ALBERT

Description: High-resolution sampling in monogenetic fields has the potential to reveal fine-scale heterogeneity of the mantle, a feature that may be overwhelmed by larger fluxes of magma, or missed by under-sampling. The Quaternary Auckland Volcanic Field (AVF) in northern New Zealand is a basaltic field of 51 small-volume volcanic centres, and is one of the best-sampled examples of a monogenetic volcanic field. We present data for 12 centres in the volcanic field. These show the large compositional variations between volcanoes as well as through single eruptive sequences. Whole-rock compositions range from subalkaline basalt in the larger centres, through alkali basalt to nephelinite in the smallest centres. Fractional crystallization has had a limited effect in many of the centres, but high-pressure clinopyroxene crystallization may have occurred in others. Three end-members are observed in Pb isotope space, indicating that distinct mantle source components are involved in the petrogenesis of the magmas. Whole-rock multi-element patterns show that the larger centres have prominent positive Sr anomalies and lack K anomalies, whereas the smaller centres have prominent negative K anomalies and lack Sr anomalies. The melting parameters and compositions of the sources involved are modelled using trace element ratios and multi-element patterns, and three components are characterized: (1) fertile peridotite with a Pb-isotope composition similar to Pacific mid-ocean ridge basalt; (2) eclogite domains with a HIMU-like isotope composition dispersed within the fertile peridotite; (3) slightly depleted subduction-metasomatized peridotitic lithospheric mantle (containing c . 3% subduction fluids). Modelling shows that melting in the AVF begins in garnet-bearing fertile asthenosphere (with preferential melting of eclogite domains) and that melts are variably diluted by melts of the lithospheric source. The U–Th isotope compositions of the end-members in the AVF show 230 Th excess [( 230 Th/ 232 Th) ratios of 1·11–1·38], with the samples of lower ( 230 Th/ 232 Th) exhibiting higher ( 238 U/ 232 Th), which we attribute to the dilution effect of the melts from the lithospheric mantle source. Modelling reveals a correlation between melting in the asthenosphere, the degree of melting and incorporation of the metasomatized lithospheric mantle source, and the resultant size of the volcanic centre. This suggests that the scale of the eruption may essentially be controlled by asthenospheric mantle dynamics.

Description: Jeju is a volcanic field that has erupted from around 1·8 Myr to c. 1 kyr ago. Activity began with dispersed, basaltic, monogenetic, phreatomagmatic eruptions. Continuing monogenetic volcanism was later joined by more voluminous lava effusion events building a central composite shield. Samples from older (〉0·7 Ma) and younger (〈0·2 Ma) monogenetic centres were analysed for their whole-rock major element, trace element and Sr–Nd–Pb isotopic compositions. Pyroclastic products from the monogenetic centres are dominantly alkali basalt to trachybasalt, whereas the more voluminous lava flows and domes of the central edifice consist of subalkali basalt and alkali basalt to trachyte. Lavas from the Early Pleistocene monogenetic centres are depleted in MgO, Cr and Ni, reflecting considerable olivine fractionation. By contrast, Late Pleistocene–Holocene monogenetic centre magmas fractionated clinopyroxene + olivine at deeper levels. Isotopic compositions show little variation across the suite; however, the Late Pleistocene–Holocene monogenetic centres have generally lower 87 Sr/ 86 Sr and 208 Pb/ 204 Pb and higher 143 Nd/ 144 Nd than the older centres and subalkali lavas. Major and trace element and isotope data suggest a common, shallower source for the high-Al alkali and subalkali lavas, in contrast to a deeper source for the low-Al alkali magmas. We propose that mantle melting was initiated under partially hydrous conditions at a pressure of near 2·5 GPa, followed by drier conditions and extension of the melting zone to 3–3·5 GPa, with a concomitant increase in the volume of melt derived from the shallower part of the system to produce subalkaline magmas. Increasing melt production at shallow depths may be related to accelerated heat transfer resulting from deepening of the melting zone, or increased mantle upwelling. Mantle lenses were uplifted, probably lubricated by shear zones created during the opening of the Sea of Japan c. 15 Myr ago, and reactivated during rotation of the Philippine Sea plate direction of subduction at around 2 Ma. This is the first hypothesized link between subduction processes and intraplate volcanism at Jeju.

Description: Tofua volcano is situated midway along the Tonga oceanic arc and has undergone two phases of ignimbrite-forming activity. The eruptive products are almost entirely basaltic andesites (52·5–57 wt % SiO 2 ) with the exception of a volumetrically minor pre-caldera dacite. The suite displays a strong tholeiitic trend with K 2 O 〈1 wt %. Phenocryst assemblages typically comprise plagioclase + clinopyroxene ± orthopyroxene with microlites of Ti-magnetite. Olivine (Fo 83 – 88 ) is rare and believed to be dominantly antecrystic. An increase in the extent and frequency of reverse zoning in phenocrysts, sieve-textured plagioclase and the occurrence of antecrystic phases in post-caldera lavas record a shift to dynamic conditions, allowing the interaction of magma batches that were previously distinct. Pyroxene thermobarometry suggests crystallization at 950–1200°C and 0·8–1·8 kbar. Volatile measurements of glassy melt inclusions indicate a maximum H 2 O content of 4·16 wt % H 2 O, and CO 2 –H 2 O saturation curves indicate that crystallization occurred at two levels, at depths of 4–5·5 km and 1·5–2·5 km. Major and trace element models suggest that the compositions of the majority of the samples represent a differentiation trend whereby the dacite was produced by 65% fractional crystallization of the most primitive basaltic andesite. Trace element models suggest that the sub-arc mantle source is the residuum of depleted Indian mid-ocean ridge basalt mantle (IDMM-1% melt), whereas radiogenic isotope data imply addition of 0·2% average Tongan sediment melt and a fluid component derived from the subducted altered Pacific oceanic crust. A horizontal array on the U–Th equiline diagram and Ra excesses of up to 500% suggest fluid addition to the mantle wedge within the last few thousand years. Time-integrated ( 226 Ra/ 230 Th) vs Sr/Th and Ba/Th fractionation models imply differentiation timescales of up to 4500 years for the dacitic magma compositions at Tofua.

Description: Ulleung Island is the top of a 3000 m (from sea floor) intraplate alkalic volcanic edifice in the East Sea/Sea of Japan. The emergent 950 m consist of a basaltic lava and agglomerate succession (Stage 1, 1·37–0·97 Ma), intruded and overlain by a sequence of trachytic lavas and domes, which erupted in two episodes (Stage 2, 0·83–0·77 Ma; Stage 3, 0·73–0·24 Ma). The youngest eruptions, post 20 ka bp, were explosive, generating thick tephra sequences of phonolitic composition (Stage 4), which also entrained phaneritic, porphyritic and cumulate accidental lithics. Major element chemistry of the evolved products shows a continuous spectrum of trachyte to phonolite compositions, but these have discordant trace element trends and distinct isotopic characteristics, excluding a direct genetic relationship between the two end-members. Despite this, the Stage 3 trachytes and some porphyritic accidental lithics have chemical characteristics transitional between Stage 2 trachytes and Stage 4 phonolites. Within the phonolitic Stage 4 tephras three subgroups can be distinguished. The oldest, Tephra 5, is considerably enriched in incompatible elements and chondrite-normalized rare earth element (REE) patterns display negative Eu anomalies. The later tephras, Tephras 4–2, have compositions intermediate between the early units and the trachyte samples, and their REE patterns do not have significant Eu anomalies. The last erupted, Tephra 1, from a small intra-caldera structure, has a distinct tephriphonolite composition. Trace element and isotopic chemistry as well as textural characteristics suggest a genetic relationship between the phaneritic lithics and their host phonolitic pumices. The Stage 4 tephras are not related to earlier phases of basaltic to trachytic magmatism (Stages 1–3). They have distinct isotopic compositions and cannot be reliably modelled by fractional crystallization processes. The differences between the explosive phonolitic (Stage 4) and effusive trachytic (Stage 2–3) eruptions are mainly due to different pre-eruptive pressures and temperatures, causing closed- versus open-system degassing. Based on thermodynamic and thermobarometric modelling, the phonolites were derived from deeper (subcrustal) magma storage and rose quickly, with volatiles trapped until eruption. By contrast, the trachytes were stored at shallower crustal levels for longer periods, allowing open-system volatile exsolution and degassing before eruption.

Description: Taranaki (Mt. Egmont) in the western North Island of New Zealand is a high-K andesite volcano with an eruptive history extending over more than 200 kyr. In general, petrological research has concentrated on the post-10 ka record of the modern edifice. This study focuses on the earlier history, which is recorded in 11 major pre-7 ka debris avalanche deposits. Each of these formed as a result of a catastrophic collapse of the edifice of the time. The clast assemblages of these deposits provide insights into the chemical compositions of magmas erupted during the earlier stages of activity of the volcano and form the basis for a new chemo-stratigraphic analysis of the pre-10 ka volcanic succession. Sample suites from the studied debris avalanche deposits show a progressive enrichment in K 2 O and large ion lithophile elements (LILE), reflecting a gradual evolution to high-K andesite. The early magmatic system (pre-100 ka) produced a wide range of compositions including relatively primitive basalts and basaltic andesites. These rocks contain phenocryst assemblages that indicate crystallization within the lower crust or mantle, including a broad range of clinopyroxene compositions, high-Al 2 O 3 hornblende, olivine and phlogopite. A higher proportion of high-silica compositions in the younger sample suites and the appearance of late-stage, low-pressure mineral phases, such as high-TiO 2 hornblende, biotite and Fe-rich orthopyroxene, reflect a gradual shift to more evolved magmas with time. These new data are interpreted to reflect a multi-stage origin for Taranaki andesites. Parental magmas were generated within a lower crustal ‘hot zone’, which formed as a result of repeated intrusions of primitive melts into the lower crust. The geochemical and mineralogical evidence indicates that prior to 100 ka this zone was relatively thin and cold, so that primitive magmas were able to rise rapidly through the crust without significant interaction and modification. As the hot zone evolved, larger proportions of intruded and underplated mafic material were partially remelted, and interaction of these melts with fractionating mantle-derived magmas generated progressively more K- and LILE-enriched compositions. A complex and dispersed magma assembly and storage system developed in the upper crust where the hot-zone melts were further modified by fractional crystallization and magma mixing and mingling.

Description: Debris avalanches caused by the collapse of volcanic flanks pose a great risk to inhabited areas and may permanently change the surrounding landscape and its drainage systems. In this research, we explored the interplay between a debris avalanche and a tectonically uplifting surrounding landscape, providing insights into the long-term consequences of volcanic edifice failures. Exposures of coarse volcaniclastic sediments along the Hautapu River ~50 km southeast of Mount Ruapehu, New Zealand, show evidence of the largest known collapse event of the stratovolcano, which was followed by a vigorous regrowth phase that produced numerous pyroclastic eruptions and pumice-rich lahars. Similar diamicton deposits are exposed within the river catchment adjacent to the west. Cover-bed stratigraphy and geochemical correlation of andesitic lava blocks within the debris-avalanche deposit with dated lavas exposed on the cone indicate that deposition occurred between 125 and 150 ka. The collapse took place during the shift from a glacial to an interglacial climate, when glaciers on the cone were in retreat, and high pore-water pressures combined with deep hydrothermal alteration weakened the cone. In addition, collapse may have been accompanied by magmatic unrest. The ~2–3 km 3 debris avalanche inundated an area of 〉260 km 2 and entered the proto-Hautapu catchment, where it was channelized within the deeply entrenched valley. Mass-wasting events associated with postcollapse volcanism continued to be channeled into the proto–Hautapu River for another ~10 k.y., producing long-runout lahars. Subsequently, the river catchment was isolated from the volcano by incision of the intervening Whangaehu River into the proximal volcaniclastic sediments, accompanied by regional faulting and graben deepening around Ruapehu. At present, the volcaniclastic deposits form a distinctive plateau on the highest topographic elevation within the Hautapu Valley, forming a reversed topography caused by preferred incision of the Hautapu River into softer Late Tertiary sediments concurrent with constant uplift.

Description: An unusual andesitic suite from the Miocene volcanic arc in Northland, New Zealand, comprises pyroxene andesite and garnet-bearing hornblende–pyroxene, hornblende and biotite–hornblende andesites. Garnet crystals occur as 1–10 mm single crystals or more commonly as two or more annealed crystals and as garnetite lenses. The andesitic rocks also contain enclaves of high-MgO pyroxenite, hornblendite, and pyroxene–hornblende gabbro as well as high-Al 2 O 3 hornblende gabbro, garnet–hornblende gabbro, and anorthosite. Garnet crystals in the andesitic volcanic rocks and in the enclaves show comparable compositional ranges, zoning patterns and inclusions, which indicate that they share a common petrogenetic history. They can be grouped into four distinct types on the basis of mode of occurrence, chemical composition and zoning patterns, which leads to their interpretation as antecrysts rather than orthocrysts. The compositions of the garnets, as well as their included mineral assemblages, reflect a petrogenetic trend from high-temperature pyroxene-bearing high-Mg garnet to low-temperature Fe-rich garnet at relatively constant pressure. Well-preserved zoning patterns, in particular those of the Ca- and Mg-rich garnets, reflect processes within a deep crustal arc environment. Later assimilation is suggested by some zoning patterns that show decreasing Ca and increasing Fe and Mn contents. The garnets are interpreted as being derived by disintegration of discrete but closely related cumulate material that formed at pressures of 8–10 kbar. The host volcanic rocks and their garnet crystals together with the enclaves thus represent a consanguineous mixture of liquid and solid components that developed where subduction-related magmas ponded and interacted at or near the base of the crust. Together they represent a rare snapshot of the processes and components that produce arc-type rocks.

Description: Ruapehu, New Zealand’s largest active andesite volcano, is located at the southern tip of the Taupo Volcanic Zone (TVZ), the main locus of subduction-related volcanism in the North Island. Geophysical data indicate that crustal thickness increases from 〈25 km within the TVZ to 40 km beneath Ruapehu. The volcano is built on a basement of Mesozoic meta-greywacke, and geophysical evidence together with xenoliths contained in lavas indicates that this is underlain by oceanic, meta-igneous lower crust. The present-day Ruapehu edifice has been constructed by a series of eruptive events that produced a succession of lava flow-dominated stratigraphic units. In order from oldest to youngest, these are the Te Herenga (250–180 ka), Wahianoa (160–115 ka), Mangawhero (55–45 ka and 20–30 ka), and Whakapapa (15–2 ka) Formations. The dominant rock types are plagioclase- and pyroxene-phyric basaltic andesite and andesite. Dacite also occurs but only one basalt flow has been identified. There have been progressive changes in the minor and trace element chemistry and isotopic composition of Ruapehu eruptive rocks over time. In comparison with rocks from younger formations, Te Herenga eruptive rocks have lower K 2 O abundances and a relatively restricted range in major and trace element and Nd–Sr isotopic composition. Post-Te Herenga andesites and dacites define a Sr–Nd isotopic array that overlaps with the field for TVZ rhyolites and basalts, but Te Herenga Formation lavas and the Ruapehu basalt have higher 143 Nd/ 144 Nd ratios. The isotopic, and major and trace element composition of Te Herenga andesite can be replicated by models involving mixing of an intra-oceanic andesite with a crustal component derived from a meta-igneous composition. Post-Te Herenga andesites show considerable variation in major and trace element and Sr and Nd isotopic compositions ( 87 Sr/ 86 Sr ranges from 0·7049 to 0·7060 and 143 Nd/ 144 Nd from 0·51264 to 0·51282). The range of compositions can be modeled by assimilation–fractional crystallization (AFC) involving meta-greywacke as the assimilant, closed-system fractionation, or by mixing of intra-oceanic andesite or basalt and a meta-greywacke crustal composition. Plagioclase and pyroxene compositions vary over wide ranges within single rocks and few of these have compositions consistent with equilibration with a melt having the composition of either the host-rock or groundmass. The 87 Sr/ 86 Sr compositions of plagioclase also vary significantly within single whole-rock samples. Glass inclusions and groundmasses of andesitic rocks all have dacitic or rhyolitic major and trace element compositions. The application of various mineral geothermometers and geobarometers indicates pre-eruption temperatures between 950 and 1190°C and pressures ranging from 1 to 0·2 GPa. These pressure estimates are consistent with those obtained from xenolith mineral assemblages and geophysical information. Plagioclase hygrometry and the paucity of amphibole are indications that melts were relatively dry (〈 4 wt % H 2 O). Magmas represented by Ruapehu andesites were dacitic or rhyolitic melts carrying complex crystal and lithic cargoes derived from the mantle and at least two crustal sources. They have evolved through a complex interplay between assimilation, crystal fractionation, crustal anatexis and magma mixing. Parental magmas were sourced in both the mantle and crust, but erupted compositions very strongly reflect modification by intracrustal processes. Geochemical variation in systematically sampled lava flow sequences is consistent with random tapping of a complex plumbing system in which magma has been stored on varying time scales within a plexus of dispersed reservoirs. Each magma batch is likely to have had a unique history with different sized magma storages evolving on varying time scales with a specific combination of AFC and mixing processes.

Description: Regular large-scale edifice collapse and regrowth is a common pattern during the long lifespans of andesitic stratovolcanoes worldwide. The 〉130 k.y. history of Mount Taranaki, New Zealand, is punctuated by at least 14 catastrophic collapses, producing debris avalanche deposits of 1 to 〉7.5 km 3 . The largest of these sudden events removed as much as one-third of the present-day equivalent cone. The resulting deposits show similar sedimentary and geomorphic features, suggesting similar proto-edifice characteristics, failure trigger mechanisms, and runout path conditions. Each collapse was followed by sustained renewed volcanism and cone regrowth, although there are no matching stepwise geochemical changes in the magma erupted; instead a stable, slowly evolving magmatic system has prevailed. Last Glacial climatic variations are also uncorrelated with the timing or magnitudes of edifice collapse. We demonstrate here that, if the magmatic composition erupted from stratovolcanoes is constant and basement geology conditions are stable, large-scale edifice collapse and the generation of catastrophic debris avalanches will be governed by the magma supply rate. Using a mass balance approach, a volume-frequency model can be applied to forecasting both the probable timing and volume of future edifice failure of such stratovolcanoes. In the Mount Taranaki case, the maximum potential size of a present collapse is estimated to be 7.9 km 3 , while the maximum interval before the next collapse is 〈16.2 k.y. The current annual collapse probability is ~0.00018, with the most likely collapse being a small one (〈2 km 3 ).