Abstract Subglacial volcanic eruptions can generate large volumes of meltwater that is stored and transported beneath glaciers and released catastrophically in jökulhlaups. At typical basaltic dyke propagation speeds, the high strain rate at a dyke tip causes ice to behave as a brittle solid; dykes can overshoot a rock–ice interface to intrude through 20–30% of the thickness of the overlying ice. The very large surface area of the dyke sides causes rapid melting of ice and subsequent collapse of the dyke to form a basal rubble pile. Magma can also be intruded at the substrate–ice interface as a sill, spreading sideways more efficiently than a subaerial flow, and also producing efficient and widespread heat transfer. Both intrusion mechanisms may lead to the early abundance of meltwater sometimes observed in Icelandic subglacial eruptions. If meltwater is retained above a sill, continuous melting of adjacent and overlying ice by hot convecting meltwater occurs. At typical sill pressures under more than 300 m ice thickness, magmatic CO 2 gas bubbles form c. 25 vol% of the pressurized magma. If water drains and contact with the atmosphere is established, the pressure decreases dramatically unless the overlying ice subsides rapidly into the vacated space. If it does not, further CO 2 exsolution plus the onset of H 2 O exsolution has the potential to cause explosive fragmentation, i.e. a fire-fountain that forms at the dyke-sill connection, enhancing melting and creating another candidate pulse of meltwater. The now effectively subaerial magma body becomes thicker, narrower, and flows faster so that marginal meltwater drainage channels become available. If the ice overburden thickness is much less than c. 300m the entire sill injection process may involve explosive magma fragmentation. Thus, there should be major differences between subglacial eruptions under local or alpine glaciers compared with those under continental-scale glaciers.

Abstract Here we describe a two-day field trip to examine Quaternary volcanism in the Canadian Cascade arc, named the Garibaldi volcanic belt. Day 1 of the trip proceeds along the Whistler corridor from Squamish to Pemberton and focuses on Quaternary glaciovolcanic deposits. Interactions between volcanoes and ice in the Garibaldi volcanic belt have been common during the past two million years and this has resulted in a diverse array of landforms, including subglacial domes, tuyas, impounded lava masses, and sinuous lavas that exploited within-ice drainage systems. On Day 2, the trip heads northwest of Pemberton, British Columbia, along logging roads to see deposits from the 2360 yr B.P. eruption of the Mount Meager volcanic complex. This eruption began Plinian-style, generating pyroclastic fall and flow deposits and ended with the production of block and ash pyroclastic flows by explosive (Vulcanian) collapse of lava domes (e.g., Soufriére Hills). Many of the traits of the deposits seen on this two day trip are a reflection of, both, the style of eruption and the nature of the surrounding landscape. In this regard, the trip provides a spectacular window into the nature and hazards of effusive and explosive volcanism occurring in mountainous terrains and the role of water and ice.

Abstract A short-lived eruption of basaltic andesite to andesite on Deception Island in 1969 occurred from a series of fissures underneath a glacier. The glacier was thin ( c. 100 m) and the eruption created a large and sudden discharge of meltwater that overflowed the glacier, severely damaging buildings on the island. The eruption was unusually well documented and it illustrates several features of subglacial eruptions that are only poorly known and not well understood. In particular, overflowing meltwater is contrary to predictions based on existing simple hydrological models for eruptions beneath thin glaciers. The eruption is analysed in this paper and used as a model for the fluid dynamics and thermodynamics of eruptions beneath a thin glacier mainly composed of impermeable ice. It is suggested that, in eruptions of relatively fluid magmas with a low magma rise rate, volatiles and magma are able to decouple and subglacial melting is strongly influenced by the superheated magmatic and hydrothermal gases (mainly steam). Thus, melting is much faster than that due solely to coupled conductive (magma) and convective (meltwater) heat transfer. The influence of gasdriven melting also has an important effect on the shape of the meltwater cavity and may be at least partly responsible for the cylindrical ice chimneys developed above vents on Deception Island. The results of the study are important for reconstructing the shapes of englacial cavities melted above a vent. They also highlight the importance of glacier structure and densification, rather than simply glacier thickness, in determining the hydraulic evolution of an eruption. Even eruptions beneath thin glaciers can generate significant meltwater floods.

The origin and evolution of Mars’s inventory of volatile elements is pivotal to a wide range of physical, chemical, geological, and biological issues and concerns. The identification of subglacially erupted volcanoes on Mars suggests that ice sheets existed at high and low latitudes repeatedly over geological time, but the importance of those volcanoes is not just as a simple Boolean climate signal. Like terrestrial subglacially erupted volcanoes, they can potentially yield a more holistic range of paleoenvironmental parameters, including ice thickness, thermal regime, and surface elevation. On Earth, at least nine different types of terrestrial subglacial volcanic successions can be identified using landform characteristics, lithofacies, and sequence architecture. The principal characteristics of each are reviewed in this paper, together with the first empirical comparative analysis of the morphometry of the landforms. All were probably erupted in association with wet-based ice and there are different implications for volcanic landforms erupted under different glacial thermal regimes (polar, subpolar). However, they represent our best sources of information with which to assess Mars analogs, some of which (as on Earth) may have been the source of megascale meltwater outburst floods. Applying the results of this paper to three different morphological types of candidate subglacial volcanoes on Mars indicates that it is difficult to suggest a plausible glaciovolcanic analogy for Mars’s tall cones ; they more closely resemble pyroclastic mounds erupted subaerially or subaqueously, under ice-free conditions. Conversely, Mars’s low-domes may be very extensive, inflated, subglacial “interface sills” formed under comparatively thick ice of any thermal regime. Finally, the very large, flat-topped constructs on Mars resemble mafic tuyas emplaced in thick (up to 2 km) temperate ice. However, because of their very large size compared to terrestrial analogs, the possibility also exists that the latter are polygenetic stratovolcanoes, formed subglacially either within very thick ice, or as multiple superimposed lava-fed deltas emplaced in much thinner ice that repeatedly re-formed on the volcanoes after each eruptive episode. A plausible terrestrial analogy for the latter is the long-lived James Ross Island stratovolcano in Antarctica.

Abstract Eruptions from partly ice-covered and ice-capped volcanic systems constitute nearly 60% of all known historical (i.e. in the past 11 centuries) eruptions in Iceland. Since the fourteenth century such eruptions have been reported in contemporary or near-contemporary documents. At least 120 historical eruptions have broken through the ice on the glaciated parts of five volcanic systems and have left tephra layers in ice and soil, or been recorded at the time. An unknown number of eruptions did not breach the overlying ice and left no record at all. Beginning as subglacial eruptions, most eruptions break through the ice in minutes, hours or days and can last from a few days to several months. A single vent or the whole length of a fissure may then emerge to emit highly-fragmented tephra in hydromagmatic explosions of varying strength. The volume of airborne tephra varies by at least four orders of magnitude with dispersal range varying from near-vent to transatlantic. In most of the eruptions the magma was of basaltic composition. Eruption frequency is highest within the Grímsvötn system where up to seven eruptions every 40 years have occurred during peaks of activity and at least 70 eruptions over historical time. The observed pattern of temporally and spatially close eruptions, separated by periods of low or no activity, leaves open the question whether pre-Holocene deposits of several closely spaced subglacial eruptions can be securely distinguished from those formed in a single subglacial eruption.

Abstract Hoodoo Mountain volcano (HMV), a Quaternary composite volcano in northwestern British Columbia, is a well-exposed example of peralkaline, phonolitic icecontact and subglacial volcanism. Its distinctive morphology and unique volcanic deposits are indicative of subglacial, within-ice, and/or ice-contact volcanic eruptions. Distinct ice-contact deposits result from three different types of lava–ice interaction: (1) vertical cliffs of lava, featuring finely jointed flow fronts up to 200 m in height, resulted from lava flows being dammed and ponded against thick masses of ice; (2) pervasively-jointed, dense lava flows, lobate intrusions, and domes associated with mantling deposits of poorly-vesiculated breccia are derived from volcanic eruptions contained beneath relatively thick ice; and (3) an association of pervasively-jointed, highly-vesicular lava flows or dykes encased by vesicular hyaloclastite of identical composition formed by eruption under and/or through relatively thin ice. The distribution of these three deposit types largely explains the distinctive morphology of Hoodoo Mountain and can be used to reconstruct variations in ice thickness surrounding the volcano since c .85 ka. Our analysis suggests that at c .85 ka Hoodoo Mountain erupted underneath ice cover of at least several hundred metres. At c .80 ka eruptions were no longer subglacial, but the edifice was surrounded by ice at least 800 m high that dammed lava flows around the perimeters of the volcano. After a period of eruptions showing no apparent evidence for ice interaction, from <80 to >40 ka, subglacial eruptions began again, signalling the build-up of regional ice levels. Local ice thickness during these eruptions may well have been over 2 km thick.

Abstract The Garibaldi Volcanic Belt (GVB) in southwestern British Columbia is dominated by intermediate composition volcanoes in a setting that has been intermittently subjected to widespread glaciation. The glaciovolcanic features produced are distinctive, and include flow-dominated tuyas, subglacial domes, and ice-marginal flows. Flow-dominated tuyas, which are intermediate in composition, are unlike conventional basaltic tuyas; they consist of stacks of flat-lying lava flows, and lack pillows and hyaloclastite. They are inferred to represent subglacial eruptions that ultimately breached the ice surface. subglacial domes occur as steep-sided masses of heavily-jointed, glassy lava, and represent eruptions that were entirely subglacial. Ice-marginal flows derive from subaerial flows that were impounded against ice. Two unique aspects of GVB glaciovolcanic products are the presence of flow-dominated tuyas and the apparent scarcity of primary fragmental deposits. These unique features result from lava composition, the minimization of direct lava-water contact during eruptions, and topography. Composition influences morphology because eruption temperature decreases, and viscosity and glass transition temperature both increase with silica content. The result of this is that silicic subglacial volcanoes melt less water and are less likely to trap it near the vent, leading to the formation of structures whose shapes are strongly influenced by the surrounding ice. Topography also enhances meltwater drainage, favours lava flow impoundment in ice-filled valleys, and may, through erosion, influence the observed distribution of fragmental glaciovolcanic deposits.

Abstract The West Antarctic Ice Sheet (WAIS) flows through the volcanically active, late Cenozoic West Antarctic rift system. Active subglacial volcanism and a vast ( > 10 6 km 3 ) extent of subglacial volcanic structures have been interpreted from aerogeophysical surveys over central West Antarctica in the past decade, combined with results from 1960s and 1970s aeromagnetic profiles over the WAIS. Modelling of magnetic anomalies constrained by radar ice sounding shows volcanic sources at the base of the ice throughout large areas, whose subglacially erupted hyaloclastite edifices have been eroded by moving ice, as in Iceland. The 1800 m-high divide of the WAIS is underlain by the 400 km-long volcanic Sinuous Ridge, which rises above sea level; most hyaloclastite edifices there have also been glacially removed, indicating migration of the ice divide through time. Northeast of the divide of the WAIS there is a 400-nT positive magnetic anomaly over the shallowest, most rugged bedrock topography (elevation +380m above sea level), probably comprising subaerially erupted flows erupted when the Sinuous Ridge area was deglaciated. Uplift of the Sinuous Ridge may have forced the advance of the WAIS. Other aspects of the subglacial volcanism in Antarctica can be observed in Iceland and have a direct bearing on our understanding of the subglacial conditions of the WAIS and its dynamics.

Abstract The long-lived (at least 0.78 Ma) Eyjafjallajökull volcano has been constructed during a period of dramatic climatic shifts, and has produced subaerial lavas and cinder cones, pillow lavas, hyaloclastite, monogenetic volcaniclastic sediments and polygenetic glacio-fluvial sediments. Subaqueous lithofacies are dominant and nine distinct cogenetic lithofacies associations are identified within the successions that are interpreted as the proximal deposits of subglacial eruptions. These associations are typically bound by unconformities and lie directly on thin diamictites or glaciated surfaces; epiclastic sedimentary rocks are absent. The lithofacies include subaqueous sheet lava flows, lobate flows, pillow lavas, breccias, hyaloclastite and hyalotuff generated by eruption of mainly basaltic lavas. Lavas associated with massive hyaloclastite breccias commonly lie on or intrude cogenetic redeposited hyalotuff, indicating rapid changes in style of activity from explosive to effusive. This suggests that the vent was initially subaqueous but became subaerial as meltwater drained away down-slope beneath temperate ice sheets. In other instances, probably under thicker ice, the eruptions were effusive throughout but lava was subjected to intense steam explosivity. There is abundant evidence that volcaniclastic material was transported downslope by mass flow, grain flow and traction currents (i.e. by running water). Lithofacies associations that ponded in ice-dammed water commonly comprise thick redeposited tephra overlain by a ‘passage zone’ of hyaloclastite and then subaerial lavas. These deposits are voluminous but conditions suitable for ponding of large volumes of water were not common in the evolution of this large volcano. During the early stages of volcano growth, eruptive deposits were emplaced on the gentle slopes of the growing cone during both glacial and interglacial periods. Glacial erosion modified the growing edifice and, as a result, younger deposits (<600 ka) tend to be valley-confined. The unconformities produced during glacial advance represent significant time gaps within the succession. Observations support the hypothesis that there were higher volcanic production rates during periods of deglaciation, probably as a result of the rapid decrease in lithostatic pressure as ice melted.

Abstract In the Gjálp eruption in 1996, a subglacial hyaloclastite ridge was formed over a volcanic fissure beneath the Vatnajökull ice cap in Iceland. The initial ice thickness along the 6 km-long fissure varied from 550 m to 750 m greatest in the northern part but least in the central part where a subaerial crater was active during the eruption. The shape of the subglacial ridge has been mapped, using direct observations of the top of the edifice in 1997, radio echo soundings and gravity surveying. The subglacial edifice is remarkably varied in shape and height. The southern part is low and narrow whereas the central part is the highest, rising 450 m above the pre-eruption bedrock. In the northern part the ridge is only 150–200 m high but up to 2 km wide, suggesting that lateral spreading of the erupted material occurred during the latter stages of the eruption. The total volume of erupted material in Gjálp was about 0.8 km 3 , mainly volcanic glass. The edifice has a volume of about 0.7 km 3 and a volume of 0.07 km 3 was transported with the meltwater from Gjálp and accumulated in the Grímsvötn caldera, where the subglacial lake acted as a trap for the sediments. This meltwater-transported material was removed from the southern part of the edifice during the eruption. Variations in basal water pressure may explain differences in edifice form along the fissure. Partial floating of the overlying ice in the northern part is likely to have occurred due to high water pressures, reducing confinement by the ice and allowing lateral spreading of the edifice. The overall shape of the Gjálp ridge is similar to that of many Pleistocene hyaloclastite ridges in Iceland. Future preservation of the Gjálp ridge will depend on the rate of glacial erosion it will suffer. Besides being related to future ice flow velocities, the erosion rate will depend on the rate of consolidation due to palagonitization and shielding from glacial erosion while depressions in the ice are gradually filled by ice flow directed towards the Gjálp hyaloclastite ridge.