur simultaneously in the solution; the thermal convection is driven by cooling at the roof and the compositional convection is driven by melting and crystallization at the floor. The roof was rapidly melted by the convective heat flux, which resulted in formation of a separate eutectic melt layer (the upper liquid layer) with negligible mixing of the underlying liquid (the lower liquid layer). On the other hand, a mushy layer formed at the floor. The compositional convection at the floor carried a low heat flux, so that the heat transfer at the floor was basically explained by simple heat conduction. The thermal boundary layer in the lower liquid layer at the interface with the upper liquid layer became thicker with time and subsequently temperature decreased upward throughout the lower liquid layer. Compositional gradient with NH₄Cl content decreasing upward formed by compositional convection in the lower liquid layer. The formation of these gradients resulted in formation of double-diffusive convecting layers in the lower liquid layer. The upward heat transfer was suppressed when compared with the case where the liquid region is homogenized by vigorous convection.These experimental results imply that, when a basalt magma is emplaced in continental crust, floor melting does not always enhance the cooling of the magma, but it may even reduce the total heat loss from the magma to the crusts due to suppression of convection by formation of a stabilizing compositional gradient.     

Experimental study on the effects of crustal temperature and composition on assimilation with fractional crystallization at the floor of magma chambers

When a hot basaltic magma is emplaced into continental crust or a pre-existing silicic magma chamber, the processes of assimilation with fractional crystallization (AFC) are likely to control the liquid line of descent of the magma. These processes are particularly important at the floor of the magma chamber because evolved light liquids generated by floor melting readily mix with the overlying basaltic magma. In order to clarify the effects of temperature and composition of the floor on the AFC processes, we experimentally investigated simultaneous melting and crystallization of a NH₄Cl–H₂O binary eutectic system. In the experiments, evolution of temperature and compositional profiles of a hot solution overlying a cold solid mixture of variable initial temperatures and compositions were measured. The initial NH₄Cl concentrations of solid and liquid are chosen to be higher than the eutectic composition, such that the density change of the experimental material by crystallization and melting is qualitatively the same as that of natural magmas and crusts. The results show that a mushy layer forms at the floor due to simultaneous crystallization and (partial) melting and that the liquid evolves due to mixing with liquids released by crystallization and melting. The ratio of melting mass to crystallization mass (M/C ratio) depends on the initial floor temperature and composition. As the initial floor temperature decreases, the rate of melting largely decreases, so that the M/C ratio becomes smaller. As the initial NH₄Cl concentration of the solid floor decreases, the degree of partial melting of the floor increases; however, it does not necessarily result in an increase in the M/C ratio. The higher melt fraction of the mushy layer increases permeability within the mushy layer, so that vertical exchange between the liquid in the mushy layer and the more concentrated overlying liquid is enhanced. This effect promotes crystallization in the mushy layer, and decreases the M/C ratio. It is suggested that the M/C ratio during AFC processes depends on details of the mixing process in the liquid layer such as spacing and meandering of buoyant plumes.     

An experimental investigation of volatile exsolution in evolving magma chambers

Previous laboratory experiments investigating the fluid dynamics of replenished magma chambers have been extended to model effects resulting from the release of gas. Turbulent transfer of heat between a layer of dense, hot and volatile-rich mafic magma overlying cooler more evolved magma can lead to crystallization and exsolution of volatiles in the lower layer. Small gas bubbles can cause the bulk density to decrease to that of the upper layer and thus produce sudden overturning and initiate mixing, followed by further exsolution of gas and explosive eruption. These processes have been modelled in the laboratory using a chemical reaction between sodium or potassium carbonate and nitric acid to release small bubbles of CO₂. We have investigated both the initial overturning produced by gas release in the lower layer, and the subsequent evolution of gas due to intimate mixing of the two layers. The latter experiments, in which the reactants remained isolated in the two layers until overturning occurred, demonstrated unambiguously that the fluxes of chemical components across the interfaces between convecting layers are very slow compared to the flux of heat. This shows that the evolution of layers of magma of different origins and composition can take place nearly independently of each other. The magmas can coexist in the same stratified chamber, until their bulk densities become equal and they mix together. The processes illustrated in these experiments could occur in H₂O-bearing magmas such as in the calcalkaline association and in CO₂-bearing mafic magmas such as in silica undersaturated suites.     

Creep of crystals: J.P. Poirier,Cambridge University Press, 1985, xii + 260 pp., £27.50, US$ 49.50, ISBN 0-521-26177-5

Fractional crystallization behaviour of a magma ocean extending to lower mantle depths was deduced from estimations of melting relations for the deep mantle and the density relationships between ultrabasic liquid and mantle minerals. The accretional growth of the Earth necessarily involves a molten zone (magma ocean) in the outer layer of the growing Earth. The fractionation by melting during accretion results in primary stratification composed of a molten ultrabasic upper mantle (magma ocean), a perovskite-rich lower mantle, and an iron core. A certain amount of Al₂O₃ and CaO was removed from the magma ocean and retained in the lower mantle due to eclogite fractionation in the early stage of accretion and the perovskite fractionation in the later stage of accretion. Models of the stratification of the upper mantle arising from fractional crystallization of the magma ocean and subsequent convective disturbance were deduced on the basis of estimations of melting relations for the deep mantle and the density relationships between the ultrabasic liquid and mantle minerals. The stratification of the mantle, which is consistent with geophysical constraints is as follows; the upper mantle is composed of two layers, the upper olivine-rich layer and the lower garnet-rich layer with a thickness around 200 km, and the lower mantle with a perovskite-rich composition. In this model, both the 400 and 650 km discontinuities are the chemical boundaries.     

Water contents, temperatures and diversity of the magmas of the catastrophic eruption of Nevado del Ruiz, Colombia, November 13, 1985

The petrology of the highly phyric two-pyroxene andesitic to dacitic pyroclastic rocks of the November 13, 1985 eruption of Nevado del Ruiz, Colombia, reveals evidence of: (1) increasingly fractionated bulk compositions with time; (2) tapping of a small magma chamber marginally zoned in regard to H₂O contents (1 to 4%), temperature (960–1090°C), and amount of residual melt (35 to 65%); (3) partial melting and assimilation of degassed zones in the hotter less dense interior of the magma chamber; (4) probable heating, thermal disruption and mineralogic and compositional contamination of the magma body by basaltic magma “underplating”; and (5) crustal contamination of the magmas during ascent and within the magma chamber. Near-crater fall-back or “spill-over” emitted in the middle of the eruptive sequence produced a small pyroclastic flow that became welded in its central and basal portions because of ponding and thus heat conservation on the flat glaciated summit near the Arenas crater. The heterogeneity of Ruiz magmas may be related to the comparatively small volume (0.03 km³) of the eruption, nearly ten times less than the 0.2 km³ of the Plinian phase of Mount St. Helens, and probable steep thermal and P_H2O gradients of a small source magma chamber, estimated at 300 m long and 100 m wide for an assumed ellipsoidal shape.     

Potassium enrichment by magma mixing and vapor exsolution: an example in the miocene volcanism of eastern Morocco

Miocene volcanism in eastern Morocco is comprised of potash rich calc-alkalic and alkaline rocks. In the southern part of the area, at the base of the Guilliz massif, basic inclusions are found in latites. From a petrological, geochemical and mineralogical study of both latites and their inclusions, it appear that the inclusions represent a basic liquid quenched in the latitic magma, inside the magma chamber. As a result of the drop in pressure and crystallization, a K-rich vapour phase separated from the inclusion-forming liquid and percolated the latitic magma, increasing the K₂O content of the latter and possibly triggering eruptions. The mixing process between the two magmas seems supported by density and viscosity estimations. Calculations show that for temperatures ranging from 700 °C to 1000 °C and H₂O contents from 1.6 to 5%, the inclusion forming liquid is less dense than the latitic liquid and can therefore ascend into the latitic magma by interface disequilibrium and flotation.     

Melt inclusion analysis of the Unzen 1991–1995 dacite: implications for crystallization processes of dacite magma

Dacitic magma, a mixture of high-temperature (T) aphyric magma and low-T crystal-rich magma, was erupted during the 1991–1995 Mount Unzen eruptive cycle. Here, the crystallization processes of the low-T magma were examined on the basis of melt inclusion analysis and phase relationships. Variation in water content of the melt inclusions (5.1–7.2 wt% H₂O) reflected the degassing history of the low-T magma ascending from deeper levels (250 MPa) to a shallow magma chamber (140 MPa). The ascent rate of the low-T magma decreased markedly towards the emplacement level as crystal content increased. Cooling of magma as well as degassing-induced undercooling drove crystallization. With the decreasing ascent rate, degassing-induced undercooling decreased in importance, and cooling became more instrumental in crystallization, causing local and rapid crystallization along the margin of the magma body. Some crystals contain scores of melt inclusions, whereas there are some crystals without any inclusions. This heterogeneous distribution suggests the variation in the crystallization rate within the magma body; it also suggests that cooling was dominant cause for melt entrapment. Numerical calculations of the cooling magma body suggest that cooling caused rapid crystal growth and enhanced melt entrapment once the magma became a crystal-rich mush with evolved interstitial melt. The rhyolitic composition of melt inclusions is consistent with this model.Editorial responsibility: H Shinohara     

Intracrustal origin of Arenal basaltic andesite in the light of solid–melt interactions and related compositional buffering

The origin of Arenal basaltic andesite can be explained in terms of fractional crystallization of a parental high-alumina basalt (HAB), which assimilates crustal rocks during its storage, ascent and evolution. Contamination of this melt by Tertiary calc-alkalic intrusives (quartz–diorite and granite, with ⁸⁷Sr/⁸⁶Sr ratios ranging 0.70381–0.70397, nearly identical with those of the Arenal lavas) occurs at upper crustal levels, following the interaction of ascending basaltic magma masses with gabbroic–anorthositic layers. Fragments of these layers are found as inclusions within Arenal lavas and tephra and may show reaction rims (1–5 mm thick, consisting of augite, hypersthene, bytownitic–anorthitic plagioclase, and granular titanomagnetite) at the gabbro–lava interface. These reaction rims indicate that complete `assimilation' was prevented since the temperature of the host basaltic magma was not high enough to melt the gabbroic materials (whose mineral phases are nearly identical to the early formed liquidus phases in the differentiating HAB). Olivine gabbros crystallized at pressure of about 5–6 kbar and equilibrated with the parental HAB at pressures of 3–6 kbar (both under anhydrous and hydrous conditions), and temperatures ranging 1000–1100°C. In particular, `deeper' interactions between the mafic inclusions and the hydrous basaltic melt (i.e., with about 3.5 wt.% H₂O) are likely to occur at 5.4 (±0.4) kbar and temperatures approaching 1100°C. The olivine gabbros are thus interpreted as cumulates which represent crystallized portions of earlier Arenal-type basalts. Some of the gabbros have been `mildly' tectonized and recrystallized to give mafic granulites that may exhibit a distinct foliation. Below Arenal volcano a zoned magma chamber evolved prior the last eruptive cycle: three distinct andesitic magma layers were produced by simple AFC of a high-alumina basalt (HAB) with assimilation of Tertiary quartz–dioritic and granitic rocks. Early erupted 1968 tephra and 1969 lavas (which represent the first two layers of the upper part of a zoned magma chamber) were produced by simple AFC, with fractionation of plagioclase, pyroxene and magnetite and concomitant assimilation of quartz–dioritic rocks. Assimilation rates were constant (r<sub>1</sub>=0.33) for a relative mass of magma remaining of 0.77–0.80, respectively. Lavas erupted around 1974 are less differentiated and represent the `primitive andesitic magma type' residing within the middle–lower part of the chamber. These lavas were also produced by simple AFC: assimilation rates and the relative mass of magma remaining increased of about 10%, respectively (r₁=0.36, and F=0.89). Ba enrichment of the above lavas is related to selective assimilation of Ba from Tertiary granitic rocks. Lava eruption occurred as a dynamic response to the intrusion of a new magma into the old reservoir. This process caused the instability of the zoned magma column inducing syneruptive mixing between portions of two contiguous magma layers (both within the column itself and at lower levels where the new basalt was intruded into the reservoir). Syneruptive mixing (mingling) within the middle–upper part of the chamber involved fractions of earlier gabbroic cumulitic materials (lavas erupted around 1970). On the contrary, within the lower part of the chamber, mixing between the intruded HAB and the residing andesitic melt was followed by simple fractional crystallization (FC) of the hybrid magma layer (lavas erupted in 1978–1980). By that time the original magma chamber was completely evacuated. Lavas erupted in 1982/1984 were thus modelled by means of `open system' AFCRE (i.e., AFC with continuous recharge of a fractionating magma batch during eruption): in this case assimilation rates were r₁=0.33 and F=0.86. Recharge rates are slightly higher than extrusion rates and may reflect differences in density (between extruded and injected magmas), together with dynamic fluctuations of these parameters during eruption. Ba and LREE (La, Ce) enrichments of these lavas can be related to selective assimilation of Tertiary granitic and quartz–dioritic rocks. Calculated contents for Zr, Y and other REE are in acceptable agreement with the observed values. It is concluded that simple AFC occurs between two distinct eruption cycles and is typical of a period of repose or mild and decreasing volcanic activity. On the contrary, magma mixing, eventually followed by fractional crystallization (FC) of the hybrid magma layer, occurs during an ongoing eruption. Open-system AFCRE is only operative when the original magma chamber has been totally replenished by the new basaltic magma, and seems a prelude to the progressive ceasing of a major eruptive cycle.     

Eutectic between wollastonite II and calcite contrasted with thermal barrier in MgO-SiO2-CO2 at 30 kilobars,with applications to kimberlite-carbonatite petrogenesis

In the join CaCO₃-CaSiO₃ at 30 kbars, calcite melts at 1615°C, wollastonite II at 1600°C, and a binary eutectic occurs at 1365°C with liquid composition 43 wt.% CaCO₃, 57 wt.% CaSiO₃. The eutectic liquid quenches to a glass with few quench crystals. In the join MgCO₃-MgSiO₃ at 30 kbars, magnesite melts at 1590°C, enstatite at 1837°C, and the fields for the primary crystallization of magnesite and enstatite are separated by a thermal barrier near 1900°C for the melting of forsterite in the presence of CO₂. Only about 10 wt.% MgSiO₃ dissolves in the carbonate liquid. These data, are considered together with incomplete results for joins CaMgSi₂O₆-CaMg(CO₃)₂, CaMgSi₂O₆-MgCO₃, CaMgSi₂O₆-CaCO₃, and other published data in the system CaO-MgO-SiO₂-CO₂. A thermal barrier separates the silicate and carbonate liquids in MgO-SiO₂-CO₂ but, in the quaternary system, silicate liquids with dissolved CO₂ can follow fractionation paths around the forsterite field to the fields for the primary crystallization of carbonates. This suggests that fractional crystallization of CO₂-bearing ultrabasic magma at 100 km depth can produce residual carbonatite magma.     

Evolution of tholeiitic diabase sheet systems in the eastern United States: examples from the Culpeper Basin, Virginia-Maryland, and the Gettysburg Basin, Pennsylvania

High-TiO₂, quartz-normative (HTQ) tholeiite sheets of Early Jurassic age have intruded mainly Late Triassic sedimentary rocks in several early Mesozoic basins in the eastern United States. Field observations, petrographic study, geochemical analyses and stable isotope data from three HTQ sheet systems in the Culpeper basin of Virginia and Maryland and the Gettysburg basin of Pennsylvania were used to develop a general model of magmatic differentiation and magmatic-hydrothermal interaction for HTQ sheets. The three sheet systems have remarkably similar major-oxide and trace-element compositions. Cumulus and evolved diabase in comagmatic sheets separated by tens of kilometers are related by igneous differentiation. Differentiated diabase in all three sheets have petrographic and geochemical signatures and fluid inclusions indicating hydrothermal alteration beginning near magmatic temperatures and continuing to relatively low temperatures. Sulfur and oxygen isotope data are consistent with a magmatic origin for the hydrothermal fluid.The three sheet systems examined apparently all had a similar style of crystal-liquid fractionation that requires significant lateral migration of residual magmatic liquid. The proposed magmatic model for HTQ sheets suggests that bronzite-laden magma was intruded in an upper crustal magma chamber, with bronzite phenocrysts collecting in the lower part of the magma chamber near the feeder dike. Early crystallization of augite and Ca-poor pyroxene before significant plagioclase crystallization resulted in density-driven migration of lighter residual magmatic liquids along lateral and vertical pressure gradients towards the upper part of the sheet. The influence of water on the physical properties of the residual liquid, including density, viscosity and liquidus temperature, may have facilitated the lateral movement more than 15 km up dip in the sheets. Exsolution of a Cl- and S-rich metal-bearing aqueous fluid from residual magma resulted in concentration and redistribution of incompatible and aqueoussoluble elements in late-stage differentiated rocks. This proposed hydrothermal mechanism has important economic implications as it exerts a strong control on the final distribution of noble metals in these types of diabase sheets.     

Phenocryst crystallization during ascent of alkali basalt magma at Rishiri Volcano, northern Japan

Phenocrysts in volcanic rocks are commonly used to deduce crystallization processes in magma chambers. A fundamental assumption is that the phenocrysts crystallized in the magma chambers at isobaric and nearly equilibrium conditions, on the basis of their large sizes. However, this assumption is not always true as demonstrated here for a porphyritic alkali basalt (Kutsugata lava) from Rishiri Volcano, northern Japan. All phenocryst phases in the Kutsugata lava, plagioclase, olivine, and augite, have macroscopically homogeneous distribution of textures showing features characteristic of rapid growth throughout the crystals. Rarely, a core region with distinct composition is present in all phenocryst phases. Phenocrysts, excluding this core, are occasionally in direct contact with each other, forming crystal aggregates. The equilibrium liquidus temperature of plagioclase, the dominant phase (35 vol%) in the Kutsugata lava, can never exceed the estimated magmatic temperature, unless the liquidus temperature increases significantly due to vesiculation of the magma during ascent. This suggests that most phenocrysts in the Kutsugata lava were formed by decompression of the magma during ascent in a conduit, rather than by cooling during residence in a magma reservoir. In the magma chamber before eruption, probably located at depth of more than 7 km, only cores of the phenocrysts were present and the magma was nearly aphyric (<5 vol% crystals), though the observed rock is highly porphyritic with up to 40 vol% crystals. The Kutsugata magma is inferred to have been rich in dissolved H₂O (>4 wt.%) in the magma chamber, and liquidus temperatures of phenocryst phases were significantly suppressed. Large undercooling caused by decompression and degassing of the magma was the driving force for significant crystallization during ascent because of the increase in liquidus temperature due to vapor exsolution. Low ascent rate of the Kutsugata magma, which is suggested by pahoehoe lava morphology and no association of pyroclastics, gave sufficient time for crystallization. Furthermore, the large degree of superheating of plagioclase in the magma chamber caused plagioclase crystallization with low population density and large crystal size, which characterizes the porphyritic nature of the Kutsugata lava. Alkali basalt is likely to satisfy these conditions and similar phenomena are suggested to occur in other volcanic systems.     

Cumulate nodules as evidence for convective fractionation in a phonolite magma chamber

The physical mechanism by which chemical zonation develops in magma chambers has been controversial partly because unambiguous geological constraints have been lacking. The 11,000 years B.P. eruption of Laacher See Volcano produced a zoned tephra deposit and also ejected crystal-rich nodules which provide a snapshot of the materials crystallising at the magma chamber margins. New data on petrography and chemical compositions of nodules, their cumulate minerals and interstitial glasses are used to deduce the chemical evolution of the phonolite melt due to fractional crystallisation of the mineral assemblages. These data, together with those on the vertical zonation of the melt in the bulk of the chamber, are shown to be consistent with a model of stratification of the chamber by convective fractionation, in which a thin boundary layer of residual melt from fractional crystallisation ascends at the chamber side and accumulates at the roof. Crystallisation could have provided buoyancy to drive convection by enriching incompatible volatile components (mainly water) in the residual melt. Available fluid dynamic studies of single- and double-diffusive boundary layers are used to assess convection in the Laacher See chamber. The boundary layer is likely to have been: (1) laminar, which implies that the density gradient in the chamber steepened upwards; (2) in the counterflow regime, in which compositional and thermal layers flow in opposite directions; and (3) thin ( 10 cm), estimated from theory for a flat wall, suggesting that wall morphology could be important in determining boundary layer characteristics. Estimates of mass transfer rates due to this mechanism suggest that the chamber could have become stratified in a time of the order of 10³ years.     

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.     

Changes in magma composition at Arenal volcano,Costa Rica, 1968–1985: Real-time monitoring of open-system differentiation

Arenal volcano in Costa Rica has been erupting nearly continuously, but at a diminishing rate, since 1968, producing approximately 0.35 km³ of lavas and tephras that have shown consistent variations in chemistry and mineralogy. From the beginning of the eruption in July 1968 to early 1970 (stage 1, vol.=0.12 km³) tephras and lavas became richer in Ca, Mg, Ni, Cr, Fe, Ti, V, and Sc and poorer in Al₂O₃ and SiO₂. Concentrations of incompatible trace elements (including Sr) decreased by 5%–20%. Phenocryst contents increased 20–50 vol%. During stage 2 (1970–1973, vol. = 0.13 km³) concentrations of compatible trace elements rose, and concentrations of incompatible trace elements either remained constant or also rose. Al₂O₃ contents decreased by 1 wt%. Phenocryst content increased slightly, principally due to increased orthopyroxene. During stage 3 (mid-1974 to the present, vol.= 0.10 km³) concentrations of SiO₂ increased by 1 wt%, compatible trace elements decreased slightly, and incompatible trace element concentrations increased by 5% to 10%. Although crystals increased in size during stage 3, their overall abundance stayed roughly constant.Our modeling suggests that early stage-1 magmas were produced by boundary layer fractionation under high-p H₂O conditions of an unseen basaltic andesitic magma that intruded into the Arenal system after approximately 500 B.P. Changes in composition during stage 2 resulted from mixing of this more mafic original magma with new magma that had a similar SiO₂ content, but higher compatible and incompatible element concentrations. The changes during stage 3 resulted from continued influx of the same magma plus crystal removal.We conclude that the eruption proceeded in the following way. Before 1968 zoned stage-1 magma resided in the deep crust below Arenal. A new magma intruded into this chamber in July 1968 causing ejection of the stage-1 magmas. The intruding magma mixed with mafic portions of the original chamber producing the mixed lavas of stage 2. Continued mixing plus crystal fractionation along the chamber and conduit walls produced stage-3 lavas. The time scales of crustal level magmatic processes at Arenal range 10⁰–10³ years, which are 3–6 orders of magnitude shorter than those of larger, more silicic systems.     

The magmatic feeding system of El Reventador volcano (Sub-Andean zone,Ecuador) constrained by texture,mineralogy and thermobarometry of the 2002 erupted products

After a 26 years long quiescence El Reventador, an active volcano of the rear-arc zone of Ecuador, entered a new eruptive cycle which lasted from 3 November to mid December 2002. The initial sub-Plinian activity (VEI 4 with andesite pyroclastic falls and flows) shifted on 6 and 21 November to an effusive stage characterized by the emission of two lava flows (andesite to low-silica andesite Lava-1 and basaltic andesite Lava-2) containing abundant gabbro cumulates. The erupted products are medium to high-K calc-alkaline and were investigated with respect to major element oxides, mineral chemistry, texture and thermobarometry. Inferred pre-eruptive magmatic processes are dominated by the intrusion of a high-T mafic magma (possibly up to 1165 ± 15 °C) into an andesite reservoir, acting as magma mixing and trigger for the eruption. Before this refilling, the andesite magma chamber was characterized by water content of 5.3 ± 1.0%, high oxygen fugacity (> NNO + 2) and temperatures, in the upper and lower part of the reservoir, of 850 and 952 ± 65 °C respectively. Accurate amphibole-based barometry constrains the magma chamber depth between 8.2 and 11.3 km (± 2.2 km). The 6 October 2002 seismic swarm (hypocenters from 10 to 11 km) preceding El Reventador eruption, supports the intrusion of magmas at these depths. The widespread occurrence of disequilibrium features in most of the andesites (e.g. complex mineral zoning and phase overgrowths) indicates that convective self-mixing have been operating together with fractional crystallization (inferred from the cognate gabbro cumulates) before the injection of the basic magma which then gave rise to basaltic andesite and low-silica andesite hybrid layers. Magma mixing in the shallow chamber is inferred from the anomalous SiO₂–Al₂O₃ whole-rock pattern and strong olivine disequilibria. Both lavas show three types of amphibole breakdown rims mainly due to heating (mixing processes) and/or relatively slow syn-eruptive ascent rate (decompression) of the magmas. The lack of any disequilibrium textures in the pumices of the 3 November fall deposit suggest that pre-eruptive mixing did not occur in the roof zone of the chamber. A model of the subvolcanic feeding system of El Reventador, consistent with the intrusion of a low-Al₂O₃ crystal-rich basic magma into an already self-mixed andesite shallow reservoir, is here proposed. It is also inferred that before entering the shallow chamber the “basaltic” magma underwent a polybaric crystallization at deeper crustal levels.     

The dynamics of magma withdrawal from a density stratified dyke

Eruptions from the top of a dyke containing two layers of magma can selectively withdraw the upper layer, leaving the dense lower layer undisturbed. Alternatively, if the upper layer is thinner than some critical depth, d, then both layers will be tapped simultaneously. Laboratory experiments yield an equation giving the draw-up depth, d, as a function of dyke geometry, eruption rate, and magma properties. This equation is valid for low to moderate Reynolds numbers and applies to dykes which are much longer than the draw-up depth. Short dykes will yield larger draw-up depths than are predicted by the equation. A large draw-up depth is favoured when the eruption rate, upper layer magma viscosity, or dyke length/breadth ratio is large or the density difference is small. Calculations show that rhyolite-capped dykes can contain several hundred metres thickness of rhyolite when a lower layer is first tapped. Draw-up depths in a dyke are as much as an order of magnitude greater than those for an identical eruption from a large cylindrical chamber tapped by a central vent. Nonetheless, for low effusion rate eruptions from small dykes, as at Inyo Domes, California, relatively small draw-up heights are calculated (e.g. 70 m). This is compatible with the small amounts of mixed magmas found at the transition between the two rhyolite magmas erupted there [11].     

Magma mixing: petrological process and volcanological tool

Magma mixing is a widespread, if not universal igneous phenomenon of variable importance. The evidence for magma mixing is found primarily in glassy tephra; the consolidation of lava obscures the evidence. Inclusions of glass in big crystals in tephra, because of their greater range in composition compared to the whole rock and the residual glass, indicate that the big crystals were derived from separate systems which mixed together prior to and during eruption. The observed or reconstructed concentration of K₂O in inclusions of glass in large crystals represent the composition of the contaminant and host systems. Selective enrichment in K₂O during entrapment of melt by growing crystals is shown to be negligible. The weight percents of K₂O in host, contaminant and residual glass and bulk rock determine the proportions of contaminant and host required to yield either the residual glass or bulk rock. In several cases the proportion of contaminant required is substantially larger than the proportion of crystals in the hybrid magma; therefore, by heat budget argument, the contaminant was partly liquid when contamination began. In some tephra individual phenocrysts contain glasses which are more silicic toward the center of the crystal indicating that the crystal grew from a melt whose composition changed in the opposite sense to that expected for progressive solidification of a closed system. Space time associations of compositionally distinct glassy tephra with contaminated magmas suggest coexistence of basaltic and silicic melts within magma systems. Evidence of contamination is present in most tephra studied so far. Magma mixing appears to be the prevalent process whereby contamination occurs. Magma mixing seems to be particularly evident in systems where there is independent evidence for a vapor-saturated magma reservoir. Probably vapor saturation promotes mixing in magma systems. Magma mixing probably is an important mechanism of compositional diversification (differentiation) of volcanic rocks from continental margin and possibly other environments.Textural evidence of the onset of magma mixing can be related to disturbance of a complex reservoir immediately before ascent and eruption. Thus, conditions before mixing can be ascribed to the reservoir. In this way it is possible to learn about the reservoir: its composition, its diversity, its depth, its walls. It is also possible to learn about the causes of eruption: whether by increase in gas pressure due to either progressive consolidation, or heating from below by an injection of hot magma, or by encounter with ground water; whether by buoyant rise. Evaluation of these problems requires also a thorough knowledge of the chronology of particular eruptions. Thus, magma mixing is a useful volcanological tool.     

Major element variation of the Skaergaard pyroxene and its petrologic implications

Pyroxenes from the Layered Series (LS), Upper Border Series (UBS), and Marginal Border Series (MBS) of the Skaergaard intrusion were analyzed using electron microprobe and mineral separation techniques to examine geochemical variations. In general, pyroxenes from all three series show similar trends in major elements vs % crystallization: SiO₂, MgO, Al₂O₃, and TiO₂ progressively decrease, FeO and MnO progressively increase, and CaO, Fe₂O₃, and P₂O₅ do not change systematically with differentiation. Pyroxenes in the LS and MBS follow a trend similar to that reported by Wager and Brown. The estimated crystallization temperatures closely follow the general trends of published temperature estimates. Major element variation in Skaergaard pyroxenes shows smooth variations with increasing differentiation, indicating that there was no volumetrically significant injection of new magma into the chamber after the initial emplacement. These results strongly support the idea that the Skaergaard intrusion represents in situ crystallization under a closed system magma chamber.     

Degassing of stratified magma by compositional convection

A new model is proposed for passive degassing from sub-volcanic magma chambers. The water content in stably stratified shallow magma chamber will be equated to its solubility at the upper boundary by convection. Water from a lower layer high in water content can enrich the contact zone of the upper layer and lead to further convective overturn of this boundary layer. A complete set of equations describing convection with bubble formation and dissolution is reduced to a simplified form by assuming a small bubble content. The development and pattern of flow driven by vesiculation is modeled numerically in a 2D magma chamber for relatively low Raleigh numbers (5×10⁵). Bubbles rising from the magma will collect near the roof in a layer of 8–10 vol% and then escape upward to fumaroles. The Stokes flux of bubbles escaping from an andesitic magma with viscosity 10⁴ P and a top surface of about 500×500 m corresponds with observed total magmatic water fluxes of 35 kg/s. Pressure within the chamber is buffered by elastic (and local visco-elastic) deformations in the solid rocks bounding the chamber to the range between ambient and close to lithostatic values. In a chamber closed to fresh magma inputs, the decrease in volume due to such gentle volatile escape lowers the reference pressure. Bubbles flux from the lower layer induced by variation of the saturation level around stratification boundary may be efficient mechanism for the water transport between layers.