V41D-0791
The Massive Compound Cofre de Perote Shield Volcano: a Volcanological Oddity in the Eastern Mexican Volcanic Belt
Cofre de Perote volcano anchors the northern end of the easternmost of several volcanic chains orthogonal to the E-W trend of the Mexican Volcanic Belt (MVB). Its structure, geochemistry, and volcanic history diverge significantly from that of the large dominantly andesitic stratovolcanoes that have been the major focus of research efforts in the MVB. Andesitic-trachyandesitic to dacitic-trachydacitic effusive activity has predominated at Cofre de Perote, forming a massive low-angle compound shield volcano that dwarfs the more typical smaller shield volcanoes of the central and western MVB. The 4282-m-high volcano overlooking Xalapa, the capital city of the State of Veracruz, has a diameter of about 30 km and rises more than 3000 m above the coastal plain to the east. Repeated edifice collapse has left massive horseshoe-shaped scarps that truncate the eastern side of the edifice. Five major evolutionary stages characterize the growth of this compound volcano: 1) emplacement of a multiple-vent dome complex forming the basal structure of Cofre de Perote around 1.9-1.3 Ma; 2) construction of the basal part of the compound shield volcano from at least two main upper-edifice vents at about 400 ka; 3) effusion of the summit dome-like lavas through multiple vents at ca. 240 ka; 4) eruption of a large number of geochemically diverse, alkaline and calc-alkaline Pleistocene-to-Holocene monogenetic cones (likely related to regional volcanism) through the flanks of the Cofre de Perote edifice; 5) late-stage, large-volume edifice collapse on at least two occasions (ca. 40 ka and ca. 10 ka), producing long-runout debris avalanches that traveled to the east. An undated tephra layer from Cofre de Perote overlies deposits likely of the youngest collapse. Cofre de Perote is one of several volcanoes in the roughly N-S-trending chain that has undergone major edifice collapse. As with Citlaltepetl (Pico de Orizaba) and Las Cumbres volcanoes, Cofre de Perote was constructed at the eastern margin of the Altiplano, with pronounced differential relief and sloping substrate promoting failures toward the Gulf of Mexico coastal plain.
V41D-0792
The Baja California Peninsula During the Last 12.5 Ma as Example of the Role of Plate Stresses in Controlling Active Volcanism
Late Cenozoic volcanism in the Baja California Peninsula records a major plate reorganization between the Farallon and North America plate boundary that started ca.12.5 Ma. A concomitant change in the style and geochemical diversity of volcanic activity during the last 12.5 Ma has been thoroughly documented, although the origin of post-subduction volcanism has remained debatable. In this work we used a finite element model to estimate the role played by stress changes in the evolution of volcanism in the Baja California Peninsula. The model consists of an elastic layer representing the North America plate resting on an incompressible viscous fluid approximating mantle rheology. Deformation in the model is driven by viscous and buoyant forces in the mantle, which in turn depend on the imposed boundary conditions. Our results show that during the active subduction stage a tensional-stress build up was present in the Comondu volcanic arc. By the end of subduction (12.5 to 6.5 Ma) the model indicates that the entire elastic plate experienced tension, therefore promoting the re- activation of old fractures and allowing the trenchward continuation of volcanic activity during this period. For younger times the model shows that tensional stresses increased one order of magnitude along the base of the elastic plate, favoring migration of volcanic activity back to the east. In all cases, the state of stress predicted by the model is in good agreement with the observed geologic record, therefore highlighting the relevance of the local stresses in controlling the location of active volcanism.
V41D-0793
Lithium in melt inclusions records crustal assimilation at Volcan Jorullo, Mexico
Trace element concentrations and Li isotope ratios have been analyzed in olivine-hosted melt inclusions found in tephras from Volcan Jorullo, a monogenetic cinder cone of the Michoacan-Guanajuato Volcanic Field, Mexico. The melt inclusions range from basalt to basaltic andesite in composition, with K2O and incompatible trace element concentrations increasing with SiO2 content. Lithium concentrations range from 2 ppm in the most primitive inclusions to 12 ppm in the most evolved. Some of the lava flows from Jorullo contain abundant granitic xenoliths, and most contain plagioclase xenocrysts, suggesting that the lavas have assimilated some amount of crustal material. Lithium concentrations in the lava flows are 10-12 ppm, while the Li concentration in a granitic xenolith from one of the flows is 29 ppm. We interpret the basaltic melt inclusions containing 2 ppm Li to represent the primitive melt, while the higher Li in the more evolved melt inclusions and the lava flows results from assimilation of granitic country rock. The positive correlation between Li/Y and Eu/Eu* in the melt inclusions supports this theory of Li enrichment by crustal assimilation. The δ7Li in the melt inclusions ranges from typical arc lava values of 1.1±1.1‰ in the most primitive (low-Li) melt inclusions to extremely light values down to -9.2±0.9‰ in the more evolved (high-Li) melt inclusions. The significance of these data lies in determining whether the light Li in the evolved melt inclusions represents the initial composition of the contaminated melt itself, or if the Li isotope ratios in these inclusions have been altered by post-entrapment processes such as diffusive exchange with olivine or with the evolving melt through the olivine host. Lithium is known to diffuse quite quickly, and furthermore 6Li has been shown to diffuse more quickly than 7Li, which could result in kinetic fractionation of Li in the melt inclusions post-entrapment. However, it is puzzling that the most evolved melt inclusions, which were presumably the last melt inclusions to be trapped prior to or during eruption, have the most fractionated δ7Li. Lithium isotopic analysis of granitic xenoliths collected from the lava flows will determine if the country rock can provide a direct source of light Li. We propose that the high mobility of Li combined with high concentrations of Li in crustal rocks relative to basalts may combine to make Li an ideal tracer of crustal interaction even in relatively primitive lavas.
V41D-0794
Pre-eruptive Volatile Contents of Mafic Magma at Popocatépetl Volcano, Mexico, from Olivine- hosted Melt Inclusions
Volcanic gases are intimately related to degassing and their study allows the characterization of the composition, concentration and origins of volatiles in magma. Exsolution and loss of volatiles, particularly H2O, CO2 and S, leads to major changes in magma crystal content, density and viscosity, which in turn can produce major shifts in eruption style (explosive to dome building). Therefore understanding degassing behavior at active volcanoes is crucial for hazard evaluation. Gas emissions at active volcanoes (mainly SO2), has commonly been attributed to dissolved sulfur in the silicate melt before eruption. However, for most eruptions for which remote sensing and melt inclusion data are available, excess eruptive S in gas emissions is observed so a pre-eruptive gas phase could be another source of sulfur emissions (Wallace, 2001). We analyzed melt inclusions in Fo 88-90 olivine from the May 11 and June 30, 1997 eruptions of Popocatépetl (Mexico). The melt inclusions are basaltic trachyandesite in composition, with 3.17 - 5.30 wt% MgO and 50.32 - 54.75 wt% SiO2. Analyses by FTIR and electron microprobe show that the melt inclusions contain 0.2 – 4.70 wt% H2O, <100 - 2413 ppm CO2, 624 - 2222 ppm S, and 821 – 1169 ppm Cl. Calculated vapor saturation pressures for these melt inclusions (<0.1 - 6 kb) demonstrate that degassing of mafic magma beneath Popocatépetl starts at a minimum depth of 25 km. In addition, the high dissolved CO2 and S contents in the melt inclusions suggest that mafic magma recharges was responsible for the high measured fluxes of CO2 and SO2 at Popocatépetl during the period from April to June, 1997.
V41D-0795
Sulfur Speciation and Oxygen Fugacity in Primitive Magmas From the Trans-Mexican Volcanic Belt
Sulfur dissolves in silicate melts in both reduced (S-2) and oxidized (S+6) forms, and the ratio of species depends on the oxygen fugacity of the melt. The more oxidized the magma, the more sulfur is disolved as sulfate and the higher the overall S solubility (e.g., Luhr, 1990). We have measured the speciation of S in olivine- hosted melt inclusions from 11 mafic cinder cones and 1 maar in the Trans-Mexican Volcanic Belt using the S Kα wavelength shift method (Carroll and Rutherford, 1988). Four of the cinder cones are from the Colima Graben and are high-Mg silica-undersaturated potassic rocks. The other cinder cones are from the Michoacan- Guanajuato Volcanic Field (MGVF) and include calc-alkaline high-Mg basalts and basaltic andesites. The maar, also in the MGVF, erupted alkali basalt. A basaltic andesite from the Colima Graben was also analyzed due to its transitional chemical composition between the potassic magmas and the MGVF calc-alkaline magmas. For the calc-alkaline melt inclusions, S+6 accounts for 65-90% of total dissolved S. Similarly, the potassic and basaltic andesite melt inclusions have 60-90% and 78-94% of total S as S+6 respectively. In contrast the alkali basalt melt inclusions have lower values, with 50-60% of total S dissolved as S+6. Some cinder cones exhibit nearly the entire range of melt S speciation whereas others have restricted ranges, but there does not appear to be a correlation between melt degassing (based on H2O contents of the melt inclusions) and either oxygen fugacity or S speciation. Based on the average value of melt inclusions from each cone and the relationship of S speciation to oxygen fugacity (Wallace and Carmichael, 1994), the calc-alkaline cinder cones, including the basaltic andesite, have oxygen fugacity values of NNO+0.9 to +1.3. The potassic cinder cones have similar values of NNO+0.9 to +1.3. The alkali basalt averages NNO+0.5. The values have 1 standard deviation uncertainties of ±0.2 log units for the calc-alkaline basalts, ±0.4 for the potassic cones, ±0.5 for the basaltic andesite and ±0.08 for the alkali basalt, which emphasizes the distinction of the alkali basalt with respect to the other magma types. The potassic and calc-alkaline magmas appear to have similar oxygen fugacities that are distinctly higher than the alkali basalt. This distinction cannot be explained by compositional differences but appears to be related to the mantle source of the magmas and the presence or absence of a subduction derived H2O-rich component.
V41D-0796
The Dynamics of Magma Chamber Processes at Popocatepetl Volcano, as Recorded by Plagioclase Phenocrysts
Popocatépetl Volcano had 5 Plinian eruptions in the last 23,000 years, and is currently undergoing passive degassing and small-scale explosions. We seek to clarify the roles of crystallization, magma mixing, and assimilation of the calcareous basement in controlling magma composition by studying the compositional and 87Sr/86Sr isotopic variations in plagioclase phenocrysts. Compositional variations were analyzed by electron microprobe (EPMA) and zones based on An content analyzed for Sr isotopes using a 213 nm laser ablation system coupled with an IsoProbe multicollector, magnetic sector ICP-MS. Based on variations of An content we can classify crystals into three groups. The first group has albite-rich cores that are mantled by nearly constant composition plagioclase of An50. The second group has anorthite-rich cores that are mantled by nearly constant composition plagioclase of An40. The third group lacks cores with disparate compositions. Despite the relatively simple grouping, isotopic trends are variable among groups. For those with albite-rich cores, some have constant isotopic values from core to rim, whereas others have more radiogenic cores than their rims (e.g. Δ c-r87Sr/86Sr > 0.0015), and yet others have less radiogenic cores than rims (e.g. Δ c-r87Sr/86Sr < 0.0022). Equally complicated zonings are found in the other two groups. Variable correlations between isotopic trends and An contents show that the magma system of Popocatepetl has been quite dynamic. Crystals with albite-rich cores related to less radiogenic rims indicate mixing with a more mafic magma. Crystals with variable An content and more radiogenic rims might reflect local basement assimilation. Crystals with minor variations of An content and nearly constant isotopic values reflect crystallization. Crystals with variable An content and nearly constant isotopic values may reflect processes as decompression, changes in temperature, or changes in water content.
V41D-0797 INVITED
Xenoliths From Isla Isabel, Nayarit, Mexico: The Nature of the Upper Mantle Underneath the Western Part of the Mexican Volcanic Belt
Isla Isabel is located ~65 km NW of San Blas (Nayarit), off the Pacific coast of central Mexico. The island is a Quaternary (Ar/Ar < 0.7 Ma) volcanic complex built atop attenuated continental crust. Isabel lies on the east side of the mouth of the Gulf of California, near the area previously occupied (early Pliocene) by Los Cabos Block. Southeast of Isabel, on the mainland, is the NW-trending Tepic-Zacoalco rift, a major volcano-tectonic structure in the western part of the Mexican Volcanic Belt. On land, the rift is the boundary between the Jalisco and Sierra Madre Occidental blocks, and Isabel lies along its projection. Immediately S of Isabel is the San Blas Trough, a swale that trends NW-SE, co-linear with a gravity lineation parallel to the Tamayo and San Blas fault zones, which are the transform boundaries between the northern Rivera and North American plates. Plio-Quaternary alkaline and calc-alkaline lavas have erupted contemporaneously in the Tepic-Zacoalco rift, but so far no mantle xenoliths have been reported in them. Isabel's rocks are intra-plate type alkaline basalts to trachybasalts, with 5-6%\ normative Ne. Primary paragenesis in the lavas is: Ol + Pl + Cpx + TMt. Small (< 5 cm) peridotite xenoliths, and xenocrysts derived from them, are ubiquitous in the rocks. Eleven xenoliths were studied comprising 3 dunites, 7 harzburgites (one Pl-bearing), and 1 gabbro. Compared to other Mexican xenolith localities N of the MVB, they are refractory as they are depleted in, or lack, Cpx. Ol crystals in xenoliths are homogenous and their Mg#\ s range as follows: peridotites (92-88), gabbro (84), and Pl-harzburgite (80). Cores of primary Ol phenocrysts (90.5-86.5) in Isabel's volcanic rocks are significantly higher in Mg#\ s than in Ol from other Mexican xenolith localities (max. 86) and overlap with the associated peridotite xenoliths. Such overlap has not been reported for other Mexican xenolith localities. Xenolith equilibration temperatures for 5 peridotites were calculated with the TBKN model. Assuming P=15 kb, temperatures are 1021-1112°C. Pressure estimates, based on Ca-exchange model between co-existing Ol and Cpx, assuming the TBKN calculated at 15 kb, range 4.2-14.8 kb, which lies at the shallow range of pressures for other Mexican xenoliths. Only two samples from the coastal locality of San Quintin match the lowest pressures obtained for Isabel. Although the 5 Isabel's xenoliths are Sp- peridotites, one Pl-harzburgite is present in the suite. The only other Mexican Pl-bearing peridotite xenolith reported is from Pinacate, which lies a short distance from the northern end of the Gulf of California. The unusual nature of the mantle beneath Isla Isabel is attributed to the combination of volcano-tectonic features in the area: a protracted history of subduction-related volcanism associated first to the Sierra Madre Occidental and later to the MVB, attenuation of the continental crust related to the opening of the Gulf of California, and inception of sea floor spreading in the nearby Alarcon Basin.
V41D-0798
Geologic Evolution of the Sierra de Nanchititla Silicic Volcanic Center, Southern Mexico
The Sierra de Nanchititla Volcanic Center (SNVC) is located at the borders of the States of Mexico, Michoacán and Guerrero, in the northern sector of Sierra Madre del Sur magmatic province (SMS) of México. Previous studies in the areas of Taxco and Tilzapotla recognized an Early Tertiary brittle deformation along WNW-ESE to NW-SE trending strike slip fault systems at the north of the SMS (Alaniz-Alvarez et al., 2002, JVGR; Morán-Zenteno et al., 2005, JVGR). This fault systems apparently favored the emplacement of several silicic volcanic centers aligned in a WNW-ESE direction during the Eocene and earliest Oligocene. These volcanic centers include Huatla, Tilzapotla, Taxco, La Goleta and Sierra de Nanchititla. We carried out a detailed geologic study of the SNVC backed by Ar/Ar geochronology and trace element geochemistry. Our results indicate that the SNVC is part of the Eocene magmatic episode recorded in the SMS. The first volcanic rocks in the area are basaltic-andesitic lavas dated at ~42 Ma. However, the activity of the SNVC proper is concentrated between 37.8±0.2 and 35.6±0.5 Ma and was characterized by the emplacement of rhyolitic sub-volcanic bodies (cryptodomes), voluminous pyroclastic deposits and endogenous domes. The oldest magmatic event is the intrusion of WNW-ESE-trending mafic aphyric dikes dated 35.6±0.5 Ma that fed the Puerto El Salitre andesite. Magmatic activity continued with the emplacement of cryptodomes with sub-volcanic texture. The main phase of the SNVC took place in late Eocene with the emplacement of three units: (1) WNW-ESE-trending felsic pyroclastic dikes feeding the large Nanchititla Ignimbrite (35.6±0.5 Ma), (2) rhyolitic domes with vents aligned in a WNW-ESE direction (35.6±0.1 Ma), the and (3) a pyroclastic flow (Tequezquite Ignimbrite) likely vented from a dome. Geochemical data are consistent with a differentiation process, starting with primary mafic magmas and originating both basaltic- andesitic and andesitic rocks. These magmas present typical arc-like features, and the oldest rhyolitic rocks tend to have a peraluminous character. Since the distribution of volcanic bodies and centers in the SNVC coincides with WNW-ESE-trending regional lineaments, and considering the significant amount of material emplaced during a relatively short time span (~2.2 Ma), it can be concluded that the formation of the SNVC was produced by the trapping and differentiation of magma along a pre-existing, and inactive, regional structure. The absence of a clear sub circular depression, intracaldera deposits, ring or radial dikes, and ring domes suggest that the magma chamber feeding the ignimbrite was likely very deep and that the ignimbrite was emplaced through WNW-ESE trending fissure with limited or negligible collapse of the venting area.
V41D-0799
A Study of the Source Processes of Colima Volcano Explosions
Colima volcano, considered as Mexico's most active volcano, has presented several intermittent effusive and explosive phases in recent years. During 2005, a sequence of explosive events with VEI less than or equal to 3 occurred. This activity presented the most intense explosions since the seismic network was deployed. Many of the explosive events were recorded by the digital three-component seismic stations operated by the University of Guadalajara and Jalisco State Civil Defense. These signals were recorded not only by stations located on the volcanic edifice, but also by stations on the northern coast of Jalisco (MCUJ, BSSJ) and Ceboruco Volcano at 184, 182 and 200 km distance, respectively. A study of these signals will be presented. Each explosion was preceded by a seismic event. Nevertheless, the located earthquakes preceding the explosions did not show a common source under the volcano structure, which suggests the existence of a complex structure with possibly more than one conduit, this is also confirmed from a first motion analysis for station F03J, located 12 km at north of the volcano. From analysis of the first ten seconds of the seismic signal on F03J using different representations of the seismic signals, such as waveforms, spectra, time-frequency and time-scale analysis, it is suggested that the source processes are non-stationary, implying that for the case of this period, a general model of the source process of the Colima volcano explosions can not be formulated. The size of the events is evaluated using different criteria. A clear relation between the magnitude of the seismic signals and the amplitude of the sonic and infrasonic waves was not observed.
V41D-0800
Evidence for Slab Melt Contributions to the Mexican Volcanic Belt and Other Young Hot Slab Arcs from Lu-Hf Isotopes
Despite major advances in delineating the processes that govern magma generation at convergent margins, the problem persists of distinguishing slab, mantle wedge, and crustal contributions. A corrollary question is whether there is significant melting of subducted ocean crust. Especially in thick crust regions, the importance of crustal versus mantle contributions to lavas represents a long-standing fundamental issue in arc magma geochemistry. We show that frontal arc magmas from the Central Mexican Volcanic Belt (CMVB), including the large andesitic stratovolcanoes Popocatepetl and Nevado de Toluca, display negligible crustal contamination, and contain substantial contributions from melting of subducted Pacific ocean crust. Despite ca. 50 km thick continental crust, the CMVB erupts near primitive lavas including "high-Nb" alkaline basalts that show negligible "subduction signatures" in their trace element patterns. These "high-Nb" basalts define the regional mantle wedge composition in isotope-trace element space. The "normal" calcalkaline lavas form a negative correlation between Hf isotopes and Lu/Hf. One endmember is like the high Nb basalts representing the regional mantle wedge. The other endmember has higher Hf isotopes (approaching values of Pacific MORB) and very low Lu/Hf of less than 0.04 (e.g. compared to typical values of ca. 0.2 in Pacific MORB). The low Lu/Hf values require low degree partial melting of a source rich in garnet. The high Hf isotopes require a depleted mantle source with isotopes like Pacific MORB. Together the Lu-Hf data indicate a substantial component derived from melting of eclogitic Pacific ocean crust. A key feature of the data is that the stratovolcano lavas showing the largest slab melt signature also show the highest Hf isotope ratios and thus are more "depleted mantle-like" than the regional mantle wedge. Thus, the integrated data allow us to clearly distinguish between mantle and crustal sources in the CMVB and point to substantial subducted slab melt contributions to these lavas. Preliminary data from the Colima region in western Mexico also indicate subducted ocean crust melt in stratovolcano lavas. The oceanic crust at the trench near Western and Central Mexico ranges in age from Pliocene to Miocene, thus represents subduction of a young hot slab. The southern Cascades are another case where a young hot slab is being subducted beneath a thick continental crust, and where both calcalkaline and high-Nb alkaline lavas are erupted. These suites show analogous trace element and isotopic relationships to Mexico that delineate contributions from melting of the subducted oceanic crust to the calcalkaline lavas. In contrast, the subducted ocean crust melt signature is absent from western Pacific arcs, where old cold slabs are being subducted, such as the Marianas, New Britain, and Kermedec. Thus the global data thus suggest that Lu- Hf isotopes represent an effective tracer for slab melting, and that young hot slabs melt and old cold slabs don't. We note that our Colima samples are from Jim Luhr, whose interest, enthusiastic support for MVB studies, and generosity are hereby acknowledged and greatly appreciated, and whose scientific counsel is greatly missed.
V41D-0801
Magnetotelluric Survey Reveals Surficial Conductor Under the Jalisco Block, Possibly Linked to Magmatism in the Area
The composition of the volcanic rocks in the Jalisco Block (JB), particularly those from the Colima rift and Los Volcanes and Mascota regions, are quite similar and show an anomalous enrichment in incompatible elements (e.g., K). Luhr et al. (1989) proposed that the alkaline and calc-alkaline lavas from those areas originated in a common source, which consisted of mantle overlying the subducting Rivera and Cocos plates. Such a source was described as heterogeneously and variably enriched in the incompatible elements. However, besides the surface location of the eruptive centers, no evidence is available for the depth and extent such a mixed mantle/slab-derived source. We conducted two magnetotelluric transects in SW-NE and NW-SE directions that sample a portion of the region where magmatism is present, in the vicinity of the San Sebastián-Mascota area, whose lengths are 200 and 135 km respectively, with stations spaced 20 km on the average. The SW-NE transect extends from the SW coast of the JB through a region 90 km in length, devoid of magmatism, crossing the boundary between the coastal region and the uplifted domain of the JB, and continuing inland. Resistivity inversion of the MT response along this line reveals the conductive character of the down going slab to depths of 60 km. There is a good correlation between the resistivity and the seismological trace of the slab. Beyond the location of the deepest portion of the slab, a sudden change in resistivity pattern is obtained, where a surficial (8- 20 km depth) conductive anomaly is detected. This "perched'' conductor corresponds to the location of volcanism in the Mascota-Los Volcanes area. The NW-SE transect also shows the presence of this conductor, which extends 90 km along the SE portion of the line, from San Sebastián to Los Volcanes, and depths from 8 to 20 km. We submit that this conductor corresponds to the enriched magmatic source proposed by Luhr et al. (1989), feeding the Quaternary volcanism of the region. Additional transects will be carried throughout the JB in order to more fully define the characteristics of this conductor.
V41D-0802
The 1793 Eruption of San Martin Volcano (Los Tuxtlas, Veracruz, Mexico)
San Martin Tuxtla Volcano is located in the State of Veracruz, Eastern Mexico (18.572N, 95.169W, 1650 masl). Its last eruption, which occurred 1793, was described by D. Jose Moziño, a naturalist sent by the Viceroy-of the then New Spain-to report on the eruption. The activity lasted for several months with distinct events of explosive character, which produced thick ash fall deposits in its vicinity. The explosions were heard, among other places, in the coasts of Tampico some 500km NW from the volcano. The ash fall reached distances up to 200 Km from the crater and covered an area of about 112,000 Km2. Following the description of Moziño and the results of field studies we make a reconstruction of the eruption. We identified the air fall deposit from this eruption and present an isopach map. We present radiocarbon ages of the paleosoils under the ash bed as an indirect evidence of its age. This data together with present day wind velocities, and a diffusion-advection model of the dispersion of ashes allow to estimate in at least 10km the altitude reached by some of the eruptive plumes. An estimation of the minimum volume of ash erupted, based on the reconstructed isopachs, is of about 1.3 x 108 m3. Microphotographs of the ashes suggest that the activity was of phreatomagmatic and strombolian nature. Finally, we address some aspects of the volcanic risk in the area derived from our study.
V41D-0803
Subduction of Young Lithosphere at Critical Thermal Ages (10-20 Ma): Incites From Thermal Models and the Trans-Mexican Subduction System with Emphasis on the Importance of Slab Travel Times
The Trans-Mexican Volcanic Belt (TMVB) is a system where the age range of subducting plates is typically thought of as critical in their relation to lithospheric thermal structure (~10-20 Ma). I refer to this age range as "critical" because it is in this range where thermal lithospheres begin to rapidly decrease their correlative influence on the thermal state of the subduction system above the crust/mantle wedge interface for most systems. After ~20 Ma the cool upper portion of downgoing lithosphere becomes sufficiently thick so that crustal reheating and corresponding heat flow in the time between trench subduction and the zone of melt generation, due to the accumulation of conductive and frictional heating, behaves somewhat similarly with little regard to age. Typical slab travel times are on the order of 1.5-2.5 My. However, low dip angles and flat-slab behavior in the Eastern end of the TMVB facilitates anomalously long travel times exceeding 6.5 My! Here I show that while the influence of plate age is clearly significant in determining the holistic thermal geodynamics of subduction systems, the influence can be dampened or enhanced by auxiliary factors. I present thermal modeling cases using a solution for the conduction of heat into an evolving semi-infinite half-space with variable boundary conditions. Preliminary results indicate that extraordinary slab travel times and flat-slab behavior, mantle wedge advection regimes, and plate age thermally enhance and dampen each other. Geochemically, the TMVB shows consistent along-arc changes in light element abundance systematics (B/Be, Li/Yb, Be/Zr). Moderately elevated B/Be (Easterly increases from ~4 to ~12 ppm B/Be) observed in the eastern shallow subduction region is thought to correlate with subduction of an older, cooler portion of the slab. However, greater slab travel times in the Eastern TVMB should simultaneously act to warm the slab and depreciate these values which may partly explain the minimal consistency and magnitude of the TMVB along-arc variations. Li/Yb has a more impressive range of correlative along-arc variation, argued to be the result of greater extents of melting in the east. This can be explained by increased dehydration melting (due to a cooler slab and longer H20 residence times in the slab), a warmer slab affect, or changes in the mean depth of amphibole and garnet crystallization. While elevated B and B/Be values in the eastern TMVB correlate with increased hydration melting, it is likely that the affect is dampened by the positive thermal affects of greater travel times. Comparison to arc rocks in subduction further to the east in Central America, particularly Guatemala, whose B abundance and B/Be ratios are much more elevated (20-70 ppm B/Be) and show convergence on those in the Eastern TMVB. I suggest that the minimal range of B/Be variation seen in the TMVB arc is a result of the extraordinary slab travel times associated with shallow subduction in the eastern TMVB. Furthermore, similar to the B/Be data, the depleted Li/Yb (relative to the rifting region encompassing the Jalisco block) of the Eastern end also continues to converge on values represented in the Central American arc, indicating that the strong variations seen in the TMVB are primarily due to crystallization instead of extents of melting. Thermal modeling results presented here show that sources aside from plate age are capable and likely have influenced the systematic correlations observed in the TMVB and this hypothesis is consistent with the LREE data.
V41D-0804
The activity of the Colima volcano and morphological changes in the summit between 2004 and 2007.
Colima Volcano, located in the West of the Volcanic Mexican Belt, has shown a new cycle of explosive activity beginning May 30 1999, and reaching its maximum in March and April of 2005. This year the explosive activity increased gradually, having the largest event on May 23, when a new dome was created. Hours later this dome was destroyed by a strong explosion, forming an ash column 5.6 km high with subsequent pyroclastic flows that reached a distance of 4.2 km flowing along the ravines of the South sector. On May 30 the most intense explosion in 1999 occurred, when the plume reached heights in excess of 4.4 km above the crater, and piroclastic flows were created. On the same year in July two explosive events occurred of characteristics similar to those in May. These constant explosions caused continuos morphological changes in the summit, the most significant being the collapse of the North and South walls of the crater, in the first week of June of 2005, and the creation of a new crater in July. In 2006 the most significant explosive activity took place during April, May and July, when the eruptive columns reached heights of more than 1500 meters above the crater, occasionally forming small pyroclastic flows. In May of 2007 morphological changes were observed in the summit. Among them a crater explosion on the East side, a new dome was formed on the West side, with 20 m in high and 50 m in diameter. The explosive events continue to date, but they have diminished in size and intensity. This activity was similar to the one observed in 1902-1903 and reported by Severo Diaz (1906), but without reaching the maximum levels of activity reported for 1903, where it had levels of three to five maximum explosive events per day. The photographs and the digital mapping have provided detailed information to quantify the dynamic evolution of the volcanic structures that developed on the summit of the volcano in the course of the last for years.
V41D-0805
The Miocene Tepoztlan Formation (Central Mexico) – Key to a Better Understanding of the Initial Phase of the Transmexican Volcanic Belt
In Miocene times, a major volcano-tectonic change took place in West and Central Mexico due to a reorganization of the tectonic plates in the western Pacific region. Since the mid-Miocene, the Transmexican Volcanic Belt (TMVB) began to form. Until present, few data exist on its initial phase since older volcanic products of the TMVB are widely covered by young volcanic rocks. Furthermore, it is hard to infer the volcano-tectonic history of a region from mostly reworked and redeposited rocks. The studied lower to mid-Miocene volcaniclastic deposits (Tepoztlan Formation) of the southern edge of the TMVB are covered by Quaternary volcanic rocks. Based on sedimentological, petrographical, palaeomagnetic, and geochemical studies we aim to establish a stratigraphic framework and a palaeoenvironmental interpretation of the Tepoztlan Formation, contributing to the decipherment of the origin of the TMVB. The 800 m thick Tepoztlan Formation consists of pyroclastic rocks (flow, surge and fall deposits), lahar deposits (debris-flow and hyperconcentrated-flow deposits), fluvial and lacustrine sediments and occasional lava flows. The clastic material is of volcanic origin exclusively, documenting the environmental response and long-term posteruptive sedimentation effects after initial explosive and effusive eruptions. Based on K/Ar analyses, the Tepoztlan Formation is preliminary dated between 21.8 ± 0.2 Ma and 19.0 ± 1.2 Ma. Palaeomagnetic data show several pole reversals within the studied sequence, allowing a more precise subdivision of the depositional period and thus a better correlation of the sections. Principally, the volcanic rocks of the Tepoztlan Formation show andesitic to dacitic composition; basaltic andesite and rhyolite samples are also present. REE patterns are homogenous with enrichment in LREE and no remarkable element anomaly is present, probably indicating a single magmatic origin. This hypothesis is supported by the relatively short period of deposition of the Tepoztlan Formation. Vertical and lateral distribution of depositional units and stratigraphical data (K/Ar, Ar/Ar of pyroclastic units, palaeomagnetic data) are used to reconstruct the evolution of the depositional environments within time, to detect the volume, type and distribution of the volcanic deposits and thus the type of volcanic eruption. Different sedimentary environments of the Tepoztlan Formation include proximal-to-vent deposits, braided streams and sandy floodplains. Reworking and resedimentation of pyroclastic debris began immediately after an eruption. The style of this response varied between eruptions and between depositional environments. Initially, depositional processes were dominated by debris- and hyperconcentrated flows. Ephemeral lakes developed in ignimbrite-dammed depressions. Later, braided streams developed, reintegrating primary and secondary pyroclastic material and incising channels and scours. The volcaniclastics of the Tepoztlan Formation accumulated for the most part in proximal to medial distance to the vent in volcanic flank and apron settings, pointing to several small volcanoes erupting into lowlands. The integrated sedimentological, geochemical and geophysical study shows that the analysis of autochtonous and reworked volcanic material leads to a better understanding of the processes during initial volcanic arc development.
V41D-0806
Volatile Contents of Olivine-Hosted Melt Inclusions From the Central Oregon High Cascades
The young (~ 16 Ma) and slowly subducting (~ 5 cm/a NNW) Juan de Fuca plate should be mostly dehydrated at typical arc magma generation depths (~ 100 km) beneath the Central Oregon High Cascades, resulting in small amounts of slab-derived fluids infiltrating the overlying mantle wedge (Green and Harry, 1999; Peacock, 2003). To investigate this hypothesis, we analyzed olivine-hosted melt inclusions in mafic tephra deposits from five volcanic centers (Blue Lake Crater, Yapoah Crater, Twin Crater, Palagonite Tuff, and North Sister Volcano) by FTIR, LA-ICPMS, and electron probe to characterize the pre-eruptive volatile contents (H2O, CO2, S, F, Cl) and their relationship to major and trace element compositions. Host olivine phenocrysts (Fo79-84) have δ18O values of 5.35 ±0.2 ‰. The melt inclusions have compositions of 49 - 54 wt. % SiO2 and 4.9 - 5.5 wt. % MgO. H2O contents for the entire dataset vary from ~ 0.2 to 4.0 wt. % and CO2 contents vary from below detection up to 2000 ppm, implying maximum pressures of entrapment between 1.7 and 3.9 kbar during ascent and degassing (Newman and Lowenstern, 2002). Additionally, within-cone averages for inclusions range from 800 - 1500 ppm F, 900 - 1400 ppm S, and 350 - 1000 ppm Cl. For the set of inclusions from each cone, we assume that the highest H2O content is least affected by degassing and therefore represents the minimum amount of H2O in the parent magma. Highly fluid mobile to less fluid mobile element ratios (e.g., Ba/La and Sr/La) correlate positively with maximum H2O/La ratios for Cascades melt inclusions (r2 ~ 0.60; n = 5 volcanic centers). High incompatible element abundances relative to NMORB closely correspond to an EMORB-type source mantle, overprinted with even higher abundances of incompatible and fluid mobile elements (e.g., Ba, U, Pb), resulting in a distinct, negative Nb-Ta anomaly. The volatile data, taken together with olivine δ18O values, trace element abundances and conservative element ratios (e.g., Nb/Y ~ 0.29 - 0.60), suggest partial melting of an EMORB-type mantle source that has been fluxed by an H2O-rich slab component even in this hot, dry end-member subduction zone.
V41D-0807
Snow Peak, OR: Miocene and Pliocene Tholeiitic Volcanism in the Cascadia Forearc
Snow Peak is a voluminous (>150 km3), glacially dissected shield volcano located approximately 50 km southeast of Salem, OR, with a summit height of 1,310 m above sea level. Snow Peak lies approximately 60 km west of the current High Cascade arc axis. Lavas from the southeast face of Snow Peak have been previously dated using K-Ar at ~3 Ma. New Ar-Ar dating indicates that lavas from the northwest face are ~5.4 Ma, and the summit plug is ~6 Ma. Snow Peak volcanics unconformably overlie western Cascade volcanics aged from middle to late Miocene (~10- 17 Ma). The age of Snow Peak is broadly contemporaneous with the initiation of modern High Cascade volcanism. Snow Peak's location provides a rare opportunity to study magmas produced within the modern High Cascades forearc region. The goal of this investigation is to characterize the composition and timing of volcanism at Snow Peak and the role of volatiles in magma genesis. Hypotheses for the formation of Snow Peak include flux melting associated with the Cascadia subduction zone and/or decompression melting associated with extensional faulting. Preliminary geochemical data on the basalts from Snow Peak indicate that they are low-to-medium-K tholeiites (SiO2 47.9-51.7 wt.%, MgO 5.5- 8.3 wt.%, K2O, 0.36-0.55 wt.%) and that they range from primitive to moderately evolved (Mg# 0.51-0.61). Common phenocryst phases are plagioclase, olivine, and clinopyroxene. Textures are typically hypocrystalline, and fine-grained to porphyritic. Mantle-normalized multi-element plots indicate Snow Peak lavas are generally HFSE depleted and LILE enriched. These data are consistent with a preliminary interpretation of a subduction zone signature, yet the major element composition most closely resembles high alumina olivine tholeiite (HAOT), more indicative of extensional environments. The degree of LILE enrichment is significantly lower than in calc alkaline lavas from the High Cascades and western Cascades. Determining the petrogenesis of this forearc center will include a comprehensive analysis of the volcano's major and trace element geochemistry, and additional age dating to constrain eruption rates. Direct measurement of volatiles in olivine-hosted melt inclusions will complement the major and trace element geochemistry in order to measure pre-eruptive water contents.
V41D-0808
Across Arc Variation in Basaltic fO2: Influence of a Subduction Component in the Cascadia Subduction Zone
Oxidation of the subarc mantle in subduction zones can greatly affect mineral phase equilibria, the speciation of volatiles, and the transfer of multivalent elements in basaltic magmas. While peridotite xenoliths provide the most direct approach to measuring mantle oxidation states, such xenoliths in continental arcs are rare. In this investigation, we applied an alternative method, the determination of sulfur speciation in olivine-hosted melt inclusions and chromite-olivine oxygen barometry. We present a first attempt to spatially correlate oxygen fugacity relative to the subduction zone in a continental arc. The overall range in oxygen fugacity, based on sulfur speciation measurements, is from <-0.25 log units to +1.9 log units (ΔFMQ). Sulfur oxidation and the concentration of fluid-mobile trace elements both generally increase from backarc to forearc. This correlation is interpreted to reflect a progressively greater proportion of fluid-rich, oxidized subduction component closer to the trench. Estimates of the amount of subduction component (up to ~6 wt%) required to generate the geochemical diversity based on flux melt modeling correlate with oxygen fugacity, with the exception of calc-alkaline basalts, with high oxygen fugacity and greater proportion subduction component closer to the trench. Two other important observations to come from the flux melt modeling are 1) as basalt oxygen fugacity increases, calculated mantle temperature decreases, and 2) shoshonitic basalts require a depleted mantle source, distinct from the more enriched mantle source of the low-K tholeiite, calc-alkaline and ocean island-like magmas. The potential mantle source for shoshonitic basalts has a predicted oxygen fugacity from +0.3 to +2.4 log units (ΔFMQ) while the mantle source for low-K tholeiite, calc-alkaline and ocean island-like basalts may range from -1.1 to +0.7 log units (ΔFMQ), consistent with estimates of oxidation state for oceanic lithosphere. Therefore, despite the volatile and fluid-mobile trace element enrichment, the subarc mantle need not be significantly oxidized relative to unmodified oceanic lithosphere to generate the diversity in Cascade arc basaltic magmas.
V41D-0809
Petrological Characteristic of Recent Eruption Events at Galeras Volcano, Colombia
On-going volcanic activity at Galeras began in 1988, and major explosive eruption events occurred in 1993. Long- period seismic events had occurred before these events. In late 2004, explosive eruptive events resumed and intermittently continued by the present. Long-period events similar to those before the 1993 explosive eruptive events have been observed since early 2006. Evaluating potential of more explosive future eruptions becomes very important to minimize volcanic disasters in cities and towns around this volcano, including the city of Pasto. Investigation of temporal changes in petrological characteristics of eruption products makes us possible to understand the magma system undergone at Galeras. Whether has it changed (or developed) from the 1993 explosive events or not? Ballistics and scoria of vulcanian explosions during 2004-2006 and of the 1991 eruption were investigated in this paper. Rocks are two pyroxene andesite with various crystallinity in groundmass. Small amount of hornblende and olivine microphenocrsyts are involved. The whole rock chemistry hardly changed with time. Lines of petrological evidence suggest that magma mixing occurred throughout the eruption products during 1991-2006; 1) bimodal populations in core compositions of plagioclase phenocrysts, 2) plagioclase microlites with the composition between the two polulations, 3) plagioclase phenocrysts rims more enriched in Fe, and 4) reverse zoning of pyroxene phenocrysts that rather show single chemical population. Melt inclusions in pyroxene phenocrysts are slightly less evolved than the groundmass glass, suggesting that most pyroxenes were derived from felsic magma. These suggest mixing of low-temperature hydrous felsic magma with high-temperature anhydrous (pyroxene-free) mafic magma. Similarity in the petrographical characteristics and temperatures with the pyroxene geothermometry among all the samples shows that nearly constant mixing processes has been operated throughout the recent eruption events, including the 1991 eruption.
V41D-0810
Precursory Activity of the 2005 Eruption of Santa Ana Volcano, El Salvador.
After a period of unrest, Santa Ana (Illamatepec) volcano in El Salvador erupted suddenly on October 1st, 2005 at 1420 UTC (0820 local time), ejecting its acidic crater lake and generating a gas-and-ash plume ~10 km above the volcano. The short-lived eruption (~1 hr duration) deposited ballistics and ash up to 5m thick at the crater rim and depositing ash up to 40 km to the west. Underlying phreatomagmatic deposits exposed in the crater suggest that larger eruptions of this type are characteristic of recent historic activity. In this study, precursory activity to the 2005 eruption is investigated by analyzing physical and chemical parameters of the crater lake. Data has been compiled on water chemistry, temperature, and color of the lake from direct sampling and ground observations from 2004-2007. Lake water data suggests three phases of activity: (1) constant, well constrained activity from Jan. 2004 to Dec. 2004 showing SO4 ~10,000 ppm, Cl ~6000 ppm, and SO4/Cl ~1.6; (2) potential precursory activity from Jan. 2005 to Oct. 2005 expressed as a ramping up of SO4 to 11,625 ppm in May with a sudden decrease to 8250 ppm one month later, increased variability in Cl, and color change from dark coffee color to green in mid-September; and (3) post- eruption activity to present showing increasing Cl to a maximum of 22340 ppm, low SO4/Cl=0.38-0.8, an increase in temperature to 65.6 degrees C, and color change to yellowish-green). Analysis of high resolution satellite imagery from the ASTER sensor (15-90m/pixel) from 2000 to 2007 provides further information on lake size, temperature, and color. ASTER images show that the lake re-established itself further to the west after the eruption, drowning the adjacent high temperature fumarole field (max. 875 degrees C) which potentially contributed to the observed post-eruption changes in the lake. The combination of synoptic satellite-based remote sensing data with ground measurements will enhance the capabilities to recognize and interpret precursory activity of explosive eruptions at Santa Ana and similar volcanoes.
V41D-0811
Sulfur Yield of the 1600 Eruption of Huaynaputina Determined by Apatite Compositions
Aerosols from the VEI 6 eruption of Huaynaputina in 1600 have been implicated as a major factor in producing the coldest summers in the Northern Hemisphere in the last 500 years. We have previously estimated the stratospheric sulfur (S) input to range from 16 to 32 Mt based on ice core data, while others have estimated S in the magma to have been ~2 to 4 Mt. Another approach has suggested a fluid phase with ~24 to 51 Mt of S in equilibrium with magma. Here we report on S contents measured in apatite grains in the pumice from the plinian eruption. We invert these results to obtain independent estimates of the magmatic S contents and reassess the magnitude and source of the atmospheric loading. Apatite occurs as distinct microphenocrysts or as inclusions in biotite and amphibole ranging in size up to 200 microns. Most of the examples we analyzed exhibited basal sections. Microprobe analyses yielded apatite SO3 concentrations of 0.09 to 0.17 wt % with no systematic difference between microphenocrysts and inclusions. Multiple analyses of single grains demonstrated homogeneity. Residual glass contains SO3 below detection limits, <100 ppm, which suggests it was almost completely degassed. Other phase equilibria yield magmatic conditions of 840±21 °C and log fO2 of ~-11.5 (considerably less oxidized than recent anhydrite bearing dacites). Using these data, we calculate a total magmatic S content of ~4.1 Mt for the 9 km3 of magma erupted during the plinian eruption. Subtracting a conservative estimate of ~100 ppm S remaining dissolved in the melt, the total erupted sulfur is ~2.1 Mt. This value is almost identical to the range of erupted S calculated from experimental equilibria (~2 to 4 Mt) and supports the idea that much of the stratospheric sulfur input from Huaynaputina was from a fluid phase in equilibrium with the melt. Our work also demonstrates that weakly oxidized dacitic magmas that never stabilized anhydrite can have climatically significant atmospheric loading.
V41D-0812
Holocene processes and petrogenesis at a long-lived trachyandesitic magmatic system
Three major Holocene explosive eruptions of Hudson volcano in southern Chile (1991, 3.6ka BP, and 6.7ka BP) have produced a range of magma compositions from trachybasalt to trachydacite. The 1991 event generated both explosive and effusive eruptions of trachybasalt (phase 1, <1km3 DRE) and trachyandesite (phase 2, ~3km3 DRE). More evolved, compositionally homogeneous trachydacite was erupted during the 3.6ka BP event (~4km3). The 6.7ka BP eruption discharged trachyandesite, with some compositional diversity towards trachydacite, but generally less evolved than the 3.6ka BP magma. These Holocene events have consistently produced a mineral assemblage including plagioclase, clinopyroxene, orthopyroxene, Fe-Ti oxides, and trace apatite ± olivine, with total crystallinities varying from ~10-40%. Geochemical modeling suggests that the dominant trachyandesite erupted in 1991 and 6.7ka BP can be produced by fractional crystallization of a parental basaltic magma similar in composition to that produced during 1991 phase 1 eruption. Similarly, the more evolved trachydacite can be derived by fractional crystallization of the trachyandesitic composition. However, variations in magma compositions between and within the major plinian eruptions require a combination of fractional crystallization and magma mixing that is likely to take place in a relatively shallow magma storage region. Evidence for magma mixing includes coexisting glasses in single samples, complex and oscillatory zoning in plagioclase phenocrysts, and more evolved melt inclusions relative to co- existing matrix glasses. Melt inclusion volatile contents from the three eruptions, as estimated by the difference method, range from <1 to 5 wt%, and when combined with comparisons of experimental phase petrology indicate pre-eruptive storage at 1200-1800 bars pressure and temperature ranges of ~940 - 972°C (log fO2 -10.68 to -10.24). Geochemical and petrologic evidence suggests that a trachyandesite magma storage region is periodically recharged with more mafic magmas. This magma mixing is a likely cause for triggering of explosive eruptions at Hudson, as shown in particular by the 1991 eruption, and can lead to the simultaneous eruption of contrasting magmas at separate vents.
V41D-0813
Recharge as an Eruption Trigger Revealed at El Misti, Southern Peru
The most recent major explosive eruption of El Misti in southern Peru occurred ~2000 yr BP, producing significant pyroclastic deposits and extensive lahars. As such this eruption has considerable significance for assessing the hazard posed for the city of Arequipa (pop. ~800,000) built on the lahars of this eruption. Throughout its ~100 kyr history, eruptions from Misti have recorded recharge events into an open magmatic system. The 2000 BP eruption is no exception and contains a spectacular record that suggests that recharge may have triggered the eruption. Juvenile blocks from the 2000 BP eruption reveal the intimate mingling of two magmas; amphibole-plagioclase rhyolite and amphibole-plagioclase andesite. Glass compositions range from ~67 wt.% SiO2 in the andesite to ~75 wt.% SiO2 in the rhyolite. Phase equilibria yield temperatures of 780±30°C in the rhyolite and 960±40°C for the andesite. Viscosity estimates yield 108 and 106 Pa s, respectively. Mingling at the cm scale is manifested in distinct bands and selvages that are linear or folded, while at the mm scale variably convoluted ribbons, wisps, and threads attest to strong viscosity contrasts between the magmas. The record in plagioclase phenocrysts in both phases is extremely complicated with eight different textural types in the rhyolite and four different types in the andesite. Plagioclase phenocrysts from the andesite have core compositions of ~An80, are normally zoned with rims to ~An60, but are texturally complex. Plagioclase phenocryst compositions in the rhyolite are much more complex; some are similar to those in the andesite, others are more typical of those that would be expected in a rhyolitic melt with cores of ~An55 to rims as low as An40. Microlites in the andesite are non-zoned and have compositions similar to rims of the phenocrysts. Microlites in the rhyolite are normally zoned, and define two populations; one with An60 cores and An50 rims, and the other with An40 cores and An30 rims, similar to the phenocrysts. The record in amphibole phenocrysts is simpler. Those in the rhyolite have dehydration reaction rims whereas the andesite-hosted amphiboles do not. Compositions of amphiboles in the two phases are broadly similar although the andesite does have a population with higher Mg# and lower AlIV. We interpret the plagioclase record to represent an integrated history extending back to pre-2000 BP, whereas the amphibole and microlite data to be revealing of the 2000 BP event. We propose that a dike of hotter, less viscous andesite intruded into a stagnant or semi-solid mass of rhyolitic composition material and its associated phenocryst suite. By virtue of its lower viscosity and momentum, the andesite forced its way through the rhyolite with little interaction except for local exchange through viscous coupling at the margins of the dike. Dehydration of amphiboles in the rhyolites was induced and may have contributed volatiles to the mixture. The microlites record decompression-induced growth during ascent of the magmas. The distinct composition of residual glass in the two phases demonstrates a lack of wholesale thermal and compositional equilibration suggesting that the mingling occurred a very short time before the subsequent eruption. We interpret the record in the 2000 BP eruption of El Misti as one where mafic recharge triggered the eruption and this may serve as a model for the recurrent explosive eruptions at this volcano.
V41D-0814
Hyperacid volcano-hydrothermal fluids from Copahue volcano, Argentina: Analogs for "subduction zone fluids"?
Hyperacid concentrated Chlorine-Sulfate brines occur in many young arc volcanoes, with pH values <1, high concentrations of volcanogenic elements (S, Cl, F, As, B) and the main rock forming elements (Ca, Al, Mg, K, Na, P). Sulfur isotope data and Silica thermometry from such fluids sampled over a ten year period from the Copahue volcanic system (Argentina) suggest reservoir temperatures of 175-300 oC, whereas the surface fluids do not exceed local boiling temperatures. These fluids are generated at much lower P-T conditions than fluids associated with a dehydrating subducted sediment complex below arc volcanoes, but their fundamental chemical compositions may have similarities. Incompatible trace element, major element concentrations and Pb isotope compositions of the fluids were used to determine the most likely rock protoliths for these fluids. Mean rock- normalized trace element diagrams then indicate which elements are quantitatively extracted from the rocks and which are left behind or precipitated in secondary phases. Most LILE show flat rock-normalized patterns, indicating close to congruent dissolution, whereas Ta-Nb-Ti show strong depletions in the rock-normalized diagrams. These HFSE are either left behind in the altered rock protolith or were precipitated along the way up. The behavior of U and Th is almost identical, suggesting that in these low pH fluids with abundant ligands Th is just as easily transported as U, which is not the case in more dilute, neutral fluids. Most analyzed fluids have steeper LREE patterns than the rocks and have negative Eu anomalies similar to the rocks. Fluids that interacted with newly intruded magma e.g., during the 2000 eruption, have much less pronounced Eu anomalies, which was most likely caused by the preferential dissolution of plagioclase when newly intruded magma interacted with the acid fluids. The fluids show a strong positive correlation between Y and Cd (similar to MORB basalts, Yi et al., JGR, 2000), suggesting that Cd is mainly a rock-derived element that may not show chalcophilic behavior. The fluids are strongly enriched (relative to rock) in As, Zn and Pb, suggesting that these elements were carried with the volcanic gas phase into the system. In summary, if these fluids are broadly similar to fluids from dehydrating subducted sediments, they tend to transport preferently the LILE, LREE, U as well as Th, while the HFSE are left behind.
V41D-0815
Geochemical Evidence for 1.9-1.8 Ga Continental Arc Magmatism in the Arequipa Massif, Southwestern Peru
The anomalous Arequipa Massif, along the coast of southwestern Peru, is part of an exotic terrane (The Arequipa- Antofalla Basement) that was likely emplaced along the margin of Amazonia at ca. 1 Ga. As such, it may be a tectonic tracer left behind by the craton that collided with Amazonia during the ca. 1 Ga Sunsas Orogeny. Correlation of the massif with its "parent craton" could constrain the position of Amazonia within the Mesoproterozoic supercontinent Rodinia. Better characterization of the "pre-1 Ga" tectonic evolution of the Arequipa Massif is necessary to evaluate potential parent cratons. Previous geochronologic work has constrained the timing of Paleoproterozoic events. Bimodal magmatism occurred between 2.02-1.82 Ga. This granite and gabbro was metamorphosed between 1.82 and 1.79 Ga. The resulting gneisses were intruded by granite at ca. 1.79 Ga. The entire sequence was metamorphosed ca. 1 Ga, generating a planar fabric in the younger granite. Based on these ages, there are numerous potential parent cratons, including Amazonia, Sao Francisco/Congo Craton, Kalahari, Laurentia, and Baltica (although previous Pb isotopic work refuted potential correlation with Laurentia). New elemental analyses from the same samples now provide better characterization of the tectonic environment in which these rocks formed. Samples include three granitic gneisses, two metagabbroic amphibolites, and three foliated megacrystic granites. The six granitic rocks have calc-alkaline compositions and the two mafic rocks are alkaline-rich. Among the granitic samples, comparison of Ta/Yb vs Th/Yb is consistent with formation in an "active continental margin" and comparison of Th-Hf/3-Ta (for all samples), Rb vs. Y+Nb, Ta vs. Nb suggest formation by volcanic arc magmatism. Although the rocks have been metamorphosed, multiple geochemical indicators are consistent with formation during continental arc magmatism.
V41D-0816
Insights Into the Formation of Deep Hydrothermal Quartz From the Porphyry-Copper- Molybdenum Deposit at Butte, Montana
We combine SEM-cathodoluminescence (SEM-CL) images of magmatic and hydrothermal quartz with trace elements and quartz precipitation temperatures to elucidate the formation of deep quartz veins in a porphyry-style deposit. Trace elements (Ti, K, Al, and Fe) were measured by EPMA along traverses crossing CL textural boundaries and quartz precipitation temperatures were calculated using the Ti in quartz (TitaniQ) geothermometer (Wark and Watson, 2006). We examined Butte granite, quartz porphyry, and five deep vein types, including biotite crackles and early dark micaceous (EDM) veins (potassic alteration), barren quartz/quartz-molybdenum veins lacking alteration, and pyrite-quartz veins with sericitic alteration. Magmatic quartz shows concentric zoning that is cross-cut by CL-dark quartz veins. Complex textures in hydrothermal quartz indicate multiple quartz-precipitation episodes corresponding to different physical conditions of quartz growth. Concentrations of Ti have a strong positive correlation with CL brightness. K and Al concentrations show a weak relationship with CL brightness but they generally vary in unison. Fe concentrations do not appear to correlate with CL brightness or concentrations of other trace elements. TitaniQ temperatures range from 710 to 730°C in plutonic quartz, 625 to 750°C in porphyry quartz, and 650 to 730°C in barren quartz veins, overlapping with magmatic quartz. Biotite crackles and EDM vein temperatures are generally cooler than magmatic quartz and barren quartz veins, ranging from 660°C to less than 480°C, a temperature limited by our Ti detection limit. SEM-CL brightness boundaries do not necessarily match quartz grain boundaries. Careful examination of these boundaries in comparison with our existing trace element data, along with new electron backscatter diffraction (EBSD) mapping will help clarify the roles of diffusion, dissolution, and recrystallization in forming the CL textures. TitaniQ geothermometer results show that magmatic and deep hydrothermal temperature regimes overlap considerably. Temperatures combined with SEM-CL textures within one sample indicate significant temperature fluctuations: an increase from 560 to 730°C then decrease to 590°C between episodes of vein formation.
V41D-0817
Experimental Study of the Partitioning of Au, Ag, Mo, W, Zn, Mn, and Ti among Pyrrhotite and Immiscible Fe-S-O and Rhyolitic Melts
Porphyry and related ore deposits are associated in space and time with shallow level intrusions. Identifying mechanisms of chemical enrichment or depletion during the events preceding shallow emplacement of an evolved arc magma may allow for better exploration for economically viable deposits. The effect of ore metal partitioning amongst a rhyolitic melt (sm), pyrrhotite (po), magnetite (mt), and an immiscible Fe-S-O melt (Fe-S-O) has been evaluated in this experimental study for Au, Ag, Mo, W, Zn, Mn and Ti. Capsules were loaded with mt, po, sm, and either Mo + W, or Au + Ag. Other metals (Zn, Mn, and Ti) were introduced with the natural starting materials. Experiments were run at 1036oC (Au, Ag) or 1042oC (Mo, W) in sealed silica tubes, for durations up to 2000 minutes. The coexistence of mt and po allows for the determination of oxygen and sulfur fugacities. fO2 was between the FMQ and NNO buffers, and log fS2 ~ -1 bar. Run products were analyzed by using EMPA and LA-ICP-MS. Experiments yielded a partition coefficient, D(po/sm)±1σ(SDM) =120±50, for Au, which compares well with the value determined by Jugo et al., 1999 (140±40) for hydrous rhyolitic melts in equilibrium with po at 850°C and 100 MPa. The experiments also yielded D(po/sm) and D(Fe-S-O/sm), respectively for Ag 58±8, 120±20; Mo 35±3, 90±10; W 1.2±0.6x10-3, 9±3; Zn 3.4±0.4, 10±1; Mn 1.1±0.1, 1.9±0.3; and Ti 0.030±0.002, 0.28±0.05. During the ascent of magma through the crust, crystal fractionation will remove these elements from the melt to different extents. Assuming po constitutes 0.1wt% of the crystalline assemblage, and the initial magma fractionates 90%, pyrrhotite crystallization can result in removal of: Au by 24%, Ag by 13%, Mo by 8%; W, Zn, Mn and Ti would remain largely unchanged. Small blebs of Fe-S-O melts may also be fractionated at higher temperatures, removing up to 24% Ag and 19% Mo, again leaving W, Zn, Mn and Ti largely unaffected. The metals removed by fractionation would be unavailable for ore formation; sub-economic deposits or the absence of significant geochemical anomalies could be the end result. Crystal fractionation of po may be augmented by the addition of sulfide to the magma by assimilation of reduced sulfide-bearing sedimentary rock. This process may lead to excess po (or, at high enough temperatures, Fe-S-O) precipitation and increased ore metal sequestration, further removing ore metals from the magmatic system, and thereby reducing the probability of porphyry-ore formation by later, high-level, magmatic-hydrothermal processes.
V41D-0818
Lead Isotope Constraints on the Sources of Ore Metals in SW Mexican Deposits
Lead isotope ratios from mineral deposits in southern Mexico increase with distance from the trench from 206Pb/204Pb values between 18.597 and 18.650 in the coastal area to values between 18.712 and 19.069 approximately 800 km east from the trench. This variation has been attributed to increasing assimilation of radiogenic lead from the crust with increasing distance from the trench. New sampling was undertaken in this area to provide a clearer picture of the potential sources of ore metals in this arc system, and also, if possible, to examine whether ore metal sources differ among the proposed tectonostratigraphic exotic terranes of southern Mexico. New TIMS lead isotope analyses are presented for samples from the metamorphic basement rocks of the Guerrero Terrane, the Late Cretaceous clastic sedimentary rocks from the Upper Mesozoic Assemblage, and for mid-Cretaceous igneous rocks, as well as for samples from the Oligocene La Verde, Esmeralda, and El Malacate copper prospects. Whole rock samples of schist from the Jurassic-Cretaceous Arteaga Complex and phyllite and slate from the Tierra Caliente Complex contain radiogenic lead relative to bulk earth models, with 206Pb/204Pb ranging from 18.981-19.256. These values are substantially more radiogenic than published values of analyses of metagabbro and charnockite from the Grenvillian-age Oaxaca Terrane. Sedimentary rocks (sandstones, siltstones, and marls) belonging to the Huetamo Sequence have 206Pb/204Pb values ranging between 18.630 to 18.998, close to the published data for the sediments from IPOD-DSDP Sites 487 and 488, Cocos Plate. Whole rock analyses of igneous rocks (granodiorite) collected from La Verde and El Malacate have 206Pb/204Pb ranging from 18.764 to 18.989, clustering between the fields represented by the sedimentary and the metamorphic rocks, suggesting assimilation of lead from these components. Ore samples from La Verde and Esmeralda have 206Pb/204Pb between 18.685 and 18.731 and plot within the field defined by the sedimentary rocks. Whereas the metamorphic and igneous rocks analyzed so far plot generally above the Stacey-Kramers (1975) reference line, the ores and sedimentary rocks define a narrow field just below this line. The distribution of data from these ores on co-variation diagrams also appears to define a steep array, suggesting a MORB-EPR source contaminated with a 207Pb- and 208Pb-rich component (possibly the metamorphic basement or a sedimentary component). The analyzed sedimentary rocks are less radiogenic than the metamorphic basement, suggesting they are not simply derived from the basement and that other rocks were involved in their provenance.
V41D-0819
Experimental Study of the Behaviour of Gold in Calc-alkaline Arc Magmas : Demonstration of the Effect of Sulphur and Geological Implications on the Formation of Gold Deposits
Copper-gold-molybdenum deposits represent major metal resources known to be spatially and temporally associated with intrusive arc magmatism (Hedenquist and Lowenstern, 1994). Although the bulk of the ore metals seem to originate from the magmas, there is as yet no consensus on processes responsible for their concentration. The fact that primary ore minerals are predominantly sulfides has led to the suggestion that sulfur may play an important role in metal enrichment processes at the magmatic stage (Sillitoe, 1997; Mungall, 2002). Previous experimental studies have shown that sulfides can largely fractionate gold from the melt, and therefore concluded that oxidized magmas are more likely to be Au-enriched. Here, we conducted high-pressure temperature experiments (4 Kbar, 1000°C) on three dacitic and two rhyolitic natural magmas from the Philippines (both adakitic and typical calc-alkaline compositions were used) over a range of fO2 corresponding to reducing (~NNO-1), moderately oxidizing (~NNO+1.5) and strongly oxidizing (~NNO+3) conditions as measured by solid Ni-Pd-O sensors. The experiments were carried out in gold capsules, the latter also serving as the source of gold, in the presence of variable amounts of H2O. The water content of experimental glasses was determined by Karl-Fisher Titration. Both S-free and S-bearing (~1 wt% S added) experiments were performed. Sulfur concentrations in experimental glasses were measured by electron microprobe, and thermodynamic calculations were used to compute the fS2 of the experimental charges. Gold concentrations in glasses were determined by LA-ICP-MS. Charges consist of dominant silicate melt quenched to glass plus minor silicate phases (mostly Cpx, rarely Opx), together with discrete magnetites in the most oxidized charges. All S-bearing charges are saturated with either sulfides or sulfates, together with discrete Au-Fe-S alloys, depending on the fO2. Our data show that gold solubility in S-free charges is low (30-240 ppb) but globally increases with fO2, suggesting the implication of oxidized species in the dissolution of gold. Gold solubility in S-bearing charges appears much higher, ranging from ~500 to 5200 ppb. Gold content is higher in reduced than in moderately oxidizing conditions, and is not affected by melt composition. In very oxidizing conditions, Au solubility is low and nearly constant, and does not seem affected by the presence of S ; it only shows a dependence on fO2 by following the same trend than S-free charges. At ~NNO+1.5, gold content of silicate melt appears to be the result of a complex competition between fO2 and fS2, whereas under more reduced conditions, the effect of fS2 appears the strongest, leading to a direct dependence of gold solubility on sulfur content. This study shows conclusively that Au transport and concentration in silicic melts can be dramatically enhanced by the presence of sulphur. A major consequence is that sulfide crystallization is not likely to scavenge all Au present in the magma, but may allow gold-enriched residual magmas, especially those that are moderately reduced (NNO-1 to NNO+1.5), to be emplaced in the upper crust. Gold enrichment in porphyry- and epithermal- type deposits can be directly linked with sulfur incorporation and abundances in magmas.
V41D-0820
Ancient mantle contribution to boninitic magma: evidence from Os isotopic compositions of Cr-spinel sands of boninites in Izu-Bonin arc
Osmium (Os) isotope is a sensitive tracer of crust and sediment, because of significant contrast between a high Os isotope ratio of crust and sediments (187Os/188Os>0.5) and a low ratio of mantle (187Os/188Os<0.15). Therefore, Os isotope provides information on slab component contribution to the mantle source of island arc lavas. Radiogenic Os isotopic compositions are commonly found in volcanic lavas (Alves S., Schiano P., Capmas F. & Allegre C. J., (2002) Earth Planet. Sci. Lett. 198, 355-369)and peridotite xenoliths (Brandon A. D., Becker, H., Carlson, R. W. & Shirey S. B., (1999), Chem. Geol. 160, 387-407) in a subduction zone setting. Though most of the authors had attributed these high 187Os/188Os to the input of the slab component to the mantle source, some pointed out that assimilation of crustal materials during magma ascent possibly reproduce the elevated Os isotope ratios of the arc lavas (Lassiter J. C. & Luhr, J. F., (2001), Geochem. Geophys. Geosyst. 2, 2000GC000116). Since then, whether the high Os isotope ratios of arc lavas are caused by contribution of slab component or by assimilation of the overlying crust has long been highly debated (Woodhead J. & Brauns M., (2004), Earth Planet. Sci. Lett. 221, 309-323). Here we report the unradiogenic Os isotopic ratios (187Os/188Os < 0.125) of Cr-spinel sands from Chichi-Jima and Yome-Shima boninites, Izu-Bonin arc. As Cr-spinel is resistant to later alteration and weathering and, more importantly, is the early stage crystal in the fractional crystallization, it preserves the chamical and isotopic compositions of very primitive magma in its melt inclusion and spinel itself without any later stage crustal contamination. The obtained Os isotopic compositions are the lowest among the previously reported 187Os/188Os of arc magma. The unradiogenic Os isotopic compositions of boninite Cr-spinels obtained in this study are in the range of anciently depleted lithospheric mantle (Parkinson I.L., Hawkesworth C.J. and Cohen A.S. (1998) Science 281, 2011-2013) which strongly suggests that such ancient mantle contribute significantly to boninite magma generation.
V41D-0821
Multiple Sources of K-rich Melts in Central Italy: Evidence From Trace-element and Pb- isotopic Signatures of Melt Inclusions in a Single Lava Flow From Latera Volcano
Latera stratovolcano is part of Vulsini, the northernmost volcanic complex of the Roman Province, Central Italy. Lava compositions range between strongly silica-undersaturated leucite-bearing High-K (HKS) and near silica- saturated Low-K (KS) products. We analyzed homogenized melt inclusions trapped in primitive olivines (Fo=91- 87) from a single KS lava for major and trace-element compositions and Pb-isotope ratios. A wide range in CaO contents of high-Mg# olivine and clinopyroxene phenocrysts points to crystallization from a diversity of primary melts, as observed elsewhere in Roman Province HKS (>0.3 wt.%) and KS (<0.3 wt.%). Some olivines have low CaO contents (<0.15 wt.%), similar to those of Tuscan lamproites (LMP). Based on the combination of major and trace element contents of MI and mineral chemistry, the lava appears to host a collection of different alkali-rich melts: KS-type compositionally similar to the host rock, HKS-type close to lavas of the surrounding Roman Province, a high-SiO2, low-CaO, low-Na2O melt type close to the lamproite compositions, and high-Na2O, low-CaO melt which have no equivalent in nearby erupted lavas. Melt inclusions were further analyzed for Pb isotope ratios, using the Nordsim Cameca-1270 ion microprobe (Swedish Museum of Natural History, Stockholm). A clear relation was found between Pb concentrations of MI and analytical uncertainty in Pb-isotope ratios, with a strong increase in error below 10 ppm. For MI with >10 ppm, 2-sigma errors were <0.003 for 207Pb/206Pb, <0.007 for 208Pb/206Pb, <0.25 for 206Pb/204Pb, <0.2 for 207Pb/204Pb and <0.5 for 208Pb/204Pb. The 25 MI of Latera analyzed contain up to 130 ppm Pb with 75% higher than 10 ppm. The MI show an extreme Pb-isotopic diversity (e.g., 207Pb/206Pb=0.815-0.86, 208Pb/206Pb=2.03-2.11, 206Pb/204Pb=18-19.2, 207Pb/204Pb=14.6-16, 208Pb/204Pb=36.5-40.5), with each group being characterized by its own signature. The Pb-isotope composition of the host lava appears to be a mixture of (at least) three isotopically distinct end- members. The melts form arrays between MORB-like mantle and different crustal/sedimentary end-members. Subduction of the Adriatic plate was probably responsible for the introduction of the one that affected other Roman Province sources as well. Different geodynamic scenarios can be invoked for the introduction of additional heterogeneity in the Latera mantle. The Pb isotopes of the high-Na2O, low-CaO group are consistent with either a contribution of deeply subducted crustal components from the Southern Alps, as has been proposed for the LMP source, or with involvement of local lower-crust lithologies. A recent thermal event must have been responsible for the simultaneous melting of the strongly heterogeneous mantle column below the volcano.
V41D-0822
Implications for the Indosinian Orogeny and closure of eastern Paleo-Tethys: A perspective from the Song Ma suture
The eastern paleo-Tethys was consumed by amalgamation/collision between the South China, Indochina and Simbumasu blocks, a process also widely considered responsible for the Indosinian Orogeny. Here we report a detailed study of ophiolitic and associated granitic rocks from the Song Ma suture between the South China and Indochina blocks. The ophiolite consists of mafic (now exposed as amphibolites) and ultramafic (as serpentinites) members. The mafic rocks show basaltic compositions, marked with depleted light rare earth elements [(La/Sm)N= 0.44 to 0.99] and high Nd isotope ratios [εNd (0)= +10 to +6], and without apparent depletion in the high field strength elements (Nb, Ta and Ti), thus representing the Paleo-Tethyan oceanic crust. Besides, the so-called plagiogranites previously reported from the Chieng Khoung complex, now exposed in the Song Ma belt, are actually felsic rocks with adakitic geochemical signatures. These Chieng Khoung rocks, dated at 255 +/- 2 Ma by single zircon U-Pb method, exhibit low Nd [εNd(0)= -1 to -2] and relative high Sr (ISr= 0.7050 to 0.7056) isotope ratios that, together with trace element constraints, imply an origin from partial melting of garnet-amphibolites in a thickened lower continental crust above the mantle wedge related to the Paleo-Tethyan subduction. In addition to the adakitic rocks in the Chieng Khoung complex, this subduction resulted in a late Permian-early Triassic (~274-240 Ma) magmatic arc that extends south of and parallel to the Song Ma suture through central Vietnam. Our study is consistent with the notion that the continental collision between the South China and Indochina blocks along the Song Ma suture took place around 245 Ma, thus reiterating the issue of existence of an early phase of the Indosinian Orogeny.
V41D-0823
Late Cretaceous volcanic activity in the Korean Peninsula
Southern part of the Korean peninsula was one of the largest Cretaceous volcanic fields in the northwest Pacific. Volcanic activity began with the eruption of silicic and minor basaltic volcanic rocks which were intercalated into sedimentary rocks of the Sindong Group and the Hayang Group. After this minor activity, large volumes of volcanic rocks were erupted and the resultant Yucheon Group consists mainly of volcanic deposits. Volcanic activity in the Gyeongsang Basin can be divided into four main stages; (1) early minor silicic and alkali-basaltic volcanism, (2) the Chaeyaksan alkali-basalt volcanism, (3) the Jusasan andesitic volcanism, and (4) the Unmunsa rhyolitic volcanism. Cretaceous volcanic rocks outside the Gyeongsang Basin are associated with volcano-sedimentary pull-apart basins or troughs which developed along a series of NE-trending sinistral strike-slip faults. The Cretaceous volcanic rocks distributed both in and outside the Gyeongsang Basin are composed of basalt, basaltic andesite, andesite, dacite and rhyolite. Their geochemical trends show successive differentiation processes of B-A-R associations. The volcanic rocks gradually changed in composition from alkalic in the early stages through alkalic/calc-alkalic in the middle stages to calc-alkalic in late stages, and the volcanic activity became more explosive with time. The trace element composition and REE patterns of the volcanis rocks, which are characterized by a high LILE/HFSE ratio and enrichment in LREE, suggest that they are typical of Andean-type continental margin-arc calc-alkaline volcanic rocks produced in a subduction environment. The initial 87Sr/86Sr ratios of volcanic rocks from the Gyeongsang Basin were progressively lowered with time from 0.7061 to 0.7049. The whole rock chemistry and Sr isotopic composition of the Jusasan and Unmunsa volcanis rocks suggest that the basaltic andesite/andesite evolved mainly by fractional crystallization from basaltic primary magma which produced by partial melting of the mantle wedge. Dacite/rhyolite magma was mostly produced by fractional crystallization including plagioclase feldspar but it might have partly undergone crustal assimilation during the fractionation in a high level magma chamber.
V41D-0824
Age and Geochemical Characteristics of the Linzizong Volcanism for Neotethyan Slab Breakoff in Southern Tibet
Formation of the Linzizong volcanic successions in the Lhasa terrane, southern Tibet has long been related to northward subduction of the Neotethyan oceanic slab under Asia. Here we report new 40Ar/39Ar age results for the volcanic rocks recovered from a large area (29°N to 32°N and 85°E to 93°E) that, together with literature data, delineate two discrete stages of volcanism. These are a widespread Cretaceous stage and an intense, but spatially confined, Paleogene stage. The latter, occurring only in the southern part of the Lhasa terrane, resulted in the Linzizong volcanic successions. Our data, furthermore, suggest southward migration and intensification of the volcanism in the Lhasa terrane with magmatic ¡§flare- ups¡¨ at ca. 50 Ma. While the volcanic successions consist dominantly of calc-alkaline rocks typical of arc lava geochemistry, those formed during the flare-up period show significant compositional variations from low-K tholeiitic through calc-alkaline to shoshonitic magma suites. These observations enable us to interpret the volcanic southward migration and following flare-up as the consequences of rollback and breakoff of the subducted Neotethyan slab that occurred ahead and in the early stage, respectively, of the India-Asia collision. Our interpretation that involves a major Eocene tectonomagmatic activity, and concomitant topographic uplift, in southern Tibet is consistent with Himalayan metamorphic constraints, regional sedimentary records and seismic tomography.
V41D-0825
Geochemical characteristics of the volcanic rocks from the Tokara Islands, Ryukyu volcanic arc, Japan.
The tectonic setting of the Ryukyu volcanic arc extended from the southern Kyushu to Ioutori-shima is changing from continental margin to island arc. Most of the oceanic islands in the Tokara Islands are Quaternary volcanoes and some of them have eruption recodes in historical time. The volume of each volcanic island above sea level is less than 10% of the total volume estimated from the precise bathymetric map. The volcanic front of the Ryukyu arc is getting closer to Okinawa Trough at the southern part of the arc indicated in a precise bathymetric map. In order to understand the geochemical variation of the volcanic rocks located in the toransitional setting, we have analyzed volcanic rocks collected from entire volcanic arc and some Okinawa Trough. The samples are included not only onland but also dredge haul samples obtaied during the cruises KT00-15, KT07-2 and KT07-21. Analyzed volcanic rocks are ranging from basalt to rhyolite. The most abundant rock samples are andesite (SiO2= 57 wt.% and MgO = less than 6wt.%). Some of the dredged rhyolites are poorly vesiculated. The volcanic rocks, except for Kuchino-shima, in which volcanic rocks have abundant hornblende phenocryst, are mainly composed of plagioclase, augite and hypersthene. Basaltic rocks are only found dredged sample from Yokoate-jima and are rich in Al2O3 (~ 18 wt. %) due to the large amount of plagioclase phenocrysts. Most of volcanic rocks are plotted in the medium-K rock series, whereas samples from Ioutori-shima and dredged samples from Daini-amami bank and Kuchino-shima are plotted in the low-K rock series. Chondrite normalized REE patterns in the volcanic rocks are slightly enriched in light REE (La/Yb = ~3.6). They does not indicate remarkable changes in the REE patterns with increasing silica. La/K2O ratios, which are indicated LILE / HFSE, define two distinct trends. Some dredged samples including Okinawa Trough are higher ratio (La/ K2O = ~9), whereas volcanic rocks of the Tokara Islands are relatively low (La/K2O = 5.7). This ratio expected to indicate relative abundance of volatile components in the source magma. Therefore, volcanic rocks from the Tokara Islands are derived from water rich magma and those of the submarine samples including Okinawa Trough are derived relatively dry magma. Chemical variations of the volcanic rock dose not indicate systematic spatial distribution along the arc. The wet magma, which represents the Tokara volcanic islands, may be coexisting with dry magma, which represents volcanic rocks of Okinawa Trough, simultaneously even in the same volcano.
V41D-0826
A Melt Inclusion Study of the Middle and Oldest Toba Tuffs, Sumatra, Indonesia: Initial Results
The Toba Caldera Complex in northern Sumatra was the site of 4 caldera forming eruptions in the past 1.2 m.y. Tuffs erupted at 1.2, 0.840, 0.501, and 0.074 Ma are known as the Haranggoal Dacite Tuff (HDT), Oldest Toba Tuff (OTT), Middle Toba Tuff (MTT), and Youngest Toba Tuff (YTT) respectively. A previous study determined the pre- eruptive volatile contents of the 2800 km3 YTT using melt inclusions in quartz crystals. In an effort to characterize the evolution of the Toba magma system, we recently studied melt inclusions from the MTT and OTT. Quartz crystals from 3 samples that spanned the compositional ranges of each tuff were chosen for study. In the MTT, melt inclusions were rhyolitic (73-77 wt. % SiO2) and overlapped with the compositional range of the bulk rock samples. Melt inclusions in the OTT were also rhyolitic (74-77 wt. % SiO2) but were far more silicic than their bulk rock samples (69-73 wt. % SiO2). In contrast to the MTT, the most silicic melt inclusions in the OTT were found in the least evolved rock samples. Water contents determined by FTIR in the MTT range from 3.0-5.5 wt. %, while those in the OTT are 2.0-5.5 wt. %. Melt inclusions with the lowest water contents are found in the least evolved rock samples from both units. Both MTT and OTT inclusions contained about 20-120 ppm CO2. In the OTT, inclusions with the highest CO2 contents had the least silicic glass composition, but were from the most evolved bulk rock samples. Cl contents of the melt inclusions range from 1200-1800 ppm in the MTT and 1000-2000 ppm in OTT. Considerable Cl degassing is suggested by matrix glass analyses with distinctly lower Cl contents. S contents are low (<20 ppm) and near detection limits in melt inclusions from both units. Comparing MTT and OTT melt inclusions with those of the YTT indicates overlapping and indistinguishable major element geochemistry. The least evolved bulk rock samples from both the OTT and YTT contain the most evolved melt inclusions, unlike the MTT. Water contents of the MTT and OTT have a wider range than the YTT (4.0-5.5 wt. %) and gradients are recorded in the OTT and MTT magma chambers, whereas none is evident in the YTT magma. Dissolved S contents were low in all 3 magmas prior to their eruption.