The evolution of andesite volcano structures: new evidence from gravity studies in Costa Rica

Published gravity data on active volcanoes generally reflecteither the low density scoriaceous/pumiceous deposits that are localized within ring-fracture collapse depressions, such as the calderas of mature silicic volcanoes,or the high density frozen magma conduits that occur beneath basaltic shields and cones. The intensive gravity surveys reported here over three complex andesite volcanoes reveal features of both types. Their multi-component gravity fields have crater-centred positive anomalies (1–2 km diameter) surrounded by broader zones of negative gravity with similar amplitudes but greater width (5–10 km). The former are thought to reflect sub-crater magma pipes ofnormal density (ca. 2.5–2.6 Mg m⁻³) surrounded by pyroclastic scoria, ashes and occasional lava flows of muchlower net density (1.8–2.4 Mg m⁻³) which, in turn, account for the negative anomalous zones because the deeper, more consolidated and older parts of these andesite volcano edifices have more normal densities (2.3–2.6 Mg m⁻³).The low density materials are particularly interesting because they appear to have filled topographic depressions to depths of several hundred metres, especially where old caldera-like structures have been postulated from the steep gravity gradients over perimeter ring faults. A model is developed whereby short periods of caldera collapse, associated with intermittent, large high level magma bodies, are interspersed by normal crater-like activity with narrow sub-surface magma pipes. Dominantly pyroclastic activity from summit craters generates the materials that gradually fill earlier-formed topographic depressions. This study demonstrates the unique value of detailed gravity surveys, combined with surface geological information, for modelling and understanding the evolution of active volcano summit regions.     

The long-term growth of volcanic edifices: numerical modelling of the role of dyke intrusion and lava-flow emplacement

The contribution of intrusive complexes to volcano growth is attested by field observations and by the monitoring of active volcanoes. We used numerical simulations to quantitatively estimate the relative contributions to volcano growth of elastic dislocations related to dyke intrusions and of the accumulation of lava flows. The ground uplift induced by dyke intrusions was calculated with the equations of Okada (Bull. Seismol. Soc. Am., 75 (1985) 1135). The spreading of lava flows was simulated as the flow of a Bingham fluid.With realistic parameters for dyke statistics and lava-flow rheology we find the contribution of dyke intrusions to the growth of a basaltic shield archetype to be about 13% in terms of volume and 30% in terms of height. The result is strongly dependent on the proportion of dykes reaching the surface to feed a lava flow. Systematic testing of the model indicates that edifices tend to be high and steep if dykes are thick and high, issued from a small and shallow magma chamber, and if they feed lava flows of high yield strength.The simulation was applied to Ko'olau (O'ahu Is., Hawai'i) and Piton de la Fournaise (Réunion Is.) volcanoes. The simulation of Ko'olau with dyke parameters as described by Walker (Geology, 14 (1986) 310; U.S. Geol. Surv. Prof. Pap., 1350 (1987) 961) and with lava-flow characteristics collected at Kilauea volcano (Hawai'i Is.) results in an edifice morphology very close to that of the real volcano. The best fit model of the Piton de la Fournaise central cone, with its steep slope and E–W elongation, is obtained by the intrusion of 10 000 short and thick dykes issued from a very small and shallow magma chamber and feeding only 700 low-volume lava flows. The same method may be applied to the growth of basaltic shields and other volcano types in different environments, including non-terrestrial volcanism.     

Geology and eruptive history of Unzen volcano, Shimabara Peninsula, Kyushu, SW Japan

During the past 500 thousand years, Unzen volcano, an active composite volcano in the Southwest Japan Arc, has erupted lavas and pyroclastic materials of andesite to dacite composition and has developed a volcanotectonic graben. The volcano can be divided into the Older and the Younger Unzen volcanoes. The exposed rocks of the Older Unzen volcano are composed of thick lava flows and pyroclastic deposits dated around 200–300 ka. Drill cores recovered from the basal part of the Older Unzen volcano are dated at 400–500 ka. The volcanic rocks of the Older Unzen exceed 120 km³ in volume. The Younger Unzen volcano is composed of lava domes and pyroclastic deposits, mostly younger than 100 ka. This younger volcanic edifice comprises Nodake, Myokendake, Fugendake, and Mayuyama volcanoes. Nodake, Myokendake and Fugendake volcanoes are 100–70 ka, 30–20 ka, and <20 ka, respectively. Mayuyama volcano formed huge lava domes on the eastern flank of the Unzen composite volcano about 4000 years ago. Total eruptive volume of the Younger Unzen volcano is about 8 km³, and the eruptive production rate is one order of magnitude smaller than that of the Older Unzen volcano.     

Seismic constraints on magma chambers at Hekla and Torfajökull volcanoes,Iceland

Hekla and Torfajökull are active volcanoes at a rift–transform junction in south Iceland. Despite their location next to each other they are physically and geologically very different. Hekla is an elongate stratovolcano, built mainly of basaltic andesite. Torfajökull is a prominent rhyolitic centre with a 12-km-diameter caldera and extensive geothermal activity. The scope of this study is to examine the propagation of body waves of local earthquakes across the Hekla–Torfajökull area and look for volumes of anomalous S-wave attenuation, which can be evidence of magma chambers. So far the magma chamber under Hekla has been modelled with various geophysical means, and its depth has been estimated to be 5–9 km. A data set of 118 local earthquakes, providing 663 seismic rays scanning Hekla and Torfajökull, was used in this study. The major part, 650 seismograms, did not show evidence for S-wave attenuation under these volcanoes. Only six seismograms had clear signs of S-wave attenuation and seven seismograms were uncertain cases. The data set samples Hekla well at depths of 8–14 km, and south part of it also at 4–8 km and 14–16 km. Western Torfajökull is sampled well at depths of 4–14 km, eastern and southern Torfajökull at 6–12 km. Conclusions cannot be drawn regarding the existence of magma beyond these depth ranges. Also, magma volumes of smaller dimensions than about 800 m cannot be detected with this method. If a considerable molten volume exists under Hekla, it must be located either above 4 km or below 14 km. The former possibility seems unlikely, because Hekla lacks geothermal activity and persistent seismicity, usually taken as expressions of a shallow magma chamber. An aseismic volume with a diameter of 4 km at the depth of 8 km in the west part of Torfajökull has been inferred in earlier studies and interpreted as evidence for a cooling magma chamber. Our results indicate that this volume cannot be molten to a great extent because S-waves travelling through it are not attenuated. Intense geothermal activity and low-frequency earthquakes are possibly signs of magma in the south part of Torfajökull, but a magma chamber was not detected there in the areas sampled by this study.Editorial responsibility: T. Druitt     

Cladistic analysis applied to the classification of volcanoes

Cladistics is a systematic method of classification that groups entities on the basis of sharing similar characteristics in the most parsimonious manner. Here cladistics is applied to the classification of volcanoes using a dataset of 59 Quaternary volcanoes and 129 volcanic edifices of the Tohoku region, Northeast Japan. Volcano and edifice characteristics recorded in the database include attributes of volcano size, chemical composition, dominant eruptive products, volcano morphology, dominant landforms, volcano age and eruptive history. Without characteristics related to time the volcanic edifices divide into two groups, with characters related to volcano size, dominant composition and edifice morphology being the most diagnostic. Analysis including time based characteristics yields four groups with a good correlation between these groups and the two groups from the analysis without time for 108 out of 129 volcanic edifices. Thus when characters are slightly changed the volcanoes still form similar groupings. Analysis of the volcanoes both with and without time yields three groups based on compositional, eruptive products and morphological characters. Spatial clusters of volcanic centres have been recognised in the Tohoku region by Tamura et al. (Earth Planet Sci Lett 197:105–106, 2002). The groups identified by cladistic analysis are distributed unevenly between the clusters, indicating a tendency for individual clusters to form similar kinds of volcanoes with distinctive but coherent styles of volcanism. Uneven distribution of volcano types between clusters can be explained by variations in dominant magma compositions through time, which are reflected in eruption products and volcanic landforms. Cladistic analysis can be a useful tool for elucidating dynamic igneous processes that could be applied to other regions and globally. Our exploratory study indicates that cladistics has promise as a method for classifying volcanoes and potentially elucidating dynamic and evolutionary volcanic processes. Cladistics may also have utility in hazards assessment where spatial distributions and robust definitions of a volcano are important, as in locating sensitive facilities such as nuclear reactors and repositories.     

Emplacement and arrest of sheets and dykes in central volcanoes

Sheet intrusions are of two main types: local inclined (cone) sheets and regional dykes. In Iceland, the inclined sheets form dense swarms of (mostly) basaltic, 0.5–1 m thick sheets, dipping either at 20–50° or at 75–90° towards the central volcano to which they belong. The regional dykes are (mostly) basaltic, 4–6 m thick, subvertical, subparallel and form swarms, less dense than those of the sheets but tens of kilometres long, in the parts of the volcanic systems that are outside the central volcanoes. In both types of swarms, the intrusion intensity decreases with altitude in the lava pile. Theoretical models generally indicate very high crack-tip stresses for propagating dykes and sheets. Nevertheless, most of these intrusions become arrested at various crustal depths and never reach the surface to supply magma to volcanic eruptions. Two principal mechanisms are proposed to explain arrest of dykes and sheets. One is the generation of stress barriers, that is, layers with local stresses unfavourable for the intrusion propagation. The other is mechanical anisotropy whereby sheet intrusions become arrested at discontinuities. Stress barriers may develop in several ways. First, analytical solutions for a homogeneous and isotropic crust show that the intensity of the tensile stress associated with a pressured magma chamber falls off rapidly with distance from the chamber. Thus, while dyke and sheet injection in the vicinity of a chamber may be favoured, dyke and sheet arrest is encouraged in layers (stress barriers) at a certain distance from the chamber. Second, boundary-element models for magma chambers in a mechanically layered crust indicate abrupt changes in tensile stresses between layers of contrasting Young’s moduli (stiffnesses). Thus, where soft pyroclastic layers alternate with stiff lava flows, as in many volcanoes, sheet and dyke arrest is encouraged. Abrupt changes in stiffness between layers are commonly associated with weak and partly open contacts and other discontinuities. It follows that stress barriers and discontinuities commonly operate together as mechanisms of dyke and sheet arrest in central volcanoes.     

Genesis of high-magnesium andesites and associated basalts from Saga–Futagoyama, northwest Kyushu, southwest Japan

High-magnesium andesites associated with basalts erupted after the opening of the Sea of Japan are present at Saga–Futagoyama in northwest Kyushu, southwest Japan. High Mg/(Mg + Fe) [=0.84] of orthopyroxene phenocrysts and bulk rock Mg–Fe–Ni compositions suggest that these high-magnesium andesites were originally primitive melts insignificantly modified in crustal magma chambers. K_D^Ca–Na [= (Ca/Na)_pl/(Ca/Na)_{bulk rock}] ranges from 1.21 to 0.97 and suggests that the high-magnesium andesite magmas would originally have contained H₂O less than 1.8 wt.%. Nb/La does not show a negative correlation with respect to SiO₂. These lines of evidence indicate that hydrous components derived from the subducting slab would not have played a significant role in the genesis of the high-magnesium andesite magmas. Instead, the normative olivine − quartz − [CaTs + Jd] compositions and a negative correlation between Sr/Nd and SiO₂ indicate that the basalt-high-magnesium andesite association would have been formed by multi-stage partial melting of relatively anhydrous source at pressure ranging from 1.5 to 0.5 GPa.     

Growth and evolution of an emergent tuff cone: Considerations from structural geology, geomorphology and facies analysis of São Roque volcano, São Miguel (Azores)

The syn-eruptive and post-eruptive history of São Roque tuff cone, its geological setting and volcanological features were studied in detail to understand the role played by the different factors that contributed to the morphological evolution of this relatively simple and small volcanic edifice.In addition, attention was also focused on the series of natural changes that affected the tuff cone during the course of the years and that finally led to its structural disassembly. A novel model is proposed to explain this process.The São Roque volcanic centre, located on the island of São Miguel (Azores), consists of two well-consolidated bodies and numerous small islets that formed more than 4700 years ago during the hydromagmatic activity that took place along an intruding dyke, whose NNW–SSE trend is in agreement with the regional tectonic pattern. The eruptive vents probably migrated progressively from SSE to NNW, forming small edifices through the rapid accumulation of sediments during alternating phases of “dry” and “wet” magmatic emissions. Syn-eruptive partial collapses greatly modified the original morphological structure of these edifices, probably allowing sea water to continuously flow into the vents. The complex interaction of these factors controlled the depth of magma fragmentation, producing different types of deposits, in which the ash-lapilli ratio varies considerably. The high-water saturation degree of these deposits caused syn-eruptive and post-eruptive remobilization which resulted in collapses and some small-scale landslides.Post-eruptive, WNW–ESE trending transtensional and extensional tectonic activities operated during the initial dissection of the cone, generating instability. Furthermore, the rapid accumulation of “wet” tephra, and its following consolidation, caused selective collapses that favoured the fragmentation of the deposit and caused the formation of numerous islets separated by radially-arranged channels. Collapses also involved the lava units emplaced in more recent times around the tuff cone, which show that brittle deformation has been significant in the area for a prolonged period.     

The magmatic system of the Klyuchevskaya group of volcanoes inferred from data on its eruptions,earthquakes, deformation,and deep structure

The study of magmatic plumbing systems of volcanoes (roots of volcanoes) is one of the main tasks facing volcanology. One major object of this research is the Klyuchevskaya group of volcanoes (KGV), in Kamchatka, which is the greatest such group that has been found at any island arc and subduction zone. We summarize the comprehensive research that has been conducted there since 1931. Several conspicuous results derived since the 1960s have been reported, emerging from the study of magma sources, eruptions, earthquakes, deformation, and the deep structure for the KGV. Our discussion of these subjects incorporates the data of physical volcanology relating to the mechanism of volcanic activity and data from petrology as to magma generation. The following five parts can be distinguished in the KGV plumbing system and the associated geophysical model: the source of energy and material at the top of the Pacific Benioff zone at a depth of about 160 km, the region of magma ascent in the asthenosphere, the region of magma storage in the crust-mantle layer at depths of 40–25 km, magma chambers and channelways in the crust, and the bases of volcanic edifices. We discuss and explain the properties of and the relationships between these parts and the mechanisms of volcanic activity and of the KGV plumbing system as they exist today. Methods for calculating magma chambers and conduits, the amount of magma in the system, and its other properties are available.     

The Shiveluch Volcanic Massif,Kamchatka: Stages in the Evolution of a Magmatic System: Results of Geochronological and Thermobarogeochemical Studies

It is shown that Shiveluch, which consists of several volcanic edifices that stand in one area and in part overlie each other, is a long-lived volcanic massif with a complex structure. The available data on the morphology of the edifice, age, rock compositions, primary melts, and types of eruptive activity were used to identify structurally-temporal units (STUs) in the Shiveluch volcanic massif. It was found that the generation of different-age STUs was due to the activity of at least four magma chambers with different parameters. The durations of the individual chambers were determined. The activities of these chambers were initiated and came to an end nearly instantaneously because of major collapse episodes in the edifice of the massif due to high-magnitude earthquakes.     

Volcanic history of Mount Sidley,a major alkaline volcano in Marie Byrd Land,Antarctica

Mount Sidley is a complex, polygenetic stratovolcano composed primarily of phonolitic and trachytic lavas and subordinate pyroclastic lithologies at the southern extremity of the Executive Committee Range, a linear chain of volcanoes in central Marie Byrd Land, Antarctica. Detailed field investigation coupled with 14 high precision ⁴⁰Ar/³⁹Ar age determinations reveal a 1.5 million year life span between 5.7 and 4.2 Ma in which three major phonolitic central vent edifices (Byrd, Weiss and Sidley volcanoes) and their calderas were developed (5.7–4.8 Ma). This was followed (4.6–4.5 Ma) by the eruption of trachytic magmas from multiple vent localities further south, and then by small volume benmoreite-mugearite lavas and tephras around 4.4–4.3 Ma at the southern end of Mount Sidley. The final phase of activity was the eruption of basanite cones at approximately 4.2 Ma. The southward migration of volcanic activity was accompanied by distinct changes in magma composition and is best explained by the sequential release of magmas stored within an intricate system of conduits and chambers in the crust by tectonically driven (magma assisted?) fracture propagation. The style of volcanic migration at Mount Sidley is emulated on a larger scale by other volcanoes in the Executive Committee Range, in which progressive southward displacement of volcanic activity corresponds with significant petrological variations between major centers.     

Eruption of kimberlite magmas: physical volcanology, geomorphology and age of the youngest kimberlitic volcanoes known on earth (the Upper Pleistocene/Holocene Igwisi Hills volcanoes, Tanzania)

The Igwisi Hills volcanoes (IHV), Tanzania, are unique and important in preserving extra-crater lavas and pyroclastic edifices. They provide critical insights into the eruptive behaviour of kimberlite magmas that are not available at other known kimberlite volcanoes. Cosmogenic ³He dating of olivine crystals from IHV lavas and palaeomagnetic analyses indicates that they are Upper Pleistocene to Holocene in age. This makes them the youngest known kimberlite bodies on Earth by >30?Ma and may indicate a new phase of kimberlite volcanism on the Tanzania craton. Geological mapping, Global Positioning System surveying and field investigations reveal that each volcano comprises partially eroded pyroclastic edifices, craters and lavas. The volcanoes stand <40?m above the surrounding ground and are comparable in size to small monogenetic basaltic volcanoes. Pyroclastic cones consist of diffusely layered pyroclastic fall deposits comprising scoriaceous, pelletal and dense juvenile pyroclasts. Pyroclasts are similar to those documented in many ancient kimberlite pipes, indicating overlap in magma fragmentation dynamics between the Igwisi eruptions and other kimberlite eruptions. Characteristics of the pyroclastic cone deposits, including an absence of ballistic clasts and dominantly poorly vesicular scoria lapillistones and lapilli tuffs, indicate relatively weak explosive activity. Lava flow features indicate unexpectedly high viscosities (estimated at >10² to 10⁶?Pa?s) for kimberlite, attributed to degassing and in-vent cooling. Each volcano is inferred to be the result of a small-volume, short-lived (days to weeks) monogenetic eruption. The eruptive processes of each Igwisi volcano were broadly similar and developed through three phases: (1) fallout of lithic-bearing pyroclastic rocks during explosive excavation of craters and conduits; (2) fallout of juvenile lapilli from unsteady eruption columns and the construction of pyroclastic edifices around the vent; and (3) effusion of degassed viscous magma as lava flows. These processes are similar to those observed for other small-volume monogenetic eruptions (e.g. of basaltic magma).     

The dyke swarm of Mount Calanna (Etna, Italy): an example of the uppermost portion of a volcanic plumbing system

A new multidisciplinary study, combining geology, petrography, and geochemistry, on the rocks of the isolated hill of Mount Calanna (Mount Etna, Italy) has provided evidence for the existence of a dyke swarm, formed by more than 200 dykes distributed over an area of ~0.7 km², with an intensity of intrusion up to 40%. All bodies are deeply altered, and the geological and mesostructural surveying of 132 dykes revealed that they intruded in E–W direction, with an average dip of 60°. The faults affecting the outcrop have in general an E–W strike and dip of ~55°: these have all normal motion and have been interpreted as coeval with the dykes. This interpretation contrasts with the previous hypothesis that considered Mount Calanna as a thrust resulting from compressive deformation resulting from the gravitational spreading of the volcanic edifice. Mount Calanna is here interpreted as the uppermost portion of a vertically extensive magmatic plexus that fed the eruptive activity of one (or more) eruptive center/s sited in the Valle del Bove area. Measurements of the apparent densities on 23 dykes and host rock samples give an average value of 2,420 kg/m³ for the entire complex, ~15% lower than the density expected for hawaiitic magma, placing an important constraint on the geophysical identification of similar structures. Considering that Mount Etna is not an old eroded edifice but an active and growing volcano, the exposure of this subvolcanic structure can be regarded as exceptional. Its geometry and physical characteristics can be thus regarded as an interesting example of the present-day shallow plumbing system of Mount Etna as well as of other basaltic volcanoes.     

Comment on " Volume of magma accumulation or withdrawal estimated from surface uplift or subsidence, with application to the 1960 collapse of Kīlauea volcano" by P. T. Delaney and D. F. McTigue

 In volcanoes that store a significant quantity of magma within a subsurface summit reservoir, such as Kīlauea, bulk compression of stored magma is an important mode of deformation. Accumulation of magma is also accompanied by crustal deformation, usually manifested at the surface as uplift. These two modes of deformation – bulk compression of resident magma and deformation of the volcanic edifice – act in concert to accommodate the volume of newly added magma. During deflation, the processes reverse and reservoir magma undergoes bulk decompression, the chamber contracts, and the ground surface subsides. Because magma compression plays a role in creating subsurface volume to accommodate magma, magma budget estimates that are derived from surface uplift observations without consideration of magma compression will underestimate actual magma volume changes. Received: 30 September 1998 / Accepted: 27 July 1999     

The Pitcairn hotspot in the South Pacific: distribution and composition of submarine volcanic sequences

Multibeam bathymetry and bottom imaging (Simrad EM12D) studies on an area of about 9500 km² were conducted over the Pitcairn hotspot near 25°10′S, 129° 20′W. In addition, 15 dives with the Nautile submersible enabled us to obtain ground-true observations and to sample volcanic structures on the ancient ocean crust of the Farallon Plate at 3500–4300 m depths. More than 100 submarine volcanoes overprint the ancient crust and are divided according to their size into large (>2000 m in height), intermediate (500–2000 m high) and small (<500 m high) edifices. The interpretation of seafloor backscatter imagery accompanied by submersible observations and sampling enabled us to infer that the total volume of submarine lava erupted during hotspot activity is about 5900 km³ within a radius of about 110 km. The most recent volcanic activities occur on both small and large edifices composed of a great variety of lava flows. These flows vary in composition, following a succession from picritic basalt to alkali basalt, trachybasalt, trachy-andesite and to trachyte. Their large range of SiO₂ (48–62%), Na₂O+K₂O (2–11%), Ba (300–1300 ppm), MgO (1–11%), Nb (19–130 ppm), Ni (4–400 ppm) and rare earth elements suggests that crystal–liquid fractionation from basanite and/or picritic melt sources was a major process. The variation in composition between the least evolved basaltic rocks and the other more evolved silicic lava is marked by a difference in their flow morphology (pillow, giant tubes, tabular to blocky flows). The lava composition and field observation indicate that several magmatic pulses giving rise to cyclic eruptions are responsible for the construction of the edifices. The two larger edifices (>2000 m high) show more extensive eruptive events and a wider range in compositional variability than the smaller (<500 m high) ones. Several (five) submersible transects made along the slope of one of the largest edifices (Bounty) enabled us to observe at least nine successive eruptive cycles progressing from pillow and giant tubular basalt to tabular/blocky trachy-andesite and trachyte flows. Pyroclasts and hyaloclastites are often found with these eruptive sequences. The smaller edifices, forming individualized cones, are built mainly of evolved silicic (SiO₂>53%) flows consisting essentially of alternating sequences of trachy-andesite and trachyte. The distribution and composition of the small edifices suggest that they are the result of sub-crustal forceful magma injection and channeling supplied from reservoirs associated with the large volcanoes.     

Volcano-stratigraphy and geochemistry of collision-related volcanism on the Erzurum–Kars Plateau, northeastern Turkey

The Eastern Anatolia Region exhibits one of the world's best exposed and most complete transects across a volcanic province related to a continental collision zone. Within this region, the Erzurum–Kars Plateau is of special importance since it contains the full record of collision-related volcanism from Middle Miocene to Pliocene. This paper presents a detailed study of the volcanic stratigraphy of the plateau, together with new K–Ar ages and several hundred new major- and trace-element analyses in order to evaluate the magmatic evolution of the plateau and its links to collision-related tectonic processes. The data show that the volcanic units of the Erzurum–Kars Plateau cover a broad compositional range from basalts to rhyolites. Correlations between six logged, volcano-stratigraphic sections suggest that the volcanic activity may be divided into three consecutive Stages, and that activity begins slightly earlier in the west of the plateau than in the east. The Early Stage (mostly from 11 to 6 Ma) is characterised by bimodal volcanism, made up of mafic-intermediate lavas and acid pyroclastic rocks. Their petrography and high-Y fractionation trend suggest that they result from crystallization of anhydrous assemblages at relatively shallow crustal levels. Their stratigraphy and geochemistry suggest that the basic rocks erupted from small transient chambers while the acid rocks erupted from large, zoned magma chambers. The Middle Stage (mostly from 6–5 Ma) is characterised by unimodal volcanism made up predominantly of andesitic–dacitic lavas. Their petrography and low-Y fractionation trend indicate that they resulted from crystallization of hydrous (amphibole-bearing) assemblages in deeper magma chambers. The Late Stage (mostly 5–2.7 Ma) is again characterised by bimodal volcanism, made up mainly of plateau basalts and basaltic andesite lavas and felsic domes. Their petrography and high-Y fractionation trend indicate that they resulted from crystallization of anhydrous assemblages at relatively shallow crustal levels. AFC modelling shows that crustal assimilation was most important in the deeper magma chambers of the Middle Stage. The geochemical data indicate that the parental magma changed little throughout the evolution of the plateau. This parental magma exhibits a distinctive subduction signature represented by selective enrichment in LILE and LREE thought to have been inherited from a lithosphere modified by pre-collision subduction events. The relationships between magmatism and tectonics support models in which delamination of thickened subcontinental lithosphere cause uplift accompanied by melting of this enriched lithosphere. Magma ascent, and possibly magma generation, is then strongly controlled by strike-slip faulting and associated pull-apart extensional tectonics.     

On the geometrical form of volcanoes

Many volcanic edifices have a remarkably symmetric geometrical form. An example is Mount Fuji in Japan. We model this form assuming that the surface of the volcano is a surface of uniform hydraulic potential; that an erupting magma will follow the path of minimum resistance to the surface. In order to model the resistance to fluid flow we assume the volcanic edifice is a uniform porous medium. The vertical flow of magma is also resisted by the gravitational body force. If the volcano becomes too tall flank eruptions will widen it; if the volcano becomes too wide summit eruptions will increase its elevation. Using the Dupuit approximation for an unconfined aquifer it is shown that the percolation equation is applicable. As magma reaches the surface it is assumed to extend the solid, porous matrix. A similarity solution is obtained to this moving boundary problem. The solution predicts a uniform shape for all volcanoes. This shape is shown to be in excellent agreement with the geometrical form of Mount Fuji.     

Evolution of the Cañadas edifice and its implications for the origin of the Cañadas Caldera (Tenerife, Canary Islands)

The volcano-stratigraphic and geochronologic data presented in this work show that the Tenerife central zone has been occupied during the last 3 Ma by shield or central composite volcanoes which reached more than 3000 m in height. The last volcanic system, the presently active Teide-Pico Viejo Complex began to form approximately 150 ka ago. The first Cañadas Edifice (CE) volcanic activity took place between about 3.5 Ma and 2.7 Ma. The CE-I is formed mainly by basalts, trachybasalts and trachytes. The remains of this phase outcrop in the Cañadas Wall (CW) sectors of La Angostura (3.5–3.0 Ma and 3.0–2.7 Ma), Boca de Tauce (3.0 Ma), and in the bottom of some external radial ravines (3.5 Ma). The position of its main emission center was located in the central part of the CC. The volcano could have reached 3000 m in height. This edifice underwent a partial destruction by failure and flank collapse, forming debris-avalanches during the 2.6–2.3 Ma period. The debris-avalanche deposits can be seen in the most distal zones in the N flank of the CE-I (Tigaiga Breccia). A new volcanic phase, whose deposits overlie the remains of CE-I and the former debris-avalanche deposits, constituted a new volcanic edifice, the CE-II. The dyke directions analysis and the morphological reconstruction suggest that the CE-II center was situated somewhat westward of the CE-I, reaching some 3200 m in height. The CE-II formations are well exposed on the CW, especially at the El Cedro (2.3–2.00 Ma) sector. They are also frequent in the S flank of the edifice (2.25–1.89 Ma) in Tejina (2.5–1.87 Ma) as well as in the Tigaiga massif to the N (2.23 Ma). During the last periods of activity of CE-II, important explosive eruptions took place forming ignimbrites, pyroclastic flows, and fall deposits of trachytic composition. Their ages vary between 1.5 and 1.6 Ma (Adeje ignimbrites, to the W). In the CW, the Upper Ucanca phonolitic Unit (1.4 Ma) could be the last main episode of the CE-II. Afterwards, the Cañadas III phase began. It is well represented in the CW sectors of Tigaiga (1.1 Ma–0.27 Ma), Las Pilas (1.03 Ma–0.78 Ma), Diego Hernández (0.54 Ma–0.17 Ma) and Guajara (1.1 Ma–0.7 Ma). The materials of this edifice are also found in the SE flank. These materials are trachybasaltic lava-flows and abundant phonolitic lava and pyroclastic flows (0.6 Ma–0.5 Ma) associated with abundant plinian falls. The CE-III was essentially built between 0.9 and 0.2 Ma, a period when the volcanic activity was also intense in the ‘Dorsal Edifice' situated in the easterly wing of Tenerife. The so called ‘valleys' of La Orotava and Güimar, transversals to the ridge axis, also formed during this period. In the central part of Tenerife, the CE-III completed its evolution with an explosive deposit resting on the top of the CE, for which ages from 0.173 to 0.13 Ma have been obtained. The CC age must be younger due to the fact that the present caldera scarp cuts these deposits. On the controversial origin of the CC (central vertical collapse vs. repeated flank failure and lateral collapse of mature volcanic edifices), the data discussed in this paper favor the second hypothesis. Clearly several debris-avalanche type events exist in the history of the volcano but most of the deposits are now under the sea. The caldera wall should represent the proximal scarps of the large slides whose intermediate scarps are covered by the more recent Teide-Pico Viejo volcanoes.     

Pahoehoe and aa in Hawaii: volumetric flow rate controls the lava structure

The historical records of Kilauea and Mauna Loa volcanoes reveal that the rough-surfaced variety of basalt lava called aa forms when lava flows at a high volumetric rate (>5–10 m³/s), and the smooth-surfaced variety called pahoehoe forms at a low volumetric rate (<5–10 m³/s). This relationship is well illustrated by the 1983–1990 and 1969–1974 eruptions of Kilauea and the recent eruptions of Mauna Loa. It is also illustrated by the eruptions that produced the remarkable paired flows of Mauna Loa, in which aa formed during an initial short period of high discharge rate (associated with high fountaining) and was followed by the eruption of pahoehoe over a sustained period at a low discharge rate (with little or no fountaining). The finest examples of paired lava flows are those of 1859 and 1880–1881. We attribute aa formation to rapid and concentrated flow in open channels. There, rapid heat loss causes an increase in viscosity to a threshold value (that varies depending on the actual flow velocity) at which, when surface crust is torn by differential flow, the underlying lava is unable to move sufficiently fast to heal the tear. We attribute pahoehoe formation to the flowage of lava at a low volumetric rate, commonly in tubes that minimize heat loss. Flow units of pahoehoe are small (usually <1 m thick), move slowly, develop a chilled skin, and become virtually static before the viscosity has risen, to the threshold value. We infer that the high-discharge-rate eruptions that generate aa flows result from the rapid emptying of major or subsidiary magma chambers. Rapid near-surface vesiculation of gas-rich magma leads to eruptions with high discharge rates, high lava fountains, and fast-moving channelized flows. We also infer that long periods of sustained flow at a low discharge rate, which favor pahoehoe, result from the development of a free and unimpeded pathway from the deep plumbing system of the volcano and the separation of gases from the magma before eruption. Achievement of this condition requires one or more episodes of rapid magma excursion through the rift zone to establish a stable magma pathway.     

Volcanic hazards from Bezymianny- and Bandai-type eruptions

Major slope failures are a significant degradational process at volcanoes. Slope failures and associated explosive eruptions have resulted in more than 20 000 fatalities in the past 400 years; the historic record provides evidence for at least six of these events in the past century. Several historic debris avalanches exceed 1 km³ in volume. Holocene avalanches an order of magnitude larger have traveled 50–100 km from the source volcano and affected areas of 500–1500 km². Historic eruptions associated with major slope failures include those with a magmatic component (Bezymianny type) and those solely phreatic (Bandai type). The associated gravitational failures remove major segments of the volcanoes, creating massive horseshoe-shaped depressions commonly of caldera size. The paroxysmal phase of a Bezymianny-type eruption may include powerful lateral explosions and pumiceous pyroclastic flows; it is often followed by construction of lava dome or pyroclastic cone in the new crater. Bandai-type eruptions begin and end with the paroxysmal phase, during which slope failure removes a portion of the edifice. Massive volcanic landslides can also occur without related explosive eruptions, as at the Unzen volcano in 1792.The main potential hazards from these events derive from lateral blasts, the debris avalanche itself, and avalanche-induced tsunamis. Lateral blasts produced by sudden decompression of hydrothermal and/or magmatic systems can devastate areas in excess of 500km² at velocities exceeding 100 m s^–1. The ratio of area covered to distance traveled for the Mount St. Helens and Bezymianny lateral blasts exceeds that of many pyroclastic flows or surges of comparable volume. The potential for large-scale lateral blasts is likely related to the location of magma at the time of slope failure and appears highest when magma has intruded into the upper edifice, as at Mount St. Helens and Bezymianny.Debris avalanches can move faster than 100 ms^–1 and travel tens of kilometers. When not confined by valley walls, avalanches can affect wide areas beyond the volcano's flanks. Tsunamis from debris avalanches at coastal volcanoes have caused more fatalities than have the landslides themselves or associated eruptions. The probable travel distance (L) of avalanches can be estimated by considering the potential vertical drop (H). Data from a catalog of around 200 debris avalanches indicates that the H/L rations for avalanches with volumes of 0.1–1 km³ average 0.13 and range 0.09–0.18; for avalanches exceeding 1 km³, H/L ratios average 0.09 and range 0.5–0.13.Large-scale deformation of the volcanic edefice and intense local seismicity precede many slope failures and can indicate the likely failure direction and orientation of potential lateral blasts. The nature and duration of precursory activity vary widely, and the timing of slope faliure greatly affects the type of associated eruption. Bandai-type eruptions are particularly difficult to anticipate because they typically climax suddenly without precursory eruptions and may be preceded by only short periods of seismicity.