Search

Highlights • New broad band magnetotelluric data covering the whole Gran Canaria island. • First 3D electrical resistivity model of Gran Canaria island at crustal scale. • Low resistivity anomalies found may be of interest for geothermal exploitation. • 3D electrical resitivity shows a good correlation with other geophysical data available at crustal scale. Abstract Gran Canaria, one of the two main islands of the Canary Archipelago off NW Africa, has been volcanically active for at least 15 million years. The island went through several volcanic cycles that varied greatly in composition and extrusive and intrusive activity. The complex orography of the island has excluded extensive land geophysical surveys on the island. A review of the available geophysical information on the island shows that it has been obtained mainly through marine and airborne geophysical surveys. A new dataset comprising 100 magnetotelluric soundings acquired on land has been used to obtain the first 3D electrical resistivity model of the island at crustal scale. The model shows high resistivity values close to the surface in the exposed Tejeda Caldera that extends at depth to the SE cutting the islands in half. Outside the inferred limits of the Tejeda Caldera the 3D model shows low resistivity values that could be explained by hydrothermal alteration at deeper levels and the presence of marine saltwater intrusion at shallower levels near the coast. The presence of unconnected vertical-like structures, with very low resistivity (<10 ohm m) could be associated to small convective cells is confirmed by the sensitivity analysis carried out in the present study. Those structures are the most likely candidates for a detailed analysis in order to determine their geothermal economic potential. A comprehensive review of existing geophysical data and models of Gran Canaria island and their comparison with the new 3D electrical resistivity model is presented.

Highlights • Loess in Northern Iran sensitively reacted on Pleistocene climate change. • Dust accumulation and soil formation followed climate cycles of the Northern Hemisphere. • Pedostratigraphic correlation provides link within Southern Eurasian loess belt. • Southern Caspian Lowlands provide key area for climate reconstruction in West Asia. The southern Caspian Lowland sensitively reacted to Pleistocene climate change and is a key area for reconstructing climate dynamics and landscape evolution in Southern Eurasia. Loess-paleosol sequences (LPS) of the northern foothills of Alborz Mountains provide detailed records of climate-induced changes of dust accumulation and soil formation correlating with relatively dry or moist conditions of the past. The LPS at Neka-Abelou (NA) was studied in detail in order to understand these dynamics and provide a base for regional pedostratigraphic correlation. We have carried out high-resolution analyses of grain-size, sediment color, mass specific and frequency-dependent magnetic susceptibility and carbonate content and established a temporal framework as based on luminescence dating using a post-IR infrared stimulated luminescence (pIRIR) protocol and fading corrections. The LPS of NA is composed of finely textured loess, which is subdivided by at least eleven paleosols. Moreover, it contains a thin loess layer with lenses of trachytic tephra which most likely originated from the Damavand volcano and was deposited during Marine Isotope Stage (MIS) 4. The lower part of the LPS at NA consists of four strongly-developed reddish-brown paleosols (Bt, Btg horizons) separated by thin layers of pedogenically-altered loess indicating moist climate conditions and low dust accumulation rates during the Middle Pleistocene or earlier. The central part contains a pedocomplex of clay-rich paleosols composed of well developed Bt, ABk and Bw horizons formed under strongly reduced dust accumulation rates and intercalated by loess layers. Pedostr

Highlights • New broad band magnetotelluric data covering the whole Gran Canaria island. • First 3D electrical resistivity model of Gran Canaria island at crustal scale. • Low resistivity anomalies found may be of interest for geothermal exploitation. • 3D electrical resitivity shows a good correlation with other geophysical data available at crustal scale. Abstract Gran Canaria, one of the two main islands of the Canary Archipelago off NW Africa, has been volcanically active for at least 15 million years. The island went through several volcanic cycles that varied greatly in composition and extrusive and intrusive activity. The complex orography of the island has excluded extensive land geophysical surveys on the island. A review of the available geophysical information on the island shows that it has been obtained mainly through marine and airborne geophysical surveys. A new dataset comprising 100 magnetotelluric soundings acquired on land has been used to obtain the first 3D electrical resistivity model of the island at crustal scale. The model shows high resistivity values close to the surface in the exposed Tejeda Caldera that extends at depth to the SE cutting the islands in half. Outside the inferred limits of the Tejeda Caldera the 3D model shows low resistivity values that could be explained by hydrothermal alteration at deeper levels and the presence of marine saltwater intrusion at shallower levels near the coast. The presence of unconnected vertical-like structures, with very low resistivity (<10 ohm m) could be associated to small convective cells is confirmed by the sensitivity analysis carried out in the present study. Those structures are the most likely candidates for a detailed analysis in order to determine their geothermal economic potential. A comprehensive review of existing geophysical data and models of Gran Canaria island and their comparison with the new 3D electrical resistivity model is presented.

Highlights • Loess in Northern Iran sensitively reacted on Pleistocene climate change. • Dust accumulation and soil formation followed climate cycles of the Northern Hemisphere. • Pedostratigraphic correlation provides link within Southern Eurasian loess belt. • Southern Caspian Lowlands provide key area for climate reconstruction in West Asia. The southern Caspian Lowland sensitively reacted to Pleistocene climate change and is a key area for reconstructing climate dynamics and landscape evolution in Southern Eurasia. Loess-paleosol sequences (LPS) of the northern foothills of Alborz Mountains provide detailed records of climate-induced changes of dust accumulation and soil formation correlating with relatively dry or moist conditions of the past. The LPS at Neka-Abelou (NA) was studied in detail in order to understand these dynamics and provide a base for regional pedostratigraphic correlation. We have carried out high-resolution analyses of grain-size, sediment color, mass specific and frequency-dependent magnetic susceptibility and carbonate content and established a temporal framework as based on luminescence dating using a post-IR infrared stimulated luminescence (pIRIR) protocol and fading corrections. The LPS of NA is composed of finely textured loess, which is subdivided by at least eleven paleosols. Moreover, it contains a thin loess layer with lenses of trachytic tephra which most likely originated from the Damavand volcano and was deposited during Marine Isotope Stage (MIS) 4. The lower part of the LPS at NA consists of four strongly-developed reddish-brown paleosols (Bt, Btg horizons) separated by thin layers of pedogenically-altered loess indicating moist climate conditions and low dust accumulation rates during the Middle Pleistocene or earlier. The central part contains a pedocomplex of clay-rich paleosols composed of well developed Bt, ABk and Bw horizons formed under strongly reduced dust accumulation rates and intercalated by loess layers. Pedostr

Highlights • Rhine River, largest river in Western Europe dammed repeatedly by Plinian fallout. • Damming mechanism resembled formation of ice jams. • Dam was a partly floating, partly grounded plug of buoyant pumice throughout many km. • Major damming site located 12 km upstream of maximum tephra loading. The Rhine River (Germany) - the largest river in Western Europe - was dammed by pyroclastic material multiple times during the major Plinian Laacher See Eruption (12,900 BP). Dams formed both upstream and downstream of the broad tectonic Lower Neuwied Basin (LNB) which interrupts the narrow Rhine canyon. Here we document upstream damming of the Rhine River at the entrance to the LNB close to the present city of Koblenz due to overloading with tephra fall into the Rhine and its major tributaries, the Moselle and the Lahn. The dam was formed repeatedly during rapid pumice tephra fall events and became breached during breaks in eruptive activity, causing extensive, high-energy flooding throughout the entire basin. The ephemeral Koblenz dams differed significantly from “normal” volcanically-induced dams by being composed principally of washed-together pumice clasts and some driftwood. The porous nature of pumice and its ability to absorb water were crucial factors. Thus, a large volume percentage of the tephra that had fallen into the Rhine floated submerged within the upper part of the water column or swam at the surface. Moreover, the absorption of the river water by the pumice clasts increased the sediment:water ratio of the two-phase flow considerably. We here present a model of dam formation resembling the formation of ice jams. We visualize the Koblenz dams to have been elongate, partly floating and partly grounded, permeable plugs several kilometers long and rising no higher than the flood plain. Damming was most plausibly initiated in the LNB within the area of maximum tephra loading and propagated upstream in a chain reaction similar to the formation of traff

Highlights • First recorded example of tephra fallout damming a major river. • Repeated massive syn-eruptive damming. • Floods triggered by multiple breaches of tephra dam caused widespread mass erosion. • Striking large-scale upper flow regime deposits. Abstract The Rhine - the largest river in Western Europe – was dammed during the Plinian Laacher See Eruption (LSE; 12,900 BP). Damming during the climactic Plinian episode of LSE occurred both upstream and downstream of the broad tectonic Lower Neuwied Basin (LNB) that interrupts the narrow Rhine canyon. We here document details of the upstream damming at the bottleneck entrance to the LNB near the present city of Koblenz. Our reconstruction is based on a high-resolution analysis and correlation of the complex intercalation of primary fallout tephra relics with fluvially reworked Laacher See Tephra in the LNB. Tephra units representing complete eruptive cycles repeatedly fell on drained ground in between one minor and 4 major flooding events - even at the base of side channels that had been active prior to the LSE and that had been flooded by a preceding flooding event. This demonstrates that flooding occurred generally during breaks and not during fallout events. The repeated formation and breach of a dam at the upstream entrance of the LNB (Koblenz Dam) consisting of fallout components and driftwood washed together convincingly explains the multiple repetition of the drainage of the channels in the LNB followed by large-magnitude flooding in rapid succession. The strongly pulsating nature of the LSE reflected in multiple interruptions of eruptive activity fundamentally controlled the damming and flooding dynamics. The Rhine became completely blocked during distinct fallout phases due to overloading with pumice that had fallen into the river and its major tributaries. The temporary dam collapsed during eruptive breaks. This is the first recorded example of tephra fallout damming a major watercourse. The extremely

Highlights • Previous age estimates of the Laacher See Eruptions (LSE) around 12,900 years are still diverging and imprecise. • The combination of dendrochronology, wood anatomy, and 14C measurements holds the potential to establish a precise LSE date. • An absolute calendric date of the LSE would improve the synchronization of European Late Glacial to Holocene archives. Abstract The precise date of the Laacher See eruption (LSE), central Europe’s largest Late Pleistocene volcanic event that occurred around 13,000 years ago, is still unknown. Here, we outline the potential of combined high-resolution dendrochronological, wood anatomical and radiocarbon (14C) measurements, to refine the age of this major Plinian eruption. Based on excavated, subfossil trees that were killed during the explosive LSE and buried under its pyroclastic deposits, we describe how a firm date of the eruption might be achieved, and how the resulting temporal precision would further advance our understanding of the environmental and societal impacts of this event. Moreover, we discuss the relevance of an accurate LSE date for improving the synchronization of European terrestrial and lacustrine Late Glacial to Holocene archives.

Highlights • Rhine River, largest river in Western Europe dammed repeatedly by Plinian fallout. • Damming mechanism resembled formation of ice jams. • Dam was a partly floating, partly grounded plug of buoyant pumice throughout many km. • Major damming site located 12 km upstream of maximum tephra loading. The Rhine River (Germany) - the largest river in Western Europe - was dammed by pyroclastic material multiple times during the major Plinian Laacher See Eruption (12,900 BP). Dams formed both upstream and downstream of the broad tectonic Lower Neuwied Basin (LNB) which interrupts the narrow Rhine canyon. Here we document upstream damming of the Rhine River at the entrance to the LNB close to the present city of Koblenz due to overloading with tephra fall into the Rhine and its major tributaries, the Moselle and the Lahn. The dam was formed repeatedly during rapid pumice tephra fall events and became breached during breaks in eruptive activity, causing extensive, high-energy flooding throughout the entire basin. The ephemeral Koblenz dams differed significantly from “normal” volcanically-induced dams by being composed principally of washed-together pumice clasts and some driftwood. The porous nature of pumice and its ability to absorb water were crucial factors. Thus, a large volume percentage of the tephra that had fallen into the Rhine floated submerged within the upper part of the water column or swam at the surface. Moreover, the absorption of the river water by the pumice clasts increased the sediment:water ratio of the two-phase flow considerably. We here present a model of dam formation resembling the formation of ice jams. We visualize the Koblenz dams to have been elongate, partly floating and partly grounded, permeable plugs several kilometers long and rising no higher than the flood plain. Damming was most plausibly initiated in the LNB within the area of maximum tephra loading and propagated upstream in a chain reaction similar to the formation of traff

Highlights • Previous age estimates of the Laacher See Eruptions (LSE) around 12,900 years are still diverging and imprecise. • The combination of dendrochronology, wood anatomy, and 14C measurements holds the potential to establish a precise LSE date. • An absolute calendric date of the LSE would improve the synchronization of European Late Glacial to Holocene archives. Abstract The precise date of the Laacher See eruption (LSE), central Europe’s largest Late Pleistocene volcanic event that occurred around 13,000 years ago, is still unknown. Here, we outline the potential of combined high-resolution dendrochronological, wood anatomical and radiocarbon (14C) measurements, to refine the age of this major Plinian eruption. Based on excavated, subfossil trees that were killed during the explosive LSE and buried under its pyroclastic deposits, we describe how a firm date of the eruption might be achieved, and how the resulting temporal precision would further advance our understanding of the environmental and societal impacts of this event. Moreover, we discuss the relevance of an accurate LSE date for improving the synchronization of European terrestrial and lacustrine Late Glacial to Holocene archives.

Highlights • First recorded example of tephra fallout damming a major river. • Repeated massive syn-eruptive damming. • Floods triggered by multiple breaches of tephra dam caused widespread mass erosion. • Striking large-scale upper flow regime deposits. Abstract The Rhine - the largest river in Western Europe – was dammed during the Plinian Laacher See Eruption (LSE; 12,900 BP). Damming during the climactic Plinian episode of LSE occurred both upstream and downstream of the broad tectonic Lower Neuwied Basin (LNB) that interrupts the narrow Rhine canyon. We here document details of the upstream damming at the bottleneck entrance to the LNB near the present city of Koblenz. Our reconstruction is based on a high-resolution analysis and correlation of the complex intercalation of primary fallout tephra relics with fluvially reworked Laacher See Tephra in the LNB. Tephra units representing complete eruptive cycles repeatedly fell on drained ground in between one minor and 4 major flooding events - even at the base of side channels that had been active prior to the LSE and that had been flooded by a preceding flooding event. This demonstrates that flooding occurred generally during breaks and not during fallout events. The repeated formation and breach of a dam at the upstream entrance of the LNB (Koblenz Dam) consisting of fallout components and driftwood washed together convincingly explains the multiple repetition of the drainage of the channels in the LNB followed by large-magnitude flooding in rapid succession. The strongly pulsating nature of the LSE reflected in multiple interruptions of eruptive activity fundamentally controlled the damming and flooding dynamics. The Rhine became completely blocked during distinct fallout phases due to overloading with pumice that had fallen into the river and its major tributaries. The temporary dam collapsed during eruptive breaks. This is the first recorded example of tephra fallout damming a major watercourse. The extremely

More than 1500 km of multi-channel seismic reflection profiles combined with ICDP (International Continental Scientific Drilling Program) drilling data, provide important insights into the stratigraphic evolution of Lake Van, eastern Turkey. Three major basins (Tatvan, Northern and the Deveboynu basins) comprise the main lake basin and are separated by morphological highs (Ahlat ridge and Northern ridge). Moreover, NE–SW faults, parallel to the general tectonic lineament of the area, dominate the entire basin and are in charge of creating graben and half-graben structures. Well-developed prograding deltaic sequences on top of the basement were recognized by seismic stratigraphy analysis. Most likely, they formed during the initial flooding of Lake Van ∼600 ka. The Tatvan basin sediments are dominated by mass-flow deposits of various origins alternating with undisturbed lacustrine sediments including distinct tephra layers. Faulting along the Tatvan basin margins may have triggered margin-wide slope failures. Ahlat ridge started to form between ca 340 ka–290 ka. Since then, Ahlat ridge was sheltered from major mass-flows due to its elevation. Hence, slow lacustrine sedimentation has prevailed throughout lake history on Ahlat ridge, which was the location of the main drill site during the ICDP. Several lake level fluctuations are evident on the eastern slope area but the deep basins were permanently covered by water. A significant lake-level low stand (ca 600 ka BP) was found at ∼610 m below present lake level. The setting of the lake changed at about 30 ka. Tectonic activity appears to have waned significantly as the mass-transport deposition decreased across the Tatvan basin while normal undisturbed lacustrine sedimentation prevailed. A different setting is found in the Northern basin from ca 90 ka to Present, especially due to the strong influx of mostly volcaniclastic turbidites causing sedimentation rates to be about 3.5 times higher (drill Site 1), than at Site

Fifteen Lateglacial to Holocene rhyolitic, dominantly primary tephra layers piston-cored and drilled (ICDP Paleovan drilling project) in western Lake Van (eastern Anatolia, Turkey) were precisely correlated to either of the two adjacent and active large volcanoes Nemrut and Süphan based on shard textures, mineralogy and mineral and glass compositions. The young peralkaline (comenditic to pantelleritic) primary rhyolitic Nemrut tephras are characterized by anorthoclase, hedenbergitic to augitic clinopyroxene, fayalitic olivine, minor quartz, and rare accessory chevkinite and zircon. Phenocrysts in subalkaline primary rhyolitic Süphan tephras are chiefly oligoclase-labradorite, with minor K-rich sanidine in some, biotite, amphibole, hypersthene, rare augitic clinopyroxene, relatively common allanite and rare zircon. Two contrasting explosive eruptive modes are distinguished from each other: episodic (Süphan) and periodic (Nemrut). The Lateglacial Süphan tephra swarm covers a short time interval of ca. 338 years between ca. 13,078 vy BP and 12,740 vy BP, eruptions having occurred statistically every ca. 42 years with especially short intervals between V-11 (reworked) and V-14. Causes for the strongly episodic Süphan explosive behavior might include seismic triggering of a volcano–magma system unable to erupt explosively without the benefit of external triggering, as reflected in pervasive faulting preceding the Süphan tephra swarm. Seismic triggering may have caused the rise of more mafic (“trachyandesitic”) parent magma, heating near-surface pockets of highly evolved magma – that might have formed silicic domes during this stage of volcano evolution – resulting in ascent and finally explosive fragmentation of magma essentially by external factors, probably significantly enhanced by magma–water/ice interaction. Explosive eruptions of the Nemrut volcano system, interpreted to be underlain by a large fractionating magma reservoir, follow a more periodic mode of (a) long

Fifteen Lateglacial to Holocene rhyolitic, dominantly primary tephra layers piston-cored and drilled (ICDP Paleovan drilling project) in western Lake Van (eastern Anatolia, Turkey) were precisely correlated to either of the two adjacent and active large volcanoes Nemrut and Süphan based on shard textures, mineralogy and mineral and glass compositions. The young peralkaline (comenditic to pantelleritic) primary rhyolitic Nemrut tephras are characterized by anorthoclase, hedenbergitic to augitic clinopyroxene, fayalitic olivine, minor quartz, and rare accessory chevkinite and zircon. Phenocrysts in subalkaline primary rhyolitic Süphan tephras are chiefly oligoclase-labradorite, with minor K-rich sanidine in some, biotite, amphibole, hypersthene, rare augitic clinopyroxene, relatively common allanite and rare zircon. Two contrasting explosive eruptive modes are distinguished from each other: episodic (Süphan) and periodic (Nemrut). The Lateglacial Süphan tephra swarm covers a short time interval of ca. 338 years between ca. 13,078 vy BP and 12,740 vy BP, eruptions having occurred statistically every ca. 42 years with especially short intervals between V-11 (reworked) and V-14. Causes for the strongly episodic Süphan explosive behavior might include seismic triggering of a volcano–magma system unable to erupt explosively without the benefit of external triggering, as reflected in pervasive faulting preceding the Süphan tephra swarm. Seismic triggering may have caused the rise of more mafic (“trachyandesitic”) parent magma, heating near-surface pockets of highly evolved magma – that might have formed silicic domes during this stage of volcano evolution – resulting in ascent and finally explosive fragmentation of magma essentially by external factors, probably significantly enhanced by magma–water/ice interaction. Explosive eruptions of the Nemrut volcano system, interpreted to be underlain by a large fractionating magma reservoir, follow a more periodic mode of (a) long

More than 1500 km of multi-channel seismic reflection profiles combined with ICDP (International Continental Scientific Drilling Program) drilling data, provide important insights into the stratigraphic evolution of Lake Van, eastern Turkey. Three major basins (Tatvan, Northern and the Deveboynu basins) comprise the main lake basin and are separated by morphological highs (Ahlat ridge and Northern ridge). Moreover, NE–SW faults, parallel to the general tectonic lineament of the area, dominate the entire basin and are in charge of creating graben and half-graben structures. Well-developed prograding deltaic sequences on top of the basement were recognized by seismic stratigraphy analysis. Most likely, they formed during the initial flooding of Lake Van ∼600 ka. The Tatvan basin sediments are dominated by mass-flow deposits of various origins alternating with undisturbed lacustrine sediments including distinct tephra layers. Faulting along the Tatvan basin margins may have triggered margin-wide slope failures. Ahlat ridge started to form between ca 340 ka–290 ka. Since then, Ahlat ridge was sheltered from major mass-flows due to its elevation. Hence, slow lacustrine sedimentation has prevailed throughout lake history on Ahlat ridge, which was the location of the main drill site during the ICDP. Several lake level fluctuations are evident on the eastern slope area but the deep basins were permanently covered by water. A significant lake-level low stand (ca 600 ka BP) was found at ∼610 m below present lake level. The setting of the lake changed at about 30 ka. Tectonic activity appears to have waned significantly as the mass-transport deposition decreased across the Tatvan basin while normal undisturbed lacustrine sedimentation prevailed. A different setting is found in the Northern basin from ca 90 ka to Present, especially due to the strong influx of mostly volcaniclastic turbidites causing sedimentation rates to be about 3.5 times higher (drill Site 1), than at Site

Thirty-two new single crystal ages document 400 000 years of widespread explosive volcanism of historically active Nemrut Volcano towering over huge alkaline Lake Van (Eastern Anatolia). The dated deposits were selected to monitor the volcanic and compositional evolution of Nemrut Volcano through time and thus to provide a rigorous temporal framework for the tephra record of the PaleoVan Drilling Project. Tephra samples were taken from large-volume deposits or those that occur in medial to distal localities, well-exposed stratigraphic sections or from the initial phase of an eruptive sequence. Mainly fallout deposits were chosen because most ignimbrites show more complex and corroded feldspar populations owing to compositional zoning and magma mixing. Moreover, fallout deposits held the promise to be more clearly identifiable with-and correlatable to->300 tephra layers in the PaleoVan drill cores, even though commonly in amounts marginal or insufficient in thickness to allow well-supported single crystal dating. The crystals dated are dominantly anorthoclase, the main phenoctyst phase in the trachytic to rhyolitic, slightly to strongly peralkaline Nemrut magmas. Ages obtained so far range from ca. 400 ka to ca. 30 ka for Nemrut Volcano. The causes of significant changes in the frequency, volume and composition of tephra layers per unit time are discussed in terms of external (erosion, climate changes, geodynamic factors) and internal forcing (changes in magma supply and composition and incubation periods preceding large volume rhyolitic eruptions). For example, the low frequency of tephra layers deposited prior to ca. 200 ka may be due to low explosive activity, severe erosion between MIS 9 and MIS 11, or both. Nevertheless, the overall frequency of explosive eruptions appears to have increased during the past ca. 200 ka. We also recognize a slight peak in explosive eruptions during warm periods (e.g. MIS 5 and MIS 7) and speculate on lithospheric unloading triggeri

Thirty-two new single crystal ages document 400 000 years of widespread explosive volcanism of historically active Nemrut Volcano towering over huge alkaline Lake Van (Eastern Anatolia). The dated deposits were selected to monitor the volcanic and compositional evolution of Nemrut Volcano through time and thus to provide a rigorous temporal framework for the tephra record of the PaleoVan Drilling Project. Tephra samples were taken from large-volume deposits or those that occur in medial to distal localities, well-exposed stratigraphic sections or from the initial phase of an eruptive sequence. Mainly fallout deposits were chosen because most ignimbrites show more complex and corroded feldspar populations owing to compositional zoning and magma mixing. Moreover, fallout deposits held the promise to be more clearly identifiable with-and correlatable to->300 tephra layers in the PaleoVan drill cores, even though commonly in amounts marginal or insufficient in thickness to allow well-supported single crystal dating. The crystals dated are dominantly anorthoclase, the main phenoctyst phase in the trachytic to rhyolitic, slightly to strongly peralkaline Nemrut magmas. Ages obtained so far range from ca. 400 ka to ca. 30 ka for Nemrut Volcano. The causes of significant changes in the frequency, volume and composition of tephra layers per unit time are discussed in terms of external (erosion, climate changes, geodynamic factors) and internal forcing (changes in magma supply and composition and incubation periods preceding large volume rhyolitic eruptions). For example, the low frequency of tephra layers deposited prior to ca. 200 ka may be due to low explosive activity, severe erosion between MIS 9 and MIS 11, or both. Nevertheless, the overall frequency of explosive eruptions appears to have increased during the past ca. 200 ka. We also recognize a slight peak in explosive eruptions during warm periods (e.g. MIS 5 and MIS 7) and speculate on lithospheric unloading triggeri

The 20–25 m thick trachyphonolitic Ayagaures ignimbrite cooling unit [(AY); 11.8 Ma] exposed over 250 km2 (onshore volume ca. 4.5 km3 DRE) is the uppermost and most voluminous cooling unit of the Middle Fataga Formation (MFF), part of the Fataga Group (ca. 13.3–ca. 9 Ma) on Gran Canaria (GC), Canary Islands (28°00′ N, 15°35′ W). Up to 19 flow units (named b–t) subdividing the AY have been identified throughout most of the area from proximally to the caldera wall to distally as far as 14 km away. Individual flow units were distinguished from each other and logged using mainly chemical criteria. Single and/or packages of flow units (A, B and C) are tentatively interpreted to correspond to compositionally distinct magma bodies erupted from the same magma reservoir. These source-controlled flow units are interpreted to reflect successive eruptive pulses during incremental subsidence of Tejeda caldera. We subdivided AY cooling unit into four welding facies. Tentative correlation with a major syn-ignimbrite turbidite drilled during ODP Leg 157 suggests a total DRE volume of > 50 km3. The cooling unit as a whole becomes less evolved upwards as shown by major elements, trace elements and REE of bulk rock and phenocrysts. All phenocryst phases, dominantly sanidine–anorthoclase (up to 20 vol.%), with minor biotite, augite, titanite, haüyne and apatite, are unzoned and show an incremental compositional zoning in the stratigraphy. The shallow level parent magma reservoir is interpreted to have undergone strong mixing prior to starting its final compositional zoning in a thermodynamically equilibrated reservoir. Compositional zoning resulted in three main bodies. This compositional and physical layering may have been triggered by rapid growth of alkali feldspar and biotite throughout the erupted part of the magma chamber. Abundant titanite and haüyne phenocrysts in basal flow units and in a locally preserved, highly evolved fallout tephra are interpreted to reflect initial evacu

Nejapa Maar (2.5 × 1.4 km, c. 120 m deep), the largest maar along the 15-km-long Holocene Nejapa–Miraflores Lineament (NML), is the source vent of the youngest relatively widespread basaltic tholeiitic tephra blanket (Nejapa Tephra: NT) in western central Nicaragua, as shown by isopachs and isopleths (Rausch and Schmincke, 2008). The NT covers an area of > 10 km2 in W/NW Managua. The minimum total magma volume erupted is estimated as 0.09 km3. Juvenile, dominantly slightly vesicular (20–40 vol.%) basically tachylitic cauliflower-shaped lapilli with an average density of 2.1 g/cm3, make up > 90 vol.% of the deposit, while lithoclasts comprise < 10 vol.% except proximally. This, the paucity of fine-grained tuffs and the dominant plane-parallel bedding all suggest fragmentation by shallow interaction of a rising magma starting to vesiculate and fragment pyroclastically with external water. The complex particles so generated erupted in moderately high eruption columns (at least 7–10 km) and were dominantly deposited as dry to damp, warm to cool fallout. Minor surge transport is inferred from fine-grained, locally cross-bedded tephra beds chiefly north of Nejapa and just west of Asososca Maars. Synvolcanic faulting along the NML is inferred. Faults in the study area indicate that activation of the N–S-trending Nejapa–Miraflores Fault (NMF), representing the western flank of Managua Graben, preceded deposition of NT and underlying Masaya Tuff (c.1.8 ka BP), Chiltepe Pumice (c. 1.9 ka BP) and Masaya Triple Layer (2.1 ka BP). The NT deposit is underlain regionally by a paleosol and topped by a soil. The basal paleosol contains pottery sherds made by the Usulután negative technique during the Late Formative period (700 BCE–300 CE) (2.7–1.7 ka BP). The soil overlying NT contains pottery related to the Ometepe technique dated as between 1350 and 1550 CE (650–450 a BP). These, and the radiocarbon dates of the pottery-bearing paleosols (1245 ± 125 and 535 ± 110 a BP) obtained by

Nejapa Maar (2.5 × 1.4 km, c. 120 m deep), the largest maar along the 15-km-long Holocene Nejapa–Miraflores Lineament (NML), is the source vent of the youngest relatively widespread basaltic tholeiitic tephra blanket (Nejapa Tephra: NT) in western central Nicaragua, as shown by isopachs and isopleths (Rausch and Schmincke, 2008). The NT covers an area of > 10 km2 in W/NW Managua. The minimum total magma volume erupted is estimated as 0.09 km3. Juvenile, dominantly slightly vesicular (20–40 vol.%) basically tachylitic cauliflower-shaped lapilli with an average density of 2.1 g/cm3, make up > 90 vol.% of the deposit, while lithoclasts comprise < 10 vol.% except proximally. This, the paucity of fine-grained tuffs and the dominant plane-parallel bedding all suggest fragmentation by shallow interaction of a rising magma starting to vesiculate and fragment pyroclastically with external water. The complex particles so generated erupted in moderately high eruption columns (at least 7–10 km) and were dominantly deposited as dry to damp, warm to cool fallout. Minor surge transport is inferred from fine-grained, locally cross-bedded tephra beds chiefly north of Nejapa and just west of Asososca Maars. Synvolcanic faulting along the NML is inferred. Faults in the study area indicate that activation of the N–S-trending Nejapa–Miraflores Fault (NMF), representing the western flank of Managua Graben, preceded deposition of NT and underlying Masaya Tuff (c.1.8 ka BP), Chiltepe Pumice (c. 1.9 ka BP) and Masaya Triple Layer (2.1 ka BP). The NT deposit is underlain regionally by a paleosol and topped by a soil. The basal paleosol contains pottery sherds made by the Usulután negative technique during the Late Formative period (700 BCE–300 CE) (2.7–1.7 ka BP). The soil overlying NT contains pottery related to the Ometepe technique dated as between 1350 and 1550 CE (650–450 a BP). These, and the radiocarbon dates of the pottery-bearing paleosols (1245 ± 125 and 535 ± 110 a BP) obtained by

The 20–25 m thick trachyphonolitic Ayagaures ignimbrite cooling unit [(AY); 11.8 Ma] exposed over 250 km2 (onshore volume ca. 4.5 km3 DRE) is the uppermost and most voluminous cooling unit of the Middle Fataga Formation (MFF), part of the Fataga Group (ca. 13.3–ca. 9 Ma) on Gran Canaria (GC), Canary Islands (28°00′ N, 15°35′ W). Up to 19 flow units (named b–t) subdividing the AY have been identified throughout most of the area from proximally to the caldera wall to distally as far as 14 km away. Individual flow units were distinguished from each other and logged using mainly chemical criteria. Single and/or packages of flow units (A, B and C) are tentatively interpreted to correspond to compositionally distinct magma bodies erupted from the same magma reservoir. These source-controlled flow units are interpreted to reflect successive eruptive pulses during incremental subsidence of Tejeda caldera. We subdivided AY cooling unit into four welding facies. Tentative correlation with a major syn-ignimbrite turbidite drilled during ODP Leg 157 suggests a total DRE volume of > 50 km3. The cooling unit as a whole becomes less evolved upwards as shown by major elements, trace elements and REE of bulk rock and phenocrysts. All phenocryst phases, dominantly sanidine–anorthoclase (up to 20 vol.%), with minor biotite, augite, titanite, haüyne and apatite, are unzoned and show an incremental compositional zoning in the stratigraphy. The shallow level parent magma reservoir is interpreted to have undergone strong mixing prior to starting its final compositional zoning in a thermodynamically equilibrated reservoir. Compositional zoning resulted in three main bodies. This compositional and physical layering may have been triggered by rapid growth of alkali feldspar and biotite throughout the erupted part of the magma chamber. Abundant titanite and haüyne phenocrysts in basal flow units and in a locally preserved, highly evolved fallout tephra are interpreted to reflect initial evacu

A model of diatreme evolution in a shallow marine setting is based on a multi-disciplinary analysis of diatremes in the Iblean Mountains (Sicily). The approach includes stratigraphic, volcanological, structural, petrologic and compositional data. We invoke a complex interplay of internal (rapid ascent and pyroclastic fragmentation of a volatile (CO2)-rich nephelinitic magma at depth) and external factors. These comprise hydroclastic explosions due to near-surface interaction of the rising particle/volatile mixture with seawater and water-saturated lime mud. Other external factors contributing to diatreme formation include regional and local tectonics (graben formation in pull-apart motion) combined with lateral pipe enlargement by bedrock-spalling and radial block subsidence into the diatreme pipe. We suggest that fragmentation of volatile-rich magma due to internal eruption forcing was fundamental in the formation of the Iblean shallow marine diatremes. Internal and external factors may act to a variable degree, however, during diatreme evolution in general.

The Masaya Caldera Complex has been the site of three highly explosive basaltic eruptions within the last six thousand years. A Plinian eruption ca. 2 ka ago formed the widespread deposits of the Masaya Triple Layer. We distinguish two facies within the Masaya Triple Layer from each other: La Concepción facies to the south and Managua facies to the northwest. These two facies were previously treated as two separated deposits (La Concepción Tephra and the Masaya Triple Layer of Pérez and Freundt, 2006) because of their distinct regional distribution and internal architectures. However, chemical compositions of bulk rock, matrix and inclusion glasses and mineral phases demonstrate that they are the product of a single basaltic magma batch. Additionally, a marker bed containing fluidal-shaped vesicular lapilli allowed us to make a plausible correlation between the two facies, also supported by consistent lateral changes in lithologic structure and composition, thickness and grain size. We distinguish 10 main subunits of the Masaya Triple Layer (I to X), with bulk volumes ranging between 0.02 and 0.22 km3, adding up to 0.86 km3 (0.4 km3 DRE) for the entire deposit. Distal deposits identified in two cores drilled offshore Nicaragua, at a distance of ∼ 170 km from the Masaya Caldera Complex, increase the total tephra volume to 3.4 km3 or ∼ 1.8 km3 DRE of erupted basaltic magma. Isopleth data of five major fallout subunits indicate mass discharges of 106 to 108 kg/s and eruption columns of 21 to 32 km height, affected by wind speeds of < 2 m/s to ∼ 20 m/s which increased during the course of the multi-episodic eruption. Magmatic Plinian events alternated with phreatoplinian eruptions and phreatomagmatic explosions generating surges that typically preceded breaks in activity. While single eruptive episodes lasted for few hours, the entire eruption probable lasted weeks to months. This is indicated by changes in atmospheric conditions and ash-layer surfaces that had become m

The Earth's mantle is chemically and isotopically heterogeneous, and a component of recycled oceanic crust is generally suspected in the convecting mantle [Hofmann and White, 1982. Mantle plumes from ancient oceanic crust. Earth Planet. Sci. Lett. 57, 421–436]. Indeed, the HIMU component (high µ = 238U/204Pb), one of four isotopically distinct end-members in the Earth's mantle, is generally attributed to relatively old (≥ 1–2 Ga) recycled oceanic crust in the form of eclogite/pyroxenite, e.g. [Zindler and Hart, 1986. Chemical geodynamics. Ann. Rev. Earth Planet. Sci. 14, 493–571]. Although the presence of the recycled component is generally supported by element and isotopic data, little is known about its physical state at mantle depths. Here we show that the concentrations of Ni, Mn and Ca in olivine from the Canarian shield stage lavas, which can be used to assess the physical nature of the source material (peridotite versus olivine-free pyroxenite) [Sobolev et al., 2007. The amount of recycled crust in sources of mantle-derived melts. Science 316, 412–417], correlate strongly with bulk rock Sr, Nd and Pb isotopic ratios. The most important result following from our data is that the enriched, HIMU-type (having higher 206Pb/204Pb than generally found in the other mantle end-members) signature of the Canarian hotspot magmas was not caused by a pyroxenite/eclogite constituent of the plume but appears to have been primarily hosted by peridotite. This implies that the old (older than ~ 1 Ga) ocean crust, which has more evolved radiogenic isotope compositions, was stirred into/reacted with the mantle so that there is not significant eclogite left, whereas younger recycled oceanic crust with depleted MORB isotopic signature (< 1 Ga) can be preserved as eclogite, which when melted can generate reaction pyroxenite.

The Masaya Caldera Complex has been the site of three highly explosive basaltic eruptions within the last six thousand years. A Plinian eruption ca. 2 ka ago formed the widespread deposits of the Masaya Triple Layer. We distinguish two facies within the Masaya Triple Layer from each other: La Concepción facies to the south and Managua facies to the northwest. These two facies were previously treated as two separated deposits (La Concepción Tephra and the Masaya Triple Layer of Pérez and Freundt, 2006) because of their distinct regional distribution and internal architectures. However, chemical compositions of bulk rock, matrix and inclusion glasses and mineral phases demonstrate that they are the product of a single basaltic magma batch. Additionally, a marker bed containing fluidal-shaped vesicular lapilli allowed us to make a plausible correlation between the two facies, also supported by consistent lateral changes in lithologic structure and composition, thickness and grain size. We distinguish 10 main subunits of the Masaya Triple Layer (I to X), with bulk volumes ranging between 0.02 and 0.22 km3, adding up to 0.86 km3 (0.4 km3 DRE) for the entire deposit. Distal deposits identified in two cores drilled offshore Nicaragua, at a distance of ∼ 170 km from the Masaya Caldera Complex, increase the total tephra volume to 3.4 km3 or ∼ 1.8 km3 DRE of erupted basaltic magma. Isopleth data of five major fallout subunits indicate mass discharges of 106 to 108 kg/s and eruption columns of 21 to 32 km height, affected by wind speeds of < 2 m/s to ∼ 20 m/s which increased during the course of the multi-episodic eruption. Magmatic Plinian events alternated with phreatoplinian eruptions and phreatomagmatic explosions generating surges that typically preceded breaks in activity. While single eruptive episodes lasted for few hours, the entire eruption probable lasted weeks to months. This is indicated by changes in atmospheric conditions and ash-layer surfaces that had become m

A model of diatreme evolution in a shallow marine setting is based on a multi-disciplinary analysis of diatremes in the Iblean Mountains (Sicily). The approach includes stratigraphic, volcanological, structural, petrologic and compositional data. We invoke a complex interplay of internal (rapid ascent and pyroclastic fragmentation of a volatile (CO2)-rich nephelinitic magma at depth) and external factors. These comprise hydroclastic explosions due to near-surface interaction of the rising particle/volatile mixture with seawater and water-saturated lime mud. Other external factors contributing to diatreme formation include regional and local tectonics (graben formation in pull-apart motion) combined with lateral pipe enlargement by bedrock-spalling and radial block subsidence into the diatreme pipe. We suggest that fragmentation of volatile-rich magma due to internal eruption forcing was fundamental in the formation of the Iblean shallow marine diatremes. Internal and external factors may act to a variable degree, however, during diatreme evolution in general.

The Earth's mantle is chemically and isotopically heterogeneous, and a component of recycled oceanic crust is generally suspected in the convecting mantle [Hofmann and White, 1982. Mantle plumes from ancient oceanic crust. Earth Planet. Sci. Lett. 57, 421–436]. Indeed, the HIMU component (high µ = 238U/204Pb), one of four isotopically distinct end-members in the Earth's mantle, is generally attributed to relatively old (≥ 1–2 Ga) recycled oceanic crust in the form of eclogite/pyroxenite, e.g. [Zindler and Hart, 1986. Chemical geodynamics. Ann. Rev. Earth Planet. Sci. 14, 493–571]. Although the presence of the recycled component is generally supported by element and isotopic data, little is known about its physical state at mantle depths. Here we show that the concentrations of Ni, Mn and Ca in olivine from the Canarian shield stage lavas, which can be used to assess the physical nature of the source material (peridotite versus olivine-free pyroxenite) [Sobolev et al., 2007. The amount of recycled crust in sources of mantle-derived melts. Science 316, 412–417], correlate strongly with bulk rock Sr, Nd and Pb isotopic ratios. The most important result following from our data is that the enriched, HIMU-type (having higher 206Pb/204Pb than generally found in the other mantle end-members) signature of the Canarian hotspot magmas was not caused by a pyroxenite/eclogite constituent of the plume but appears to have been primarily hosted by peridotite. This implies that the old (older than ~ 1 Ga) ocean crust, which has more evolved radiogenic isotope compositions, was stirred into/reacted with the mantle so that there is not significant eclogite left, whereas younger recycled oceanic crust with depleted MORB isotopic signature (< 1 Ga) can be preserved as eclogite, which when melted can generate reaction pyroxenite.

High precision measurements of 226Ra–230Th–238U disequilibria and Ba concentrations are reported for samples from two chemically zoned trachyte deposits from Fogo volcano, São Miguel, Azores. High-precision U-series disequilibria measurements by plasma ionization multicollector mass spectrometry were performed on pumice lapilli and volcanic glass separates from Fogo 1563 A.D. (∼ 0.14 km3) and the ∼ 4.7 ka Fogo A (∼ 0.7 km3) deposits in order to quantify the time scales of magmatic processes. Observed (226Ra)/Ba–(230Th)/Ba relationships in Fogo 1563 are compatible with a conventional instantaneous fractional crystallization model and a pre-eruptive magma residence time of ∼ 50 y. However, the Fogo A data cannot be explained by instantaneous fractional crystallization, and require a prolonged crystallization history. Continuous differentiation models better explain the observed 226Ra–230Th variations within the Fogo deposits and may be more realistic in general. Such models suggest magma residence times prior to eruption of ∼ 50–80 years for Fogo 1563, and ∼ 4.7 ka for the larger volume Fogo A eruption. These time scales represent liquid residence ages rather than the crystallization ages documented in most previous magmatic time scale studies, and allow constraints to be placed on the time scales necessary for the development of chemical zonation within the Fogo magma chamber. Our results indicate that calculated time scales are relatively insensitive to the precise nature of the continuous differentiation models, and further indicate that meaningful magma differentiation time scales can be obtained despite open system behavior, because the Ra–Th disequilibria are overwhelmingly controlled by feldspar fractionation. Calculated time scales are, however, extremely sensitive to DRa/DBa ratios. We therefore emphasize the crucial importance to better constrain the relative partitioning of Ra and Ba when employing Ra–Th disequilibrium data to constrain the rates and time s

The stratigraphic succession of widespread tephra layers in west-central Nicaragua was emplaced by highly explosive eruptions from mainly three volcanoes: the Chiltepe volcanic complex and the Masaya and Apoyo calderas. Stratigraphic correlations are based on distinct compositions of tephras. The total tephras combine to a total on-shore volume of about 37 km3 produced during the last ∼ 60 ka. The total erupted magma mass, including also distal volumes, of 184 Gt (DRE) distributes to 84% into 9 dacitic to rhyolitic eruptions and to 16% into 4 basaltic to basaltic–andesitic eruptions. The widely dispersed tephra sheets have up to five times the mass of their parental volcanic edifices and thus represent a significant albeit less obvious component of the arc volcanism. Eruption magnitudes (M = log10(m) − 7 with m the mass in kg), range from M = 4.1 to M = 6.3. Most of the eruptions were dominantly plinian, with eruption columns reaching variably high into the stratosphere, but minor phreatomagmatic phases were also involved. Two phreatomagmatic eruptions, one dacitic and one basaltic–andesitic, produced mostly pyroclastic surges but also fallout from high eruption columns. Comparison of fallout tephra dispersal patterns with present-day, seasonally changing height-dependant wind directions suggests that 8 eruptions occurred during the rainy season while 5 took place during the dry season. The tephra succession documents two major phases of erosion. The first phase, > 17 ka ago, appears to be related to tectonic activity whereas the second phase may have been caused by wet climatic conditions between 2 to 6 ka ago. The Apoyo caldera had two large plinian, caldera-forming eruptions in rapid succession about 24 ka ago and should be considered a silicic volcano with long repose times. Three highly explosive basaltic eruptions were generated at the Masaya Caldera within the last 6 ka. Since then frequent but small eruptions and lava effusion were largely limited to the cal

High precision measurements of 226Ra–230Th–238U disequilibria and Ba concentrations are reported for samples from two chemically zoned trachyte deposits from Fogo volcano, São Miguel, Azores. High-precision U-series disequilibria measurements by plasma ionization multicollector mass spectrometry were performed on pumice lapilli and volcanic glass separates from Fogo 1563 A.D. (∼ 0.14 km3) and the ∼ 4.7 ka Fogo A (∼ 0.7 km3) deposits in order to quantify the time scales of magmatic processes. Observed (226Ra)/Ba–(230Th)/Ba relationships in Fogo 1563 are compatible with a conventional instantaneous fractional crystallization model and a pre-eruptive magma residence time of ∼ 50 y. However, the Fogo A data cannot be explained by instantaneous fractional crystallization, and require a prolonged crystallization history. Continuous differentiation models better explain the observed 226Ra–230Th variations within the Fogo deposits and may be more realistic in general. Such models suggest magma residence times prior to eruption of ∼ 50–80 years for Fogo 1563, and ∼ 4.7 ka for the larger volume Fogo A eruption. These time scales represent liquid residence ages rather than the crystallization ages documented in most previous magmatic time scale studies, and allow constraints to be placed on the time scales necessary for the development of chemical zonation within the Fogo magma chamber. Our results indicate that calculated time scales are relatively insensitive to the precise nature of the continuous differentiation models, and further indicate that meaningful magma differentiation time scales can be obtained despite open system behavior, because the Ra–Th disequilibria are overwhelmingly controlled by feldspar fractionation. Calculated time scales are, however, extremely sensitive to DRa/DBa ratios. We therefore emphasize the crucial importance to better constrain the relative partitioning of Ra and Ba when employing Ra–Th disequilibrium data to constrain the rates and time s

The stratigraphic succession of widespread tephra layers in west-central Nicaragua was emplaced by highly explosive eruptions from mainly three volcanoes: the Chiltepe volcanic complex and the Masaya and Apoyo calderas. Stratigraphic correlations are based on distinct compositions of tephras. The total tephras combine to a total on-shore volume of about 37 km3 produced during the last ∼ 60 ka. The total erupted magma mass, including also distal volumes, of 184 Gt (DRE) distributes to 84% into 9 dacitic to rhyolitic eruptions and to 16% into 4 basaltic to basaltic–andesitic eruptions. The widely dispersed tephra sheets have up to five times the mass of their parental volcanic edifices and thus represent a significant albeit less obvious component of the arc volcanism. Eruption magnitudes (M = log10(m) − 7 with m the mass in kg), range from M = 4.1 to M = 6.3. Most of the eruptions were dominantly plinian, with eruption columns reaching variably high into the stratosphere, but minor phreatomagmatic phases were also involved. Two phreatomagmatic eruptions, one dacitic and one basaltic–andesitic, produced mostly pyroclastic surges but also fallout from high eruption columns. Comparison of fallout tephra dispersal patterns with present-day, seasonally changing height-dependant wind directions suggests that 8 eruptions occurred during the rainy season while 5 took place during the dry season. The tephra succession documents two major phases of erosion. The first phase, > 17 ka ago, appears to be related to tectonic activity whereas the second phase may have been caused by wet climatic conditions between 2 to 6 ka ago. The Apoyo caldera had two large plinian, caldera-forming eruptions in rapid succession about 24 ka ago and should be considered a silicic volcano with long repose times. Three highly explosive basaltic eruptions were generated at the Masaya Caldera within the last 6 ka. Since then frequent but small eruptions and lava effusion were largely limited to the cal

The dacite to andesite zoned Mateare Tephra is the fallout of a predominantly plinian eruption from Chiltepe peninsula at the western shore of Lake Managua that occurred 3000–6000 years ago. It comprises four units: Unit A of high-silica dacite is stratified, ash-rich lapilli fallout generated by unsteady subplinian eruption pulses affected by minor water access to the conduit and conduit blocking by degassed magma. Unit B of less silicic dacite is well sorted, massive pumice lapilli fallout from the main, steady plinian phase of the eruption. Unit C is andesitic fallout that is continuous from unit B except for the rapid change in chemical composition, which had little influence on the ongoing eruption except for a minor transient reduction of the discharge rate and access of water to the conduit. After this, discharge rate re-established to a strong plinian eruption that emplaced the main part of unit C. This was again followed by water access to the conduit which increased through upper unit C. The lithic-rich lapilli to wet ash fallout of unit D is the product of the fully phreatomagmatic terminal phase of the eruption. A massive well-sorted sand layer, the Mateare Sand, replaces laterally variable parts of unit A and lowermost part of unit B in outcrops up to 32 m above present lake level. The corresponding interval missing in the primary fallout can be identified by comparing the composition of pumice entrained in the sand, and pumice from the local base of unit B on top of the sand, with the compositional gradient in undisturbed fallout. The amount of fallout entrained in the sand decreases with distance to the lake. The Mateare Sand occurs at elevations well above beach levels and its widespread continuous distribution defies a fluviatile origin. Instead, it was produced by lake tsunamis triggered by eruption pulses during the initial unsteady phase of activity. Such tsunamis could threaten areas not affected by fallout, and represent a hazard of particular

The dacite to andesite zoned Mateare Tephra is the fallout of a predominantly plinian eruption from Chiltepe peninsula at the western shore of Lake Managua that occurred 3000–6000 years ago. It comprises four units: Unit A of high-silica dacite is stratified, ash-rich lapilli fallout generated by unsteady subplinian eruption pulses affected by minor water access to the conduit and conduit blocking by degassed magma. Unit B of less silicic dacite is well sorted, massive pumice lapilli fallout from the main, steady plinian phase of the eruption. Unit C is andesitic fallout that is continuous from unit B except for the rapid change in chemical composition, which had little influence on the ongoing eruption except for a minor transient reduction of the discharge rate and access of water to the conduit. After this, discharge rate re-established to a strong plinian eruption that emplaced the main part of unit C. This was again followed by water access to the conduit which increased through upper unit C. The lithic-rich lapilli to wet ash fallout of unit D is the product of the fully phreatomagmatic terminal phase of the eruption. A massive well-sorted sand layer, the Mateare Sand, replaces laterally variable parts of unit A and lowermost part of unit B in outcrops up to 32 m above present lake level. The corresponding interval missing in the primary fallout can be identified by comparing the composition of pumice entrained in the sand, and pumice from the local base of unit B on top of the sand, with the compositional gradient in undisturbed fallout. The amount of fallout entrained in the sand decreases with distance to the lake. The Mateare Sand occurs at elevations well above beach levels and its widespread continuous distribution defies a fluviatile origin. Instead, it was produced by lake tsunamis triggered by eruption pulses during the initial unsteady phase of activity. Such tsunamis could threaten areas not affected by fallout, and represent a hazard of particular

The dacite to andesite zoned Mateare Tephra is the fallout of a predominantly plinian eruption from Chiltepe peninsula at the western shore of Lake Managua that occurred 3000–6000 years ago. It comprises four units: Unit A of high-silica dacite is stratified, ash-rich lapilli fallout generated by unsteady subplinian eruption pulses affected by minor water access to the conduit and conduit blocking by degassed magma. Unit B of less silicic dacite is well sorted, massive pumice lapilli fallout from the main, steady plinian phase of the eruption. Unit C is andesitic fallout that is continuous from unit B except for the rapid change in chemical composition, which had little influence on the ongoing eruption except for a minor transient reduction of the discharge rate and access of water to the conduit. After this, discharge rate re-established to a strong plinian eruption that emplaced the main part of unit C. This was again followed by water access to the conduit which increased through upper unit C. The lithic-rich lapilli to wet ash fallout of unit D is the product of the fully phreatomagmatic terminal phase of the eruption. A massive well-sorted sand layer, the Mateare Sand, replaces laterally variable parts of unit A and lowermost part of unit B in outcrops up to 32 m above present lake level. The corresponding interval missing in the primary fallout can be identified by comparing the composition of pumice entrained in the sand, and pumice from the local base of unit B on top of the sand, with the compositional gradient in undisturbed fallout. The amount of fallout entrained in the sand decreases with distance to the lake. The Mateare Sand occurs at elevations well above beach levels and its widespread continuous distribution defies a fluviatile origin. Instead, it was produced by lake tsunamis triggered by eruption pulses during the initial unsteady phase of activity. Such tsunamis could threaten areas not affected by fallout, and represent a hazard of particular

The 1998 block-and-ash flow (BAF) deposits on the western flanks of Merapi were described both immediately following emplacement, and after one and two wet seasons of erosion. This afforded a unique combination of data on deposit morphology and internal structure from proximal to distal reaches. The deposits of the basal avalanches from individual BAF events were classified into three major and two minor facies in channelised reaches, primarily resulting from effects of channel confinement and slope. Termination reaches were classified into three facies, related to preevent topography. The deposit facies architecture was interpreted to understand particle transport and deposition within the framework of granular flow systematics. The transport model we propose for these basal avalanches involves unsteady flow conditions, generated by pulses of collapse as the source, and/or by development of kinetic waves during the granular flow. The coarsest clasts along with low-density material migrated to the wave/pulse fronts within these flows. Whereas, in bulk flow, grain dispersive pressures and kinetic sieving result in coarsest clasts being buoyed to the flow top, with this tendency being greatest on higher slopes (=greater kinetic energy) and with greater flow confinement (maximising particle collisions). Deposition begins as the flow reaches near its full extent. An initial pulse or flow wave freezes rapidly from its base upward and its rear forward as it looses kinetic energy and dispersive pressure, such that an upper higher energy part of the flow is still able to move downstream over the frozen base. Later waves or pulses then accrete progressively horizontally upslope banking up onto or overtopping earlier deposited sediment. Hence, the overall flow freezes from its front to its tail. At termination reaches in unconfined areas, the overflowing upper part of flow gradually deposits from its base upward, resulting in a slightly tapering sheet formed deposit, marked b

The 1998 block-and-ash flow (BAF) deposits on the western flanks of Merapi were described both immediately following emplacement, and after one and two wet seasons of erosion. This afforded a unique combination of data on deposit morphology and internal structure from proximal to distal reaches. The deposits of the basal avalanches from individual BAF events were classified into three major and two minor facies in channelised reaches, primarily resulting from effects of channel confinement and slope. Termination reaches were classified into three facies, related to preevent topography. The deposit facies architecture was interpreted to understand particle transport and deposition within the framework of granular flow systematics. The transport model we propose for these basal avalanches involves unsteady flow conditions, generated by pulses of collapse as the source, and/or by development of kinetic waves during the granular flow. The coarsest clasts along with low-density material migrated to the wave/pulse fronts within these flows. Whereas, in bulk flow, grain dispersive pressures and kinetic sieving result in coarsest clasts being buoyed to the flow top, with this tendency being greatest on higher slopes (=greater kinetic energy) and with greater flow confinement (maximising particle collisions). Deposition begins as the flow reaches near its full extent. An initial pulse or flow wave freezes rapidly from its base upward and its rear forward as it looses kinetic energy and dispersive pressure, such that an upper higher energy part of the flow is still able to move downstream over the frozen base. Later waves or pulses then accrete progressively horizontally upslope banking up onto or overtopping earlier deposited sediment. Hence, the overall flow freezes from its front to its tail. At termination reaches in unconfined areas, the overflowing upper part of flow gradually deposits from its base upward, resulting in a slightly tapering sheet formed deposit, marked b

The 1998 block-and-ash flow (BAF) deposits on the western flanks of Merapi were described both immediately following emplacement, and after one and two wet seasons of erosion. This afforded a unique combination of data on deposit morphology and internal structure from proximal to distal reaches. The deposits of the basal avalanches from individual BAF events were classified into three major and two minor facies in channelised reaches, primarily resulting from effects of channel confinement and slope. Termination reaches were classified into three facies, related to preevent topography. The deposit facies architecture was interpreted to understand particle transport and deposition within the framework of granular flow systematics. The transport model we propose for these basal avalanches involves unsteady flow conditions, generated by pulses of collapse as the source, and/or by development of kinetic waves during the granular flow. The coarsest clasts along with low-density material migrated to the wave/pulse fronts within these flows. Whereas, in bulk flow, grain dispersive pressures and kinetic sieving result in coarsest clasts being buoyed to the flow top, with this tendency being greatest on higher slopes (=greater kinetic energy) and with greater flow confinement (maximising particle collisions). Deposition begins as the flow reaches near its full extent. An initial pulse or flow wave freezes rapidly from its base upward and its rear forward as it looses kinetic energy and dispersive pressure, such that an upper higher energy part of the flow is still able to move downstream over the frozen base. Later waves or pulses then accrete progressively horizontally upslope banking up onto or overtopping earlier deposited sediment. Hence, the overall flow freezes from its front to its tail. At termination reaches in unconfined areas, the overflowing upper part of flow gradually deposits from its base upward, resulting in a slightly tapering sheet formed deposit, marked b

Laacher See volcano erupted about 12,900 years ago, resulting in a tephra blanket widely dispersed in central Europe. The eruption tapped a chemically and mineralogically strongly zoned phonolitic magma reservoir. In order to better constrain the pre-eruptive storage of that magma, hydrothermal phase equilibria experiments have been performed on a powdered sample from the highly evolved phonolite of the Lower Laacher See Tephra (LLST). Experiments were run at water pressures of 75–175 MPa and temperatures of 725–800 °C. Stability fields of sanidine and amphibole, the most common crystal phases in LLST, suggest that the magma last equilibrated at temperatures of 750–760 °C and pressures of 115–145 MPa. Those conditions suggest the magma should have contained between 5 and 6 wt.% water, which agrees well with water contents of 5.7 wt.% found in glass inclusions. That agreement, along with the observation that Cl and S vary with the composition of the inclusions, suggest that the Laacher See magma was saturated at depth with a water-rich fluid, and thus was stored at pressures near those estimated from our experiments. Such pressures roughly correspond to depths of 5–6 km. Clinopyroxene and plagioclase did not occur in the experimental runs, although present in the natural pumice sample. It seems likely that these phases are not inherent to the LLST. They were formed in other parts of the chemically stratified column and later redistributed. A recalculation of the magma density, together with a new estimate of the temperature and pressure range suggests a steep density gradient and a thin withdrawal layer, which is typical for volatile-enriched magma bodies and confirms earlier models of convective fractionation in the Laacher See magma reservoir.

Laacher See volcano erupted about 12,900 years ago, resulting in a tephra blanket widely dispersed in central Europe. The eruption tapped a chemically and mineralogically strongly zoned phonolitic magma reservoir. In order to better constrain the pre-eruptive storage of that magma, hydrothermal phase equilibria experiments have been performed on a powdered sample from the highly evolved phonolite of the Lower Laacher See Tephra (LLST). Experiments were run at water pressures of 75–175 MPa and temperatures of 725–800 °C. Stability fields of sanidine and amphibole, the most common crystal phases in LLST, suggest that the magma last equilibrated at temperatures of 750–760 °C and pressures of 115–145 MPa. Those conditions suggest the magma should have contained between 5 and 6 wt.% water, which agrees well with water contents of 5.7 wt.% found in glass inclusions. That agreement, along with the observation that Cl and S vary with the composition of the inclusions, suggest that the Laacher See magma was saturated at depth with a water-rich fluid, and thus was stored at pressures near those estimated from our experiments. Such pressures roughly correspond to depths of 5–6 km. Clinopyroxene and plagioclase did not occur in the experimental runs, although present in the natural pumice sample. It seems likely that these phases are not inherent to the LLST. They were formed in other parts of the chemically stratified column and later redistributed. A recalculation of the magma density, together with a new estimate of the temperature and pressure range suggests a steep density gradient and a thin withdrawal layer, which is typical for volatile-enriched magma bodies and confirms earlier models of convective fractionation in the Laacher See magma reservoir.

Laacher See volcano erupted about 12,900 years ago, resulting in a tephra blanket widely dispersed in central Europe. The eruption tapped a chemically and mineralogically strongly zoned phonolitic magma reservoir. In order to better constrain the pre-eruptive storage of that magma, hydrothermal phase equilibria experiments have been performed on a powdered sample from the highly evolved phonolite of the Lower Laacher See Tephra (LLST). Experiments were run at water pressures of 75–175 MPa and temperatures of 725–800 °C. Stability fields of sanidine and amphibole, the most common crystal phases in LLST, suggest that the magma last equilibrated at temperatures of 750–760 °C and pressures of 115–145 MPa. Those conditions suggest the magma should have contained between 5 and 6 wt.% water, which agrees well with water contents of 5.7 wt.% found in glass inclusions. That agreement, along with the observation that Cl and S vary with the composition of the inclusions, suggest that the Laacher See magma was saturated at depth with a water-rich fluid, and thus was stored at pressures near those estimated from our experiments. Such pressures roughly correspond to depths of 5–6 km. Clinopyroxene and plagioclase did not occur in the experimental runs, although present in the natural pumice sample. It seems likely that these phases are not inherent to the LLST. They were formed in other parts of the chemically stratified column and later redistributed. A recalculation of the magma density, together with a new estimate of the temperature and pressure range suggests a steep density gradient and a thin withdrawal layer, which is typical for volatile-enriched magma bodies and confirms earlier models of convective fractionation in the Laacher See magma reservoir.

This paper reviews major results of the recent German research priority program “Changes of the geo-biosphere during the last 15,000 years”, a contribution to PAGES focussing on changes of the geo-biosphere during the Weichselian Lateglacial. Different continental archives such as annually laminated lacustrine sediments, floodplain sediments, and speleothems were used to reconstruct environmental response to climatic changes and the Laacher See eruption event at ca 12,900 cal BP. Special emphasis is paid to establish a reliable time control using varve counting, high-precision radiocarbon dating of tree-ring series, and AMS radiocarbon dating of terrestrial plant macrofossils recovered from lacustrine sediments to correlate and synchronize large-scale environmental changes and events in central Europe.

This paper reviews major results of the recent German research priority program “Changes of the geo-biosphere during the last 15,000 years”, a contribution to PAGES focussing on changes of the geo-biosphere during the Weichselian Lateglacial. Different continental archives such as annually laminated lacustrine sediments, floodplain sediments, and speleothems were used to reconstruct environmental response to climatic changes and the Laacher See eruption event at ca 12,900 cal BP. Special emphasis is paid to establish a reliable time control using varve counting, high-precision radiocarbon dating of tree-ring series, and AMS radiocarbon dating of terrestrial plant macrofossils recovered from lacustrine sediments to correlate and synchronize large-scale environmental changes and events in central Europe.

This paper reviews major results of the recent German research priority program “Changes of the geo-biosphere during the last 15,000 years”, a contribution to PAGES focussing on changes of the geo-biosphere during the Weichselian Lateglacial. Different continental archives such as annually laminated lacustrine sediments, floodplain sediments, and speleothems were used to reconstruct environmental response to climatic changes and the Laacher See eruption event at ca 12,900 cal BP. Special emphasis is paid to establish a reliable time control using varve counting, high-precision radiocarbon dating of tree-ring series, and AMS radiocarbon dating of terrestrial plant macrofossils recovered from lacustrine sediments to correlate and synchronize large-scale environmental changes and events in central Europe.

Seismic P-wave travel times collected during METEOR cruise M24 are inverted to derive a three-dimensional model of the P-wave velocity structure of the northern part of Gran Canaria, Canary Islands. The data consist of 6689 P-wave travel times from 1487 offshore air-gun shots which were recorded by both land-based seismometers and ocean bottom hydrophones. The crustal structure is well imaged by the data set as demonstrated by analysis of the resolution and tests with synthetic data. The volcanic island is characterized by generally high P-wave velocities (>5.5 km/s) and a heterogeneous structure with large lateral velocity variations. High P-wave velocities are found around the centers of the Miocene shield volcanoes in the vicinity of Agüimes, San Nicolas, and Agaete as well as the center of the Pliocene Roque Nublo volcano. The velocity structure suggests a high percentage of dense intrusive rocks. Some of the intrusive rocks were emplaced during the eruption of >1000 km3 of Miocene felsic magmas following the basaltic shield phase. The velocity structure beneath La Isleta peninsula and its submarine continuation is interpreted as a volcanic rift zone with abundant dikes. The velocities decrease to <5 km/s north of the coastline. A high velocity zone thinning away from the central edifice is interpreted as the massive island flank extending up to 50 km off the coast which is underlain by prevolcanic Neogene–Jurassic sediments. The igneous part of the oceanic crust exhibits an anomalous structure with a relatively small thickness (∼3 km) layer 3 and a 2–4-km-thick layer 2, probably reflecting a modification of the crust due to long-lasting magmatic intrusive activity during the evolution of the Canary Islands. The Moho north of Gran Canaria is found at a depth of ∼15 km. The structure of Gran Canaria and the adjacent ocean basin is thought to be the result of a diffuse mantle upwelling under a slowly moving plate.

Seismic P-wave travel times collected during METEOR cruise M24 are inverted to derive a three-dimensional model of the P-wave velocity structure of the northern part of Gran Canaria, Canary Islands. The data consist of 6689 P-wave travel times from 1487 offshore air-gun shots which were recorded by both land-based seismometers and ocean bottom hydrophones. The crustal structure is well imaged by the data set as demonstrated by analysis of the resolution and tests with synthetic data. The volcanic island is characterized by generally high P-wave velocities (>5.5 km/s) and a heterogeneous structure with large lateral velocity variations. High P-wave velocities are found around the centers of the Miocene shield volcanoes in the vicinity of Agüimes, San Nicolas, and Agaete as well as the center of the Pliocene Roque Nublo volcano. The velocity structure suggests a high percentage of dense intrusive rocks. Some of the intrusive rocks were emplaced during the eruption of >1000 km3 of Miocene felsic magmas following the basaltic shield phase. The velocity structure beneath La Isleta peninsula and its submarine continuation is interpreted as a volcanic rift zone with abundant dikes. The velocities decrease to <5 km/s north of the coastline. A high velocity zone thinning away from the central edifice is interpreted as the massive island flank extending up to 50 km off the coast which is underlain by prevolcanic Neogene–Jurassic sediments. The igneous part of the oceanic crust exhibits an anomalous structure with a relatively small thickness (∼3 km) layer 3 and a 2–4-km-thick layer 2, probably reflecting a modification of the crust due to long-lasting magmatic intrusive activity during the evolution of the Canary Islands. The Moho north of Gran Canaria is found at a depth of ∼15 km. The structure of Gran Canaria and the adjacent ocean basin is thought to be the result of a diffuse mantle upwelling under a slowly moving plate.

Seismic P-wave travel times collected during METEOR cruise M24 are inverted to derive a three-dimensional model of the P-wave velocity structure of the northern part of Gran Canaria, Canary Islands. The data consist of 6689 P-wave travel times from 1487 offshore air-gun shots which were recorded by both land-based seismometers and ocean bottom hydrophones. The crustal structure is well imaged by the data set as demonstrated by analysis of the resolution and tests with synthetic data. The volcanic island is characterized by generally high P-wave velocities (>5.5 km/s) and a heterogeneous structure with large lateral velocity variations. High P-wave velocities are found around the centers of the Miocene shield volcanoes in the vicinity of Agüimes, San Nicolas, and Agaete as well as the center of the Pliocene Roque Nublo volcano. The velocity structure suggests a high percentage of dense intrusive rocks. Some of the intrusive rocks were emplaced during the eruption of >1000 km3 of Miocene felsic magmas following the basaltic shield phase. The velocity structure beneath La Isleta peninsula and its submarine continuation is interpreted as a volcanic rift zone with abundant dikes. The velocities decrease to <5 km/s north of the coastline. A high velocity zone thinning away from the central edifice is interpreted as the massive island flank extending up to 50 km off the coast which is underlain by prevolcanic Neogene–Jurassic sediments. The igneous part of the oceanic crust exhibits an anomalous structure with a relatively small thickness (∼3 km) layer 3 and a 2–4-km-thick layer 2, probably reflecting a modification of the crust due to long-lasting magmatic intrusive activity during the evolution of the Canary Islands. The Moho north of Gran Canaria is found at a depth of ∼15 km. The structure of Gran Canaria and the adjacent ocean basin is thought to be the result of a diffuse mantle upwelling under a slowly moving plate.

We compiled a global data set of volcanic degassing during both explosive and quiescent volcanic events. The data set comprises estimates of gas emissions of volcanoes from Europe (e.g. Etna), Asia (e.g. Merapi), the Americas (e.g. Fuego), Africa (e.g. Erta Ale) and ocean islands (e.g. Kilauea) over the past 100 yr. The set includes 50 monitored volcanoes and ∼310 extrapolated explosively erupting volcanoes. Among the ∼360 volcanoes, 75% are located in the Northern and 25% in the Southern Hemisphere. We have estimated the total annual global volcanic sulfur emission into the atmosphere to be on the order of 7.5–10.5×1012 g/yr S (here as SO2), amounting to 10–15% of the annual anthropogenic sulfur output (∼70×1012 g/yr S during the decade 1981–1990) and 7.5–10.5% of the total global sulfur emission (e.g. biomass burning, anthropogenic, dimethylsulfide) with ∼100×1012 g/yr S. The estimates of other volcanic gases emitted (e.g. H2S, HCl) are based on the assumption that the different gas components emitted by a volcano are in equilibrium with each other. Accordingly, the molar ratios of the gas species in high-temperature fumaroles are similar to molar ratios equilibrated at depth where the gas separates from the magma. Thus, we can use the directly measured SO2 fluxes and known molar ratios (e.g. H2S/SO2) for a semi-quantitative estimate of other gas components emitted (e.g. H2S). The total annual emission of HCl is 1.2–170×1012 g/yr, that of H2S 1.5–37.1×1012 g/yr, of HF 0.7–8.6×1012 g/yr, of HBr 2.6–43.2×109 g/yr, and of OCS 9.4×107–3.2×1011 g/yr. We estimate an emission of 1.3×107–4.4×1010 g/yr for CS2.

We compiled a global data set of volcanic degassing during both explosive and quiescent volcanic events. The data set comprises estimates of gas emissions of volcanoes from Europe (e.g. Etna), Asia (e.g. Merapi), the Americas (e.g. Fuego), Africa (e.g. Erta Ale) and ocean islands (e.g. Kilauea) over the past 100 yr. The set includes 50 monitored volcanoes and ∼310 extrapolated explosively erupting volcanoes. Among the ∼360 volcanoes, 75% are located in the Northern and 25% in the Southern Hemisphere. We have estimated the total annual global volcanic sulfur emission into the atmosphere to be on the order of 7.5–10.5×1012 g/yr S (here as SO2), amounting to 10–15% of the annual anthropogenic sulfur output (∼70×1012 g/yr S during the decade 1981–1990) and 7.5–10.5% of the total global sulfur emission (e.g. biomass burning, anthropogenic, dimethylsulfide) with ∼100×1012 g/yr S. The estimates of other volcanic gases emitted (e.g. H2S, HCl) are based on the assumption that the different gas components emitted by a volcano are in equilibrium with each other. Accordingly, the molar ratios of the gas species in high-temperature fumaroles are similar to molar ratios equilibrated at depth where the gas separates from the magma. Thus, we can use the directly measured SO2 fluxes and known molar ratios (e.g. H2S/SO2) for a semi-quantitative estimate of other gas components emitted (e.g. H2S). The total annual emission of HCl is 1.2–170×1012 g/yr, that of H2S 1.5–37.1×1012 g/yr, of HF 0.7–8.6×1012 g/yr, of HBr 2.6–43.2×109 g/yr, and of OCS 9.4×107–3.2×1011 g/yr. We estimate an emission of 1.3×107–4.4×1010 g/yr for CS2.

The compositionally zoned San Juan eruption on La Palma emanated from three eruptive centers located along a north–south-trending rift zone in the south of the island. Seismic precursors began weakly in 1936 and became strong in March 1949, with their foci progressing from the north of the rift zone towards its south. This suggests that magma ascended beneath the old Taburiente shield volcano and moved southward along the rift. The eruption began on June 24, 1949, with phreatomagmatic activity at Duraznero crater on the ridgetop (ca. 1880 m above sea level), where five vents erupted tephritic lava along a 400-m-long fissure. On June 8, the Duraznero vents shut down abruptly, and the activity shifted to an off-rift fissure at Llano del Banco, located at ca. 550 m lower elevation and 3 km to the northwest. This eruptive center issued initially tephritic aa and later basanitic pahoehoe lava at high rates, producing a lava flow that entered the sea. Two days after basanite began to erupt at Llano del Banco, Hoyo Negro crater (ca. 1880 m asl), located 700 m north of Duraznero along the rift, opened on July 12 and produced ash and bombs of basanitic to phonotephritic composition in violent phreatomagmatic explosions (White and Schmincke, 1999). Llano del Banco and Hoyo Negro were simultaneously active for 11 days and showed a co-variance of their eruption rates indicating a shallow hydraulic connection. On July 30, after 3 days of quiescence at all vents, Duraznero and Hoyo Negro became active again during a final eruptive phase. Duraznero issued basanitic lava at high rates for 12 h and produced a lava flow that descended towards the east coast. The lava contains ca. 1 vol.% crustal and mantle xenoliths consisting of 40% tholeiitic gabbros from the oceanic crust, 35% alkaline gabbros, and 20% ultramafic cumulates. The occurrence of xenoliths almost exclusively in the final lava is consistent with their origin by wall-rock collapse at depth near the end of the eruption. T

The compositionally zoned San Juan eruption on La Palma emanated from three eruptive centers located along a north–south-trending rift zone in the south of the island. Seismic precursors began weakly in 1936 and became strong in March 1949, with their foci progressing from the north of the rift zone towards its south. This suggests that magma ascended beneath the old Taburiente shield volcano and moved southward along the rift. The eruption began on June 24, 1949, with phreatomagmatic activity at Duraznero crater on the ridgetop (ca. 1880 m above sea level), where five vents erupted tephritic lava along a 400-m-long fissure. On June 8, the Duraznero vents shut down abruptly, and the activity shifted to an off-rift fissure at Llano del Banco, located at ca. 550 m lower elevation and 3 km to the northwest. This eruptive center issued initially tephritic aa and later basanitic pahoehoe lava at high rates, producing a lava flow that entered the sea. Two days after basanite began to erupt at Llano del Banco, Hoyo Negro crater (ca. 1880 m asl), located 700 m north of Duraznero along the rift, opened on July 12 and produced ash and bombs of basanitic to phonotephritic composition in violent phreatomagmatic explosions (White and Schmincke, 1999). Llano del Banco and Hoyo Negro were simultaneously active for 11 days and showed a co-variance of their eruption rates indicating a shallow hydraulic connection. On July 30, after 3 days of quiescence at all vents, Duraznero and Hoyo Negro became active again during a final eruptive phase. Duraznero issued basanitic lava at high rates for 12 h and produced a lava flow that descended towards the east coast. The lava contains ca. 1 vol.% crustal and mantle xenoliths consisting of 40% tholeiitic gabbros from the oceanic crust, 35% alkaline gabbros, and 20% ultramafic cumulates. The occurrence of xenoliths almost exclusively in the final lava is consistent with their origin by wall-rock collapse at depth near the end of the eruption. T

Gabbro and minor metabasalt fragments of MORB composition were found on three of the seven Canary Islands. On Gran Canaria, they occur as metamorphosed (greenschist facies) metabasalt and metagabbro clasts in Miocene fanglomerates and sandstones overlying the shield basalts. On Lanzarote and La Palma, MORB gabbros occur as xenoliths in Pleistocene and historic basanite scoria cones and lava flows. The MORB xenoliths are interpreted as fragments of layers 2 and 3 of the underlying Mesozoic oceanic crust, based on mineral compositions (An-rich plagioclase, Ti- and Al-poor clinopyroxene, ± orthopyroxene ± olivine), depleted major and trace element signatures, and Jurassic ages (ca. 180 Ma) determined on single primary plagioclase and secondary amphibole crystals using the 40Ar/39Ar laser technique. The Lanzarote gabbros are very mafic (mg# 87 to 89 in clinopyroxene), moderately deformed, and highly depleted. Gran Canaria gabbros are more evolved (mg# 69 to 83 in clinopyroxene) and texturally mostly isotropic. La Palma MORB gabbros have a range of compositions (mg# 68 to 83 in clinopyroxene), some rocks being strongly metasomatized by interaction with basanite magma. The occurrence of MORB fragments on Lanzarote provides definite evidence that oceanic crust beneath the Canary Island archipelago continues at least as far east as the eastern Canary Islands. We postulate that MORB gabbros on Lanzarote which are commonly associated with peridotite xenoliths, represent the base of oceanic layer 3 where gabbros and peridotites were possibly tectonically interleaved. Such tectonic mixing would explain the enigmatic seismic velocities in this area. Gabbro xenoliths from La Palma were derived from within layer 3, probably from wall rock close to magma reservoirs emplaced during the Pleistocene/Holocene growth of La Palma. The Gran Canaria xenoliths are interpreted to represent the metamorphosed layer 2 and upper layer 3. The abundance of lower crustal xenoliths emphasizes the im