V52A-01
Origin of high-Mg andesites at White Island
On White Island, New Zealand, the intensified period of strombolian-volcanian and phreatomagmatic explosive activity that commenced in March 1977 led to eruption of unusually primitive, high-Mg andesites. These are Fo80- 93 olivine-saturated rocks that have MgO contents up to 10 percent (Mg# = 65-71) and SiO2 of 56-58 percent. They have incompatible trace element characteristics that are typical of arc rocks. 87Sr/86Sr, 143Nd/144Nd and 176Hf/177Hf ratios (0.7049-0.7053, 0.51282-0.51266 and 0.28301-0.28298, respectively) are consistent with subducted sediment addition and/or crustal input but there is no clear correlation of either isotope ratio with MgO. The rocks have modest (3-10 percent) 238U excesses at low (230Th/232Th) ratios (0.697 to 0.722). 226Ra- 230Th disequilibria is also restricted but, unusually, includes both 226Ra excesses and deficits with (226Ra/230Th) = 0.94-1.07. (210Pb/226Ra)o ranges from 0.98 to 1.52 requiring gas accumulation that may increase over time and with decreasing MgO. Sr/Y and Tb/Yb ratios are both low and relatively invariant at 8 and 0.3, respectively, and along with the 238U excesses preclude an origin in which residual garnet was involved. The occurrence of some 226Ra deficits suggests the presence of residual amphibole during partial melting for some samples. Rapid magma ascent (to preserve the 226Ra disequilibria) limits the amount of possible melt - wall rock interaction that might reduce source-derived Tb/Yb ratios and in the mantle or raise 87Sr/86Sr in the crust. The White Island high-Mg andesites did not form by partial melting of eclogite in the subducting Pacific plate. Their primitive, olivine-saturated compositions suggest that their source was peridotitic and experimental data suggest that melting at low temperatures at 0.5-1.5 GPa and in the presence of elevated alkalis can reconcile the high SiO2 and MgO of the rocks. These conditions appear to be favoured by the location beneath continental, rather than oceanic lithosphere.
V52A-02
Generating variable Mg-numbers during melt segregation of TTG arc magmas
Major, trace, and REE compositions of both Archean TTGs and modern adakite-like magmas have been used in conjunction with batch melting experiments and models to infer source rock compositions, depths of melting, and tectonic setting. However, the impact of physical melt segregation processes on the magma geochemistry has not been considered to date. When melt initially migrates, it can interact with its partially molten host and this process may have a profound impact on the composition of the segregated melt as it leaves the source region. In this study, we are investigating through numerical modeling and experimental testing, the hypothesis that TTG arc crust formation is not only a function of partial melting of a mafic source region, but the time and length scales for melting and extraction, and melt segregation mechanisms themselves. In the experimental investigation we have designed melt segregation equilibrium (MSE) experiments to reproduce the local changes in bulk composition that are predicted to occur in response to buoyancy-driven melt segregation along grain edges and associated compaction of the solid residue. We conducted the experiments between 925-1000°C at 1.4 GPa; the same conditions as an earlier set of direct partial melting experiments on the same mafic amphibolitic starting material. The MSE experimental results show distinct differences in the melt and solid phase compositions and solid phase stability when compared with the results from the earlier direct partial melting experiments. Overall, the resulting melt compositions in the MSE experiments are lower in the An component and have higher Mg-numbers when compared with the direct partial melting results. Modally, the charges have changed too, with a reduction in hornblende and plagioclase and an increase in garnet and clinopyroxene as a function of increasing temperature. One interesting aspect is that hornblende is stabilized and is a new phase in a number of the experiments at the mid-temperature range. The results suggest that if dynamic melt segregation and equilibrium processes are active, they may modify the normally robust geochemical indicators, such as Mg-numbers, which are typically used to develop models of TTG petrogenesis.
V52A-03
Subduction Contributions in the Trans-Mexican Volcanic Belt: Implications from Lava Chemistry and Hf-Nd-Pb Isotopes
Despite thick continental crust, near primitive lavas erupt throughout the Trans-Mexican Volcanic Belt (TMVB). In order to distinguish and better constrain subduction contributions and effects of crustal contamination, we analyzed samples representing subducting sediments from DSDP Site 487, and Quaternary lavas from stratovolcanoes and cinder cones, including alkaline "high-Nb" lavas from the Sierra Chichinautzin Volcanic Field (SCVF) showing negligible subduction signature in its trace element chemistry and representing melts of the mantle wedge. Our primary observations and implications are: (1) The high-Nb SCVF ‘intraplate' lavas define a linear trend along the "Nd-Hf mantle-crust array", defining the composition of the mantle wedge. (2) Popocatepetl and Nevado de Toluca stratovolcanoes show the highest Nd and Hf isotope ratios, higher than the ‘intraplate' lavas, indicating their sources are more "depleted mantle-like" than the regional mantle wedge. (3) The Popo and Toluca chemical and isotopic trends sharply contrast with Pico de Orizaba, which shows classic indications of crustal contamination (e.g. high 207Pb/204Pb, low Nd-Hf isotope ratios), consistent with contamination by local Precambrian crust. (4) Higher Nd-Hf isotopes in Popo and Toluca lavas also correlate with lower Pb isotope ratios, and lower Lu/Hf and Zr/Hf. Together, these data indicate contributions from subducted Pacific oceanic crust and hydrothermal sediment. (5) Popo and Toluca are also enriched in Th/LREE compared with ‘intraplate' lavas, reflecting subducted sediment contributions. (6) Nd-Hf isotope ratios of hydrothermal sediment from DSDP Site 487 lie on the "seawater array", with high Hf isotope ratios compared to the "mantle-crust array". Popo and Toluca Nd-Hf isotopes display a shallower slope than the "intraplate lava Nd-Hf array", reflecting contributions from hydrothermal sediment. Popocatepetl and Toluca lavas therefore avoid substantial crustal contamination of mantle wedge-derived melts, despite traverse through a ca. 35 km thick continental crust. Their compositions can be simply modeled, and reflect components from a composite slab melt (represented by altered Pacific ocean crust and DSDP 487 sediment), plus the regional mantle (represented by high-Nb SCVF lavas). The best estimate of the slab melt is a ca. 5 percent sediment melt plus a low-degree (ca. 0.5 percent) altered oceanic crust melt, with a mixing ratio of about 1:9. Thus, the integrated data allow us to clearly distinguish between mantle and crustal sources and point to substantial subducted slab contributions to TMVB lavas.
V52A-04
Implications for Melt Differentiation Processes in the Central Mexican Volcanic Belt from 'Zoned' Monogenetic Volcanoes
Across-arc mass balances require the knowledge of composition and the origins of the mass of melt. This "melt mass" may derive from slab, mantle, or crustal sources, whereby these possible sources will be reflected in magma chemistry. In order to identify the arc melt sources, we are studying a broad range of alkaline ('intraplate- type') to calc-alkaline ('arc type') Holocene volcanic rocks in the Central Mexican Volcanic Belt (CMVB) constructed on thick continental crust (~40-47 km). Sr-Nd-Pb isotope ratios similar to those from subarc mantle xenoliths in all erupted magmas, together with trace elements, point towards a dominant mantle origin of melts along with some slab-derived component, but negligible crustal additions. In order to understand how isotope and trace elements relate with major element diversity, we focus on monogenetic volcanoes with a significant zonation in major elements (e.g. Chichinautzin, Guespalapa, Suchiooc). Monogenetic volcanoes build in a single event and hence their magmas must derive from similar sources. Within zoned monogenetic centers, two different trends are recognized: (1) a 'source trend'; and (2) 'differentiation trends'. The 'source trend' describes the trend of decreasing FeO* and TiO2 with increasing Mg#, which precedes high-Mg# (>60-75) calc- alkaline and alkaline magmas. The 'source trend' parallels the tholeiitic trend of Fe-enrichment with decreasing Mg# observed in oceanic basalts, but culminates at lower overall maxima of FeO* (~9 wt%) and TiO2(~2 wt%) (FeO* ~16 wt% and TiO2 ~3.5 wt%). High-Ni olivines contained in alkaline and calc-alkaline 'source trend' magmas suggest that the 'source trend' is generated through repeated mantle melting induced by fluid addition from slab (Straub et al., G3, submitted). This supports models of mantle origin of primary high-Mg# andesite melts. However, the majority of CMVB magma have lower Mg#- numbers and plot on trajectories that emanate from any point of the 'source trend' towards lower FeO* and TiO2 with decreasing Mg#. These 'differentiation trends' are preserved fully or partially within individual monogenetic centers. Our preliminary trace element and isotope data suggest that these 'differentiation trends' can neither by explained simply by fractional crystallization nor by crustal assimiliation but require more complex petrogenesis that may involve serial addition of slab components to mantle sources. These competing models are testable by comprehensive trace elements and isotope studies in combination with mineral and melt inclusions work on zoned monogenetic centers.
V52A-05
Andesite Petrogenesis in Nevado de Toluca Stratovolcano, Central Mexico
A popular model for andesite petrogenesis in continental arcs involves a hydrous parental basalt that fractionates and probably assimilates continental crust during ascent. Nevertheless, andesites erupted from polygenetic volcanoes in the Trans-Mexican Volcanic Belt are not entirely consistent with this scenario because: (1) they display a shift to higher SiO2 contents at similar Mg# than true basalts, (2) they trend to lower HREE and HFSE contents with increasing SiO2, and (3) they often show correlated isotopic compositions with proxies for slab inputs but not with fractionation indexes. Young andesites from Toluca stratovolcano (1-0.042 Ma) also display modest adakite-like features (Sr/Y<60) that have been associated to melts from the subducted oceanic crust (Martínez-Serrano et al., 2004). More extensive sampling of the Toluca rock-suite revealed that some other andesites (2.6-1 Ma) also exhibit very strong negative Ce (Ce/Ce~0.25) anomalies, fractionated HREE patterns (Gd/Yb~4.2), as well as low Zr/Sm (~12.6), and Sr/Y (~12) ratios. These features are not easily explained by low or even high pressure differentiation from a common primitive magma, unless enormous quantities of fractionating accessory minerals are taken into account. And yet these geochemical signals are almost identical to those observed in the pelagic sedimentary horizon of the subducted Cocos plate sampled at DSDP site 487, and thus provide strong evidence for slab-derived sediment contributions to the petrogenesis of Toluca andesites. Since sediment transfer to the Toluca source must have occurred in the form of a silicate melt, the new evidence brings further support to the slab melting hypothesis in the Mexican subduction zone. Interestingly, Ce/Pb ratios of the Toluca rocks display a linear positive correlation with Pb isotopes, that departs from the pelagic sediment values, and extends to the enriched Pb isotopic compositions of intraplate-type volcanic rocks from the Chichinautzin volcanic field. Thus the Toluca rocks likely represent discrete slab-derived melts coming from different portions of the subducted slab that have interacted with the distinctively enriched Mexican mantle wedge.
V52A-06
New Constraints on the Petrogenesis and Time Scales of High Mg Andesite Evolution, White Island, New Zealand
Unusually primitive, high Mg andesites have been erupting since 1977 on White Island, which lies off shore of New Zealand in the southern part of the Tonga-Kermadec island arc. They are Fo80-93 olivine saturated rocks which have MgO contents up to 10% (Mg# = 65-71) with SiO2 of 56-58% and contained 1.4-4.4% H2O. 143Nd/144Nd and 176Hf/177Hf ratios of 0.51282-0.51266 and 0.28301-0.28298, respectively, are consistent with subducted sediment addition and there is no clear correlation of either isotope ratio with MgO. They have incompatible trace element characteristics that are typical of arc rocks and flat rare earth element patterns. Sr/Y and Tb/Yb ratios are both low and relatively invariant at 8 and 0.3 respectively and along with the 238U- 230Th disequilibria preclude an origin in which residual garnet was involved. The occurrence of 226Ra deficits and the preservation of a negative correlation between (226Ra/230Th) and (230Th/238U) suggest the presence of residual amphibole during partial melting followed by rapid magma ascent. Recent compilations of peridotite experimental data suggest the presence of high MgO (7-10%) magmas will equilibrate with high SiO2 contents (52-57%)when melting takes place at shallow levels and hence do not require an eclogitic component. This would suggest that the source of the high-Mg andesites from White Island was peridotitic with a small component of amphibole rather than eclogitic. Such observations are also consistent with the U-series disequilibria data.
V52A-07
High-Mg basalts as a Signal of Magma System Replenishment at Lopevi Island, Vanuatu
Lopevi is is a basalt to basaltic andesite island stratovolcano in central Vanuatu and is part of a long-lived, mature Island Arc chain. Central Vanuatu is tectonically influenced by the subduction of the D'Entrecasteaux zone. Primitive rock types that have been identified from the arc include picrites, ankaramites and high MgO basalts. High MgO rocks are generally considered to be a relatively rare component of arc-type magma suites but as detailed sequence sampling of individual volcanoes occurs, they have been identified more often. Here we report on the occurrence of high-Mg basalts in a sequence of lavas erupted in the last 100 years from Lopevi volcano. Activity at Lopevi is characteristically intermittent with eruptive sequences occurring over a c. 6 year period, separated by longer periods of repose. A major eruptive episode in 1939 caused evacuation of the island and the next eruptive episode in the 1960's also led to evacuation. The 1960's cycle of activity ended in 1982. The most recent phase of activity commenced in 1998 with a return to eruption of more siliceous, high alumina basaltic andesite. Geochemical data show that the 1960's lavas were different from those erupted earlier and later. They are olivine basalts with up to 9 wt percent MgO, 70 ppm Ni and 300 ppm Cr; Al2O3 content is about 12 wt percent. The 2003 lavas and pre-1960's lavas, in contrast, are basaltic andesites with c. 4 wt percent MgO, less than 25 ppm Ni, less than 100 ppm Cr and c. 20 wt percent Al2O3. The 1960's Lopevi sequence of eruptions represents an injection of a more primitive, high MgO magma at the end of a 21 year quiescent period after the major eruptions of 1939. Injection of small batches of more primitive magmas over decadal time periods at Lopevi marks the initiation of a new magmatic cycle. The occurrence of high MgO magmas as part of a cycle that includes typically low MgO arc type rocks demonstrates a consanguineous relationship and shows that high MgO arc type rocks are part of a genetically linked suite rather than a distinct magma type. Their comparative scarcity in many subduction related associations is probably a function of tectonic environment rather than of fundamental petrological factors.
V52A-08
High-Magnesium Andesites in south-eastern Papua New Guinea: tectonic control of primitive Arc magmas?
The late Cenozoic high-K arc-type volcanic association in south-eastern Papua New Guinea developed in an environment of complex tectonic processes including obduction, subduction, rifting, sea floor spreading and uplift of metamorphic core complexes. However, although the arc is demonstrably active there is no seismic evidence for current subduction.. The volcanoes of the association define an arc extending a distance of 250 km from the Papuan Peninsula south-eastward through the D'Entrecasteaux Islands into the Louisiade Archipelago. Rock types are predominantly basaltic andesite and andesite but include basalt, dacite and rhyolite. These rocks constitute a high-K arc-type volcanic suite typical of many subduction related volcanic associations. A sub-set of the Papuan suite is a group ranging from basalt to dacite which, although comparable in most other aspects of their geochemical compositions, are higher in MgO, Cr and Ni. These relatively high-Mg rocks (up to 12 wt. percent MgO at the basaltic end of the spectrum) are less porphyritic and have simple olivine- or clinopyroxene- dominated phenocryst assemblages compared with the ‘normal' rocks of the association. The ‘normal' rocks are typical plagioclase phyric arc-type rocks containing augite and hypersthene with or without olivine, hornblende and biotite. Both high-Mg and ‘normal' rock types are spatially and temporarily interlinked. The high-Mg rocks are interpreted to represent magmas derived by partial melting of subduction modified mantle which have risen rapidly from their mantle source. In contrast the ‘normal' low-Mg rocks represent magmas which were modified by shallow intracrustal processes. In this case the source of the magmas is subduction-modified mantle rendered fertile by subduction although tapping of this source is not directly linked to active subduction. The unusual abundance of high-Mg in south-eastern Papua is linked to extensional tectonics which allowed deep sourced magmas to pass through the crust without significant modification. This model is consistent with the current tectonic setting in which sea floor spreading of the Woodlark Rise is impinging on the continental crust of Papua. This explanation supports the hypothesis that high-Mg andesites can represent primitive melts in subduction settings that have avoided the trap of intracrustal processes.