Description: The Diamante Ignimbrite, which covers an area of 23.000km2 in Chile and Argentina, was emplaced 0.45 Ma ago by a catastophic eruption that caused subsidence of the Diamante Caldera. The ignimbrite is underlain by two fallout deposits, one immediately beneath the ignimbrite, the other separated by sediments. All three deposits have rhyolitic compositions but the older fallout contains higher concentrations of incompatible elements at slightly lower silica contents, reflecting a somewhat different crystal fractionation history. Five juvenile components can be identified in the ignimbrite and its underlying fallout: (1) The dominant white to pink rhyolitic pumice that contains plagioclase, biotite, quartz and ore phenocrysts; alkalifeldspar crystals occur in the matrix. (2) Light-gray rhyolitic pumice, also with plagioclase, biotite, quartz and ore phenocrysts, contains significant amounts of broken crystals and lithic fragments, and mostly occurs mixed with the white component in banded pumice clasts. (3) A highly vesicular, glassy rhyolitic pumice with very few phenocrysts. (4) Foliated, silky shinning white to grey pumice with strongly elongated vesicles and parallel-aligned phenocrysts. (5) A minor component of brown pumice distinct in both its dacitic bulk composition and the presence of amphibole phenocrysts next to plagioclase, biotite and ore. Dark-brown regions in these pumice clasts contain higher amounts of amphibole and plagioclase microlites. There is no systematic vertical compositional zonation through the fallout and ignimbrite but pumice type (5) only occurs in the ignimbrite. Pumice types (1) to (4) share the same mineral and glass compositions. We interpret type (2) as magma that ascended along the conduit walls where high shear stresses caused crystal fragmentation and entrainment of lithic fragments. The dominant type (1) pumice, in contrast, reflects the main, less sheared magma ascending near conduit center. Foliated type (4) pumice may have been magma from in-between these conduit regions. The higher vesicularity of type (3) relative to type (1) indicates that vesiculation in the magma was spatially heterogeneous. The ignimbrite also contains abundant plutonic lithic fragments mostly of andesitic to dacitic bulk compositions. While some of these lie on a fractionation trend leading to the rhyolitic pumice compositions, others have distinct chemical compositions. A fraction of the lithics contains amphibole, and these plutonic rocks have the same dacitic composition as the amphibolebearing pumice (5), which is distinct from the other pumice and lies off the main fractionation trend. We conclude that magma (5) represents a molten crustal component that was added to the rhyolitic magma reservoir.

Description: We have applied a combination of fluid inclusion and amphibole thermobarometry to felsic tephras from highly explosive volcanic eruptions along the Central American volcanic arc (CAVA) from Guatemala through Nicaragua in order to constrain pre-eruptive magma ascent and storage conditions. We note that this is the first time a combination of pressure estimates from fluid inclusions and amphibole chemistry have been used to quantify multi-stage magma chamber processes and magma ascent velocities of large eruptions. Our data document a stepwise ascent of magmas through the crust, typically involving at least two levels of stagnation. Amphibole and fluid inclusion thermobarometry both indicate a shallow preeruptive magma storage level at 80 to 200 MPa (3-8 km depth) along the entire arc. The deeper levels of magma storage vary along-arc, with a tendency to greater maximum depths of up to 25 km in Guatemala and El Salvador, compared to maximum depths of 15 km in Nicaragua. We assume that the continental crust of about 45 km thickness in Guatemala, compared to the 30km thickness of the largely oceanic crust of Nicaragua, allowed for deeper positions of the magma chambers. Thus the observed along-arc changes in mid-crustal magma storage depths indicate a dependence between magma chamber formation and the composition and probably density of the local crust. The average composition of the pre-eruptive fluid phase for highly explosive eruptions in Central America amounts to 90% water, 5% CO2 and 5% NaCl equivalents, and show no systematic alongarc variations. The pressures obtained from the earliest fluid inclusions were taken as the pressures of fluid oversaturation and thus for the beginning of degassing. They range between 150 and 400 MPa, and do not show systematic along-arc variations. Such fluid oversaturation pressures correspond to water contents between 4-8 wt% in the felsic melts. Our results show that the depths of fluid saturation are mostly independent of crustal properties. Degassing typically started at pressures 150 to 300 MPa higher that those corresponding to the last stagnation level, providing evidence for the pre-eruptive criticality of the systems.

Description: The Central American Volcanic Arc (CAVA) is, and has been, one of the most active volcanic regions and generated numerous Plinian eruptions along his 1200 km extension. The best preserved archive of this volcanism can be found as ash layers in the marine sediments downwind from the volcanic sources on the Pacific floor. Numerous ash layers up to 8 Mio old, which occur in ODP and DSDP cores of Legs 66, 67, and 202, originated in Central America and southern Mexico. The cores lie across the ash distribution areas expected from dominant wind directions as identified by mapped fallout deposits. We have chosen 145 ash layers of all three Legs for first detailed analysis of these sites to built up a data base for upcoming IODP cruise 334: Costa Rica Seismogenesis Project. The ash layers commonly have sharp contacts at the bottom and diffuse transitions to terrigenous and pelagic sediments at the top. Ash layer thickness ranges from 0.5 to 60 cm with typical grain sizes from medium silt to coarse sand. The mineral assemblages are typical for arc volcanism (plagioclase, pyroxene, hornblende, and olivine). The most evolved tephras also contain biotite. Electron microprobe analyses of 1300 glass shards yield compositions ranging from basaltic andesite to rhyolite and trachyte. Felsic ashes can be divided into seven compositional groups by means of silica and potassium contents. Correlations between marine ashes and on-land tephras are constrained by petrographical and stratigraphical criteria, major element geochemistry of glasses and minerals, and trace element data from LA-ICP-MS analyses. Due to limited exposure on land, such correlations with individual tephras are only possible for deposits of late Pleistocene to Holocene age. Older ash layers, however, can be correlated with regional arc segments making use of systematic along-arc variations of trace-element characteristics (Zr/Nb, Ba/La, Ce/Yb, La/Yb and Ba/Zr) of the arc rocks. Results show that source areas of the ash layers are distributed along the entire CAVA, as well as at the Southern Mexican Arc. The marine tephra record provides important data for ongoing studies of CAVA volcanism: (a) dating of undated land tephra by correlation with marine ashes and the ages derived by sedimentation rates; (b) stratigraphic correlations along the entire arc can be traced much more completely in the marine sediment cores than by limited onshore outcrops alone; (c) long-term changes in magmatic evolution of volcanic complexes can be reconstructed by using the marine archive of ash layers.

Description: Volcanic eruptions on the deep sea floor have traditionally been assumed to be non-explosive as the high-pressure environment should greatly inhibit steam-driven explosions. Nevertheless, occasional evidence both from (generally slow-) spreading axes and intraplate seamounts has hinted at explosive activity at large water depths. Here we present evidence from a submarine field of volcanic cones and pit craters called Charles Darwin Volcanic Field located at about 3600 m depth on the lower southwestern slope of the Cape Verdean Island of Santo Antão. We examined two of these submarine volcanic edifices (Tambor and Kolá), each featuring a pit crater of 1 km diameter, using photogrammetric reconstructions derived from ROV-based imaging followed by 3D quantification using a novel remote sensing workflow, aided by sampling. The measured and calculated parameters of physical volcanology derived from the 3D model allow us, for the first time, to make quantitative statements about volcanic processes on the deep seafloor similar to those generated from land-based field observations. Tambor cone, which is 2500 m wide and 250 m high, consists of dense, probably monogenetic medium to coarse-grained volcaniclastic and pyroclastic rocks that are highly fragmented, probably as a result of thermal and viscous granulation upon contact with seawater during several consecutive cycles of activity. Tangential joints in the outcrops indicate subsidence of the crater floor after primary emplacement. Kolá crater, which is 1000 m wide and 160 m deep, appears to have been excavated in the surrounding seafloor and shows stepwise sagging features interpreted as ring fractures on the inner flanks. Lithologically, it is made up of a complicated succession of highly fragmented deposits, including spheroidal juvenile lapilli, likely formed by spray granulation. It resembles a maar-type deposit found on land. The eruption apparently entrained blocks of MORB-type gabbroic country rocks with diameters of up to 20 cm, probably abraded by fluidization within the vent, that were laterally transported for hundreds of meters through water. In spite of the great depth, both edifices feature dense but highly fragmented volcanic deposits with an unexpected combination of large clast sizes and wide clast dispersal. This suggests an energetic eruptive environment, which may have similarities with that seen in pyroclastic eruptions on land.

Description: Drill cores recovered during several ODP and IODP Expeditions offshore Central America contain an extensive Early Cenozoic ash layer record. These ash layers have been deposited by plinian eruptions that originated either at the Central American Volcanic Arc (CAVA) or at the Galápagos Hot Spot. While plinian eruptions are well known from the CAVA, volcanism from the Galápagos region is dominantly recorded in effusive and strombolian deposits from subaerial and submarine eruptions although rare large explosive eruptions of evolved trachytic or dacitic compositions did occur in the Pleistocene (e.g., Geist et al., 1994).We have established a tephrostratigraphy from recent through Miocene times from the unique archive of ODP/IODP sites offhore Central America in which we identify tephra source regions by geochemical compositions of the glass shards. Thus we found numerous CAVA-derived tephra layers characterized by typical arc signatures (e.g., Nb-Ta troughs, LILE enrichments), but more surprisingly also an extensive record of tephra layers mostly of Miocene age featuring ocean island geochemical compositions (e.g., low La/Nb and Ba/La ratios, high Nb/Rb ratios). At this geographical setting the only plausible source for these layers is the Galápagos archipelago. Such Miocene ash layers occur in the cores of ODP Sites 1039, 1241, and 1242. At IODP Site U1381, on the Cocos Ridge offshore Costa Rica, 67 primary Miocene (~8 Ma to ~16.5 Ma) fallout ash layers have been recovered. Inferred transport distances of at least 50to 450 km from their vents imply Plinian eruptions, although two-thirds of the ash beds formed from basaltic magmas and only one-third from rhyolitic magmas that are typically associated with plinian eruptions. Our age model for Site U1381 based on sediment accumulation rates, 40Ar/39Ar dating and biostratigraphic ages, reveals a distinct increase in eruption frequency at around 14 Ma. We interpret this as an increase in magma production rates due to changes in interactions between Galápagos plume and spreading ridge.

Description: Llaima is one of the most active volcanoes of the Chilean volcanic front with recent explosive eruptions in 2008 and 2009. Understanding how the volcano evolved to its present state is essential for predictions of its future behavior. The post-glacial succession of explosive volcanic eruptions of Llaima stratovolcano started with two caldera-forming eruptions at ∼16 and ∼15 ka, that emplaced two large-volume basaltic-andesitic ignimbrites (unit I). These are overlain by a series of fall deposits (unit II) changing from basaltic-andesitic to dacitic compositions with time. The prominent compositionally zoned, dacitic to andesitic Llaima pumice (unit III) was formed by a large Plinian eruption at ∼10 ka that produced andesitic surge deposits (unit IV) in its terminal phase. The following unit V represents a time interval of ∼8,000 years during which at least 30 basaltic to andesitic ash and lapilli fall deposits with intercalated volcaniclastic sediments and paleosols were emplaced. Bulk rock, mineral, and glass chemical data constrain stratigraphic changes in magma compositions and pre-eruptive conditions that we interpret in terms of four distinct evolutionary phases. Phase 1 (=unit I) magmas have lower large ion lithophile (LIL)/high field strength (HFS) element ratios compared to younger magmas and thus originated from a mantle source less affected by slab-derived fluids. They differentiated in a reservoir at mid-crustal level. During the post-caldera phase 2 (=units II–IV), relatively long residence times between eruptions allowed for increasingly differentiated magmas to form in a reservoir in the middle crust. Fractional crystallization led to volatile enrichment and oversaturation and is the driving force for the large Plinian eruption of the most evolved (unit III) dacite at Llaima, although replenishment by hot andesite probably triggered the eruption. During the subsequent phase 3 (=unit V 〉3 ka), frequent mafic replenishments at mid-crustal storage levels favored shorter residence times limiting erupted magma compositions to water-undersaturated basaltic andesites and andesites. At around 3 ka, the magma storage level for phase 4 (=unit V 〈3 ka to present) shifted to the uppermost crust where the hot magmas partly assimilated the granitic country rock. Although water contents of these basaltic andesites were low, the low-pressure storage facilitated water saturation before eruption. The change in magma storage level at 3 ka was responsible for the dramatic increase in eruption frequency compared to the older Llaima history. We suggest that the change from middle to upper crust magma storage is caused by a change in the stress regime below Llaima from transpression to tension.