V31B-2128
Multi-stage water enrichment processes in the wake of the Farralon slab (UNESCO IGCP 557)
In this study we supply petrological evidence for a hydrous upwelling beneath the Colorado Plateau in the Western United States. We trace a hydrous upwelling by its signature of multi-stage water enrichment processes in garnet- and Cr-spinel bearing lherzolithes collected from the sample volcano "the Thumb". Multi-stage hydration is witnessed from a depth of 150 km to just below Moho depth by various constraints. We use high-resolution synchrotron based FTIR, to resolve the hydration at 150 km depth (T= 1200°C) by mapping the water content in homogenous olivine crystals, around spinel inclusions and fully embedded cracks from lherzolites samples. We interpret these lherzolites to originate from the lower part of the Farallon slab. These xenoliths are presumably released during break off of the slab. In these samples background hydration in the homogenous olivine is 140 ppm H2O wt but around 2-D defects the water content can rise up to 800 ppm H20 wt. A second set of samples of harzburgitic composition is interpreted to be derived from the subcontinental Mantle (T= 600°C). In these samples hydration is visible through cracks that are serpentinized. From this observation we conclude that hydration causes multi-stage enrichment of the mantle wedge through a process that is dominated by the growth of ubiquitous crack. Our results do not confirm the hypothesis of pervasive hydration through point defect diffusion. Cracks in the mantle are a key element in the subduction zone water cycle.
V31B-2129
Sr, Nd, Pb and Os Isotopic Compositions of Lavas From the Mount Baker Volcanic Field, Cascade Arc
We present the results of a trace element and Sr, Nd, Pb and Os isotopic study of the Mt. Baker volcanic field (MBVF), part of the northern segment of the Cascade magmatic arc known as the Garibaldi Belt. To date, only 4 Sr isotopic ratios (all from the Sulphur Creek flow) have been published. The Mount Baker volcanic field extends to 3.72 Ma and a case can be made for continuous magmatic activity in this region extending from 34 Ma to present. Our goal is to use isotope ratios to characterize the mantle source regions that underlie the Garibaldi Belt, to document the chemical inputs of slab fluid/melt, sediment, and lower crust, and to assess temporal and spatial variations in these factors. We measured 29 Sr and Nd isotopic ratios, 8 Pb isotopic ratios, and 9 Os isotopic ratios, representing the full age range and compositional diversity (calc- alkaline basalt through rhyolite) of the MBVF, including all known MBVF basalts. A 22.86-Ma gabbronorite from the adjacent Chilliwack batholith was analyzed as an analog for the modern mafic lower crust. All Mt. Baker lavas are calc-alkaline with the arc-characteristic signatures of HFSE depletion and LILE enrichment. MBVF 87Sr/86Sr values (0.703932 to 0.703057) and εNd (+4.71 to +7.79) are well correlated and lie within the mantle array. Mt. Baker Sr and Nd data are indistinguishable from other Garibaldi belt lavas (Green & Harry 1999, Green & Sinha 2005), and also overlap data from the neighboring Chilliwack batholith (Tepper 1996; Tepper et al. 1993). In contrast, central and southern Cascade arc lavas with similar Sr ratios have corresponding εNd values that are lower by ~2 epsilon units. The Garibaldi Belt and Chilliwack magmas are tapping a mantle source distinct from that of the rest of the Cascade arc. 206Pb/204Pb, 207Pb/204Pb and 208Pb/204Pb ratios of MBVF basalts plot close to the NHRL, in a linear trend between Juan de Fuca MORB and Pacific sediment, indicating a sediment contribution to the MBVF magmas. With the exception of the Sulphur Creek flow, Pb isotopic values are well correlated with Nd and Sr. (Sr/P)N, commonly used as an indicator of a slab fluid-derived component, ranges from 1.4 to 5 but correlates poorly with Sr, Nd and Pb isotopic ratios. The similarity of isotopic and trace element data among Mt. Baker and Chilliwack magmas is consistent with the hypothesis that magmas of the MBVF represent a modern continuation of the Chilliwack magmatic system. Mafic and felsic MBVF rocks are indistinguishable in their Sr and Nd isotopic compositions, most likely because the lower crust beneath Mt. Baker is primarily Cenozoic in age and mafic in composition (the root of a long-lived arc system). In an attempt to 'see through' this problem, we measured Os isotopic ratios on whole rocks and mineral separates as only a few million years are required for crustal 187Os/188Os ratios to evolve from mantle values. MBVF basalt γOs values range from +70 to +522. The most primitive sample is Table Mountain andesite (γOs 65). The γOs values and Os concentrations (1.56 to 17.7 ppt) do not correlate with any major or trace element trends or with Sr, Nd, or Pb isotope ratios. Given that primitive arc mantle-derived magmas have γOs values of ~10, we hypothesize that Cenozoic lower crust has contributed substantially to the modern Mount Baker magmas via assimilation, melting and mixing. It is likely that Cenozoic lower crust would have variable Os isotopic compositions but would be virtually homogeneous in terms of Sr, Nd, and Pb isotopic ratios.
V31B-2130
Mafic Lava Shields in the Central Cascade Arc: Tectonic Implications of Late Miocene to Early Pliocene Forearc HAOT Volcanism at Snow Peak, Oregon
Study of mafic magmatism within subduction zones provides important insights into sub-arc magma generation processes. An anomalously high density of mafic lava shields exists within the central Cascade arc due to the influence of two separate extensional systems. First, development of a mafic platform, the base of the modern High Cascades, is associated with intra-arc rifting continued from the late Miocene (~8 to 10 Ma) to the early Pliocene (~5Ma) when a graben developed along the eastern side of the western Cascades. Second, mafic volcanism is associated with the Brother's Fault zone, a diffuse set of NW-trending faults marking the intersection of the extensional Basin and Range Province with the Cascade arc in central Oregon. Snow Peak is a mafic shield volcano that sits within the modern Cascade forearc, ~50 km west of the current High Cascades, and represents the westernmost forearc mafic volcanism of the early High Cascades. Snow Peak lavas consist of low-to-medium-K, high-alumina olivine tholeiite (HAOT) basalt (SiO2: 49.2- 52.1 wt.%, MgO: 5.5-8.8 wt.%, K2O: 0.12-0.66 wt.%, Al2O3: 16.4-18.7 wt.%) that ranges from primitive to moderately evolved (Mg# 51 - 62). Mantle-normalized multi-element plots indicate Snow Peak lavas are generally HFSE depleted and LILE enriched (Ba/Nb: 36.3 - 53.5). 40Ar/39Ar dating constrains the volcanic activity at Snow Peak between 6.05±0.12 Ma and 5.31±0.07 Ma. This age range indicates that Snow Peak was active during the late Miocene to early Pliocene transition from the western Cascades to the early High Cascades, and is contemporaneous to the basaltic volcanism of the Deschutes Formation east of the current arc. Two tectonic models for HAOT-type magmatism within the forearc of the Cascadia subduction zone are considered. The first model presumes that the Snow Peak lavas derived from the lateral transport of magma associated with the initial phases of intra-arc rifting during late Miocene to early Pliocene time. The second model relates the magmatism at Snow Peak to the impingement of the Basin and Range extensional province onto the arc.
V31B-2131
Oligocene and Miocene arc volcanism in northeastern California
The Warner Range in northeastern California exposes over 3 km of Tertiary rocks and offers a unique opportunity to study the long-term history of Cascade arc volcanism in an area typically covered by younger lavas. The oldest locally-sourced volcanic rocks in the Warner Range are Oligocene (26-30 Ma) and include a sequence of basalt and basaltic andesite lava flows overlain by hornblende and pyroxene andesite pyroclastic flows and minor lava flows. Both sequences vary in thickness (0-1 km) along strike and are inferred to be the erosional remnants of one or more large, partly overlapping composite volcanoes, possibly derived from the same magma chamber. A largely amagmatic period lasted from about 25-16 Ma, although distally-derived silicic tuffs were deposited in the region during this time. Arc volcanism resumed at about 16 Ma with basalt and basaltic andesite lavas sourced from local eruptive centers 5-10 km south of the relict Oligocene centers. Post-16 Ma volcanism continued into the late Miocene and Pliocene, forming numerous eroded but well-preserved shield volcanoes to the west and south of the Warner Range. Geochemically, Oligocene through late Miocene volcanic rocks in and around the Warner Range are calc-alkaline basalts to andesites (48-61% SiO2) that display negative Ti, Nb, and Ta anomalies consistent with an arc setting. Middle Miocene lavas in the Warner Range are distinctly different from the Steens Basalt (with which they were previously correlated), which has higher Ti, lower Al, and flatter REE patterns. Middle to late Miocene shield volcanoes west of the Warner Range consist of homogenous basaltic andesites (53-57% SiO2) that are geochemically similar to older rocks in the Warner Range. They are distinctly different from younger (late Miocene to Pliocene) high-Al, low-K olivine tholeiites, which are more mafic (46-49% SiO2), did not build large edifices, and are thought to be related to back-arc extension. The Warner Range is about 100 km east of the axis of the modern arc in northeastern California, suggesting that the arc was either more broad in the middle Miocene, or migrated west during the late Miocene due to a change in dip of the subducting slab.
V31B-2132
Two Contrasting Volatile Element Compositions in Primary Melt Inclusions From Mount Shasta
In order to get the pre-eruptive volatile contents of Mount Shasta lavas, we selected primary melt inclusions from samples 85-38 (high alumina olivine tholeite, HAOT), 85-47, 85-1a and 95-15 (basaltic andesites, BA; Baker et al., 1994, CMP; Grove et al., 2002, CMP). We analyzed the H2O, CO2, F, Cl and S contents from olivine-hosted (Fo86-91) primitive melt inclusions (SiO2 <50 wt% and MgO >8 wt%) using the CAMECA 1280 ion probe of WHOI (MA, USA) and a broad range of standard compositions. As the melt inclusions were partially crystallized, they were experimentally heated in order to melt the daughter minerals. A preliminary study show that our heating procedure did not create any significant loss in the CO2, F, Cl, and S content of the melt inclusions and may not even result in a partial loss of the H2O content of the melt inclusions. The major element compositions of the Shasta melt inclusions fall on a continuous extension of the trend defined by the whole rocks and represent a more primitive endmember. Regarding the volatile contents, they can be divided into two groups: HAOT melt inclusions have low and clustered volatiles compositions (H2O: ~0.04 wt%; F: ~120 ppm; Cl: ~30 ppm; CO2: 250-450 ppm and S: 800- 1100 ppm), similar to primitive MORB compositions, whereas BA melt inclusions show higher and more variable volatile compositions (H2O: 0.1-2.6 wt%; CO2: 70-840, F: 120-1180 ppm; Cl: 480-1230 ppm and S: 480-4600 ppm). They are enriched in volatile elements compared to the primitive MORB compositions. Moreover, the BA melt inclusions are enriched in mobile elements such as K, Ba and B. The results suggest that HAOT melt inclusions and more enriched BA melt inclusions represent two endmembers of a mixing process between a depleted mantle melt and a slab-derived fluid component. The variations in the BA volatiles compositions can indicate an addition of various amounts of aqueous fluid into the mantle source, at different stage of the lavas formation. The division of the melt inclusions into two groups, a dry one (HAOT melt inclusions) and a wet one (BA melt inclusions), is in good agreement with the previous results on the formation of the Mount Shasta lavas.
V31B-2133
The Galapagos-OIB signature of the central Costa Rican volcanic front: arc-hotspot interaction
Although most Central American magmas have a typical arc geochemical signature, magmas in southern Central America have isotopic and trace element compositions with an OIB affinity, similar to the Galapagos hotspot lavas. Our new data for Costa Rica suggest that this signature, unusual for a convergent margin, has a relatively recent origin (Late Miocene-Pliocene ca. 6 Ma). We also show that there was a transition from typical arc magmas (analogous to the modern Nicaraguan volcanic front) to OIB-like magmas. The geographic distribution of the Galapagos signature in recent lavas from southern Central America is present landward from the subduction of the Galapagos hotspot tracks (the Seamount Province and the Cocos/Coiba Ridges) at the Middle American Trench. The higher Pb isotopic ratios, relatively low Nd isotopic ratios and enriched geochemical signature of central Costa Rican magmas can be explained by arc-hotspot interaction. The isotopic ratios of central Costa Rican lavas require the subducting Seamount Province (Northern Galapagos Domain) component, whereas the isotopic ratios of the adakites and alkaline basalts from southern Costa Rica and Panama are in the geochemical range of the subducting Cocos/Coiba Ridges (Central Galapagos Domain). Geological, geochemical, and isotopic evidence collectively indicate that the relatively recent Galapagos-OIB signature in southern Central America represents a geochemical signal from subducting Galapagos hotspot tracks, which started to collide with the margin ~8 Ma ago. The Galapagos hotspot contribution decreases systematically along the volcanic front from central Costa Rica to NW Nicaragua.
V31B-2134
Baseline for estimates of mantle serpentinization in the subducting Cocos plate offshore Costa Rica and Nicaragua from OBS wide-angle seismic data
We present new OBS seismic refraction data gathered on a transect on the Cocos plate, parallel to and seaward of the Middle America Trench offshore Nicaragua and Costa Rica. The data were shot in March and April 2008 with the new seismic vessel R/V Marcus Langseth. The purpose of this wide-angle seismic data set is to quantify the amount of lower plate serpentinization along the strike of the subduction zone. Geochemical studies of volatile output of the Central American Volcanic arc indicate efficient recycling of water and sediments that are subducted with the Cocos plate. The slab signature is stronger in Nicaragua than in Costa Rica, which suggests that more water is subducted to the north. Indeed, the subducting lithosphere appears more faulted offshore Nicaragua, which allows more water to penetrate the upper mantle, where it may react with peridotites to form serpentinite. These serpentinites are expected to break down under high pressures and temperatures experienced beneath the volcanic arc. Recent marine seismic refraction studies (Grevemeyer et al., 2007; Ivandic et al., 2008) offshore Nicaragua have shown that the mantle of the downgoing oceanic plate has mantle seismic velocities 5 to 7 percent lower than in normal oceanic lithosphere, which can be explained by 15 percent serpentinization in the uppermost mantle. If this low-velocity anomaly is smaller offshore Costa Rica, we can confirm that the degree of basement faulting in the downgoing lithosphere correlates with the amount of mantle serpentinization. This finding would be important for our understanding of water recycling in subduction zones.
V31B-2135
Reconstruction of the geologic history of the north side of Poás Volcano
The Costa Rica Institute of Electricity (ICE), has drilled about 155 cores on the NE and NW and NE flanks of Poás volcano as part of the site characterization study for the construction of several hydroelectric projects: Hule, Toro 1, 2, 3 and Cariblanco. The cores, up to 275 m in depth, sample the last 400 ka of eruptive and erosional history of Poás. In addition to the cores, there is also a 6 km horizontal tunnel from the Cariblanco project that has been logged and sampled in detail. A detailed stratigraphic section of the NE flank of Poás volcano has been constructed using petrographic descriptions from the cores and tunnel logs. The stratigraphic section is composed of four geological units, which, according to preliminary dating from the oldest to the youngest, are: Río Sarapiqui (>400 ka) composed principally of breccias and tuffs with subordinate lavas and an ignimbrite deposit; Paleo-Poás (400 ka aprox), consisting primarily of lavas, breccias, and tuffs. One of the cores through Paleo-Poas samples seven different lava flows separated by paleosoils, which together account for 260 m of total thickness. The von Fantzius unit (40-10ka) with a maximum thickness of 70 m is made of lava flows and some pyroclastic deposits in the upper part of the unit. The Congo unit is about 6 ka and is made of pyroclastic and explosive deposits. Petrologically, the lavas of these units range from basalts to andesites. The initial correlations and profiles based of these units will be tested and refined using geochemical data, 40Ar/39Ar, 14C dating and paleomagnetic inclination measurements of the lava flows contained within these four units. We have collected more than 350 oriented samples for paleomagnetic studies and an additional 150 samples for geochemistry and radiometric dating. The samples are from fifteen cores from Cariblanco project, five cores from Hule project and eight cores from the Toro-3 project. In addition, the walls of the Toro-3 horizontal tunnel have been sampled. With the integration of these data it should be possible to constrain the eruption rates, volume fluxes, and erosion events (indicated by soil horizons) and produce a corrected extrusive volcanic flux-rate model of the Poás volcano. The flux-rate model of the Poás volcano will further constrain the element fluxes from the volcanic front of Nicaragua and Costa Rica.
V31B-2136
Mantle-Type Granites From Jilotlan, Jalisco, Mexico: Geochemical and Isotopic Evidence
The Jilotlan Intrusive, located in the State of Jalisco, SW Mexico, almost 100 km off the Pacific coast and the actual Middle American trench, occupies a semicircular area of nearly 6,000 km2. The composite plutonic body consists of several phases: from gabbros and diorites to tonalites, granodiorites and granites. Locally, the granitic member intrudes the gabbro. The area is extensively intruded by several dike generations of mafic composition, although minor acidic dikes are observed as well. The Jilotlan pluton intrudes mid- Cretaceous volcanosedimentaries and carbonates. Representative units of the gabbroic members were sampled for petrographic studies, major, trace geochemical data and Rb-Sr and Sm-Nd isotopic investigations. Textural variations of gabbroic rocks in the field are ubiquitous. Main mineral phases consist of plagioclase with amphiboles and clinopyroxene. Chemically, products are quite heterogeneous too, displayed by a light spread of SiO2, MgO, Na2O and K2O (46.6-52.4, 4.0-9.9, 1.8-4.1, and 0.3-2.7 wt.%., respectively) at a magmatically primitive level. Total alkalis vary from 2.5 to 5.2 characterizing the rocks as belonging to the calc-alkaline suite. Multielemental diagrams show similar patterns with Cs, Rb and Ba enrichment and positive anomalies of Pb and Sr. REE diagrams exhibit a relatively flat pattern together with a light LREE enrichment. Initial isotopic signatures (at 70 Ma, Rb-Sr, WR) of gabbros are very low, with 87Sr/86Sr between 0.70328 and 0.70369 and similar epsilon-Nd values from +4.7 to +6.8. Tonalites and granodiorites have slightly higher, less dispersed 87Sr/86Sr ratios and epsilon-Nd values (0.70369-0.70379, and +2.6 to +3.7 respectively). Isotopic signatures of gabbros are similar to rocks from the close by Manzanillo batholith. Both gabbros from Jilotlan and Manzanillo form one of the largest gabbro provinces known in Mexico, with Sr-Nd isotopic ratios even more primitive than most of Trans Mexican Volcanic Belt volcanics. These results indicate a mantle origin of these continental arc intrusives, associated to Cordilleran subduction processes, which constitutes a rare scenario in magmatic arcs. There is no evidence for accretion or for significant crustal assimilation.
V31B-2137
Evolution of the North Central Peruvian Subduction Zone: Evidence From Late-Stage Adakite-Like Ignimbrites
The Miocene Yungay and Fortaleza Ignimbrites of the Peruvian Cordillera Blanca are characterised by restricted geochemical variations, with SiO2 >70 wt %, Na2O >4 wt %, Mg numbers from 4 to 30, and strong depletions in Y, Yb and Lu. La/LuN and Sr/Y ratios are high, at >20 and >65 respectively. These geochemical indicators are typical of Phanerozoic adakite-like rocks and Archaean Trondhjemote- Tonalite-Granodiortite (TTG) suites which are often interpreted as partial melts derived from subducted oceanic slab. 40Ar/39Ar dating on Yungay feldspars yields an age range of 4.1 to 7.5 Ma, while K- Ar dating on Fortaleza biotites places them at 4.9 to 5.84 Ma. Both Ignimbrites were erupted during major tectonic modifications to the Peruvian subduction margin, and it is against this complex tectonic backdrop, effectively marking the end of the current 200 Ma Andean Cycle, that the adakite-like Yungay and Fortaleza magmas were generated in rapid succession. There are only two possible source regions for the Peruvian Ignimbrite melts: thick underplated lower crust or the downgoing oceanic slab. Resolving between these allows a full understanding of the magmatic evolution of the Peruvian margin, and carries implications for formation of TTG melts. Radiogenic isotope compositions of the Yungay and Fortaleza Ignimbrites point to a juvenile crustal source (87Sr/86Sr = 0.70547 to 0.70631, 143Nd/144Nd = 0.51245 to 0.51257, ε Hf = -3.2 to 0.88, 206Pb/204Pb = 18.767 to 18.783). Direct observation of source mineralogy and melting conditions, through experimental petrology, indicates that pressures of >2.2 GPa and temperatures in excess of 1025° C are needed to generate the chemical compositions of the Ignimbrite melts. Initial melts at source were in equilibrium with a fluid-absent, hydrous source mineralogy consisting of quartz, garnet and clinopyroxene (and possibly zoisite). Quartz-hosted melt inclusions show the Ignimbrite magmas were volatile-rich, and suggest the magmas experienced a short period of crystallisation near 40 km depth after initial generation. The Yungay and Fortaleza Ignimbrites represent the final melts of a short-lived, deep lower crust at ~80 km depth, created by progressive shallowing of the Nazca Plate after break-up of the Farallon Plate at 27 Ma. Adakite-like melts may simply reflect a natural progression in arc geochemistry as a mature subduction system evolves and crust is thickened via tectonic and magmatic events during the final stages of a magmatic cycle.
V31B-2138
Devonian-Ordovician Magmatism in Chiapas Massif, Southern Maya Block, Mexico
The Chiapas Massif (CM) is located in SE Mexico and extends over an area of more than 20,000 Km2
parallel to the Pacific coast between the Isthmus of Tehuantepec and the Guatemalan border. It constitutes
the largest batholitic complex in Mexico and belongs to the Maya Block. The CM is predominantly formed by
igneous, metaigneous and metasedimentary rocks. In the central CM several magmatic and metamorphic
events have been identified in igneous rocks between Late Permian and Triassic times (220-240 Ma),
together with a Jurassic tectonothermal event. Recent geochronological studies have been focused on
granitic rocks from the central-southeastern CM and from this area, single grain zircon ages were also
obtained from metasedimentary rocks. The results suggest that the basement rocks of the eastern CM and
the Maya Block underwent several tectonothermal events since Ordovician and Devonian times. In this work,
we present new data of magmatic and metamorphic rocks from the easternmost part of CM which confirms
this hypothesis. Additionally, we present a semi-detailed geological map of this area showing some field
relationship between the different units. We identified a new basal sequence significantly older than the
Santa Rosa Formation, which has been considered up to now as forming the major underlying sequence of
the CM, as well as the occurrence of several magmatic events in the Maya Block and in rocks from
surrounding areas. The petrological, geochemical and geochronological features of these rocks show
continuous crust recycling, the occurrence of within-plate magmatism in some parts with inherited
Greenvillian and Archaean zircon grains as well as the occurrence of arc magmatism previous to accretion,
deformation and terrane separation of the crustal blocks. Our new geochronological results obtained from
the south easternmost part of the CM include a Rb-Sr biotite-muscovite age of 392+/-9 Ma and an Ar-Ar
muscovite age of 406+/- 4 Ma from a tectonized granitic body together with a Sm-Nd garnet age of 474+/-24
Ma and an U-Pb single grain zircon age of 482+/-3 Ma from a S type granite body. From a close by
amphibolite unit we obtained an age of 456+/-14 Ma (U-Pb, single grain zircons). These ages represent a key
tool for understanding the old assemblage between the crustal blocks at the Maya-Chortís boundary
during Paleozoic times focusing the tectonic scenario, temporal rock relationships, metamorphic conditions
and magma emplacement mechanisms.
http://www.geologia.unam.mx
V31B-2139
Stable isotope constraints on fluid-pressure buildup cycles and fluid–rock interaction at the frontal part of the early Eocene-middle Miocene convergent system of the Northern Apennines (Italy)
Understanding the mechanical behavior of subduction megathrusts is intimately related to the characterization of their fluid regime. Stable isotopes help in characterizing the fluid regime that existed during megathrust development. Here we analyze the late Eocene-mid Miocene, erosive boundary between the European and Adriatic plates, now exhumed in the N. Apennines. 13C and 18O composition of 125 calcite veins and 83 host rocks have been analyzed from 4 transects of the 500 m-thick fault zone. Stable isotopes have been coupled with structural analysis. Field and microstructural observations suggest that the evolution of structures in the deeper portion of the outcropping megathust, correspondent to 5 km depth, was genetically linked to episodic high Pf, resulting in cyclic fluid-induced brittle deformation interspersed with aseismic ductile creep. δ18O and δ13C values vary from samples of host rock and veins. For the hangingwall rocks the δ18O values show a great variability for each transect if they relate to the previous accretionary prism, from -8.23 to -2.16‰ PDB. If they relate to the slope sediments the isotopic values within each transect are more uniform as well as shown for the footwall samples. The accretionary prism-related rocks are also characterized by a high variability of δ13C, from -1.28 to 2.29‰ PDB, as well as the slope sediment related rocks, from -1.82 to 0.92‰ PDB. Systematic variations of isotopic values are instead observed in the vein samples. In general they show a depletion in δ18O and a shift from mostly positive to negative δ13C values with respect to both the hangingwall and footwall. For the shear zone veins O- and C-isotopic composition tend to be more uniform irrespective of the host rock values. Different localities, though, have different linear trends of δ13C and δ18O values. In general the isotopic data are consistent with a fault architecture where the deformation results concentrated on an array of discrete faults.
V31B-2140
Geochemistry of the Chagai Arc, Pakistan: Complex arc Dynamics Spanning the Cretaceous to the Quaternary.
The Chagai arc is located in western Pakistan and extends into Iran and Afghanistan. The arc forms a elongate body trending EW and is roughly 500km long by 150km wide. Activity along the arc began in the Late Cretaceous and continued through into the Quaternary. The oldest volcanic rocks in the arc belong to the Sinjrani and Kuchakki Formations. These rocks are primarily basalts and basaltic-andesites which are form both pillow sequences and massive flows. Geochemically these units are very similar. They are tholeiitic lavas with typical island arc characteristics and an N-MORB source. For example when normalized to N-MORB they are LILE enriched, HFSE depleted, and have negative Nb and Ti anomalies. Intruded into the oldest units are younger diorite and granodiorite intrusions which are most likely Oligocene to Miocene. These rocks are calc-alkaline and but again have island arc affinities with a more depleted N- MORB source than the older lavas. These rocks are more evolved and show a negative europium anomaly on a chondrite normalized spider diagram. The youngest units in this study belong to the Quaternary Koh-e-Sultan and the Miocene Koh-e-Dalil These units have a calc-alkaline fractionation trend and contain more silicic lavas, including dacites, than the older lavas. Chemically these units are very similar; they both contain continental arc signatures and were generated from an E-MORB source. Current theories to explain the multiple phases of volcanism in the Chagai arc propose that these lavas are the result of intra-oceanic convergence in the Neo-Tethys. During the Latest Paleocene the arc collided with the Afghan microplate. Our preliminary data supports this model in that the initial phases of volcanism are entirely oceanic arc, followed by much younger continental arc volcanism.
V31B-2141
Mantle Wedge Serpentinization : New Constraints on Fluid Mobile Elements (As, Sb, B, Li) Behaviour During Subduction Processes
Serpentinization of dry mantle wedge by fluid released from slabs is an important process in the element transfer during subduction processes. The deep dehydration of mantle serpentinites is partly responsible for the trace-element and isotopic signature of arc magmas (Hattori and Guillot, 2003). In situ analysis (Laser HR ICP-MS) of exhumed mantle wedge serpentines (Himalaya, Ultra-high pressure Tso Morari unit) shows a significant enrichment in fluid mobile elements (FME; e.g. Sb, As, B, U, Pb), but this enrichment was not observed in the associated iron oxides. Rare earth elements (REE) and some compatible elements such as Sc, Co or Ti remain immobile during hydration processes, allowing the identification of the primary minerals (olivine, clinopyroxene or orthopyroxene). Some serpentines (associated with metamorphic olivine) display a strong fractionation of the Nb/Ta ratio (30-1000 with Nb ranging from 0.18 to 0.38 ppm) that could reflect the presence of secondary minerals such as clinohumite (Garrido et al., 2005). Depending on the nature of the primary minerals, FME enrichments differ. Serpentinized olivines display strong enrichment in Sb, As and B (10 to up to 1000 x PM), whereas serpentinized orthopyroxenes are enriched in Pb, Cs and Li (2 to up to 10 x PM). Our observations are consistent with the experiments performed by Allen and Seyfried (2003). Below 400°C, at shallow depths, olivines are preferentially serpentinized and incorporate elements that are fluid soluble at low temperature, such as Sb, As and B. Above 400°C, pyroxenes are destabilized and incorporate Pb, Cs, Li and possibly Ba. This contrasted enrichment in natural serpentines reflects the temperature increase occurring during the downward movement of the serpentinite layer along the subduction plane. The serpentine acts as a sink for water, but also for fluid mobile elements and transports them to deeper and hotter levels in the mantle, down to the isotherm 600-650°C where dehydration occurs. Allen and Seyfried, 2003, GCA, 67, 1531-1542. Garrido et al., 2005, G3, 6, doi: 10.1029/2004GC000791. Hattori and Guillot, 2003, Geology, 31, 525-528.
V31B-2142
Mercury and Iodine systematics of volcanic arc fluids
The mantle has low Mercury and Iodine concentrations, but these elements occur in volcanic gases and hydrothermal fluids at ppb (Hg) and ppm (Iodine) levels. Possibly, the Hg and Iodine concentrations in volcanic fluids reflect subducted sediment sources in arc magmas. Iodine is a biophilic element, and I129/I values indicate that subducted sediment (especially organic matter) is an important Iodine source for arc magmas. It is uncertain if this is true for Hg as well, although in the surface environment Hg is commonly associated with organic matter. We present 60 new analyses of Hg and I in fluids from volcanoes in Central America, New Zealand, Japan, and the Cascades. A first assessment suggests that Iodine is released to some degree in the early stage of subduction in the forearc, whereas Hg may be released largely below the main volcanic arc. Isotope and trace element signatures of volcanic rocks of the investigated volcanoes show no simple correlation with Hg or Iodine abundances. The acid hot spring fluids of Copahue volcano (Argentina) carried ~ 200 ppt Hg in January 1999, ~80 ppt Hg in March 2008, and 90 ppt Hg in the crater lake in March 1997. The dissolved Hg fluxes from the Copahue hydrothermal system are ~300 gr Hg/year in 1999 and ~130 gr Hg/year in 2008. The bulk hydrothermal Hg flux (particle bound+dissolved) in 2008 was ~ 350 gr Hg/year. The potential Mercury evasion from these hydrothermal spring fluids into the air has not yet been incorporated in these estimates.
V31B-2143
Experimental demonstration of lithium-boron depletion during magma degassing
It has been speculated that some volatile elements partition into a gas phase and leave the host magma during a degassing event. We have analyzed two experimentally degassed natural rhyolite sample (Güney Dagi) and they show that significant lithium and boron depletion in the host lava can take place within 12 min of isothermal decompression at 800 °C. The starting glass material was presaturated in H2O at 200 MPa, 800 °C, then it was isothermally decompressed at the rate of 1000 and 27.8 kPa/s. Initial water content at the starting condition is approximately 6 wt%. The duration of bubble growth was 13 second and 12 min depending on the decompression rate and the final pressure before quench is approximately 50 MPa. The fast and slow decompression resulted in heterogeneous lithium and boron abundance in the charge. The areas further from bubbles have lithium and boron abundance indistinguishable from the starting glass, and the areas close to bubbles show 43 and 18 % depletion of lithium and boron, respectively, while H2O depletion is 50 %. Our result suggest that significant quantity of lithium and boron can leave a natural magma system during water degassing event. The lithium depletion of the host lava was close to that of water, and more extensive than that of boron. We interpret that the efficient depletion of lithium compared to boron was due to its fast diffusivity. This also signals that the interpretation of lithium and boron geochemical observations in subaerial lavas, which are common in subduction zones, must consider degassing events.
V31B-2144
Tracking Trace Elements Exchanges Between Fluids and Minerals in Situ at High Pressure and High Temperature
In subduction zones, deep aqueous fluids play an important role in chemical partitioning and mass transfer. In order to constrain the chemical budget in subduction zones, the properties and the composition of the fluids, as well as their evolution with depth are required. Among experimental set-up, the Diamond Anvil Cell (DAC) is very powerful for in situ investigation of materials at High Pressure and High Temperature (HP-HT). While physical properties of condensed matter have been widely documented using X-Ray Diffraction (XRD), chemical analyses at HP-HT in situ in DACs are still scarce, since there are still demanding. Only a few attempts have been reported, using hard X-Ray fluorescence spectroscopy (XRF) to characterize mineral dissolution, fluid-melt interactions using either membrane DAC [1] (up to 5GPa, 400°C) or hydrothermal DAC [2](up to 1.5 GPa, 800°C). Most of these experiments report data for concentrated solutions (a few 100 ppm) and focused only on one element at a time. Detecting several trace elements in the ppm range in the same time could offer a better characterization of fluid-mineral interactions. For this purpose, we carried out several technical developments at ID22 beamline (ESRF, France). We have built a HP-HT set-up adapted for multi-technique analyses such as XRF, XRD, X-Ray Absorption Spectroscopy (XAS). We designed: i) a special membrane DAC for High Temperature XRF measurements at 90° from the incoming beam, equipped with ii) a new heating system based on tungsten wires [3] in iii) a vacuum chamber. We recently performed simultaneous and in situ XRD and XRF measurements up to ca. 3 GPa, 1000°C, using a two hole gasket. In the first hole, we measured XRF from a 28 element standard aqueous solution, while XRD pattern from gold powder used as P gauge was recorded in the second hole dedicated to calibrants. Concentrations of 30ppm and a minimum detection limit better than 1 ppm were estimated with XRF. The solution composition and concentrations were a posteriori calibrated and crosschecked by ICPMS. The proposed HP-HT experimental set-up offers unique capabilities: P and T can be tuned and measured independently, to ca. 1000°C, 3 GPa. While 1000°C is probably the maximum temperature for such a cell, there is in principle no pressure limitation. Special care was given to DAC material and geometry to limit both the background originated from the DAC, and the Compton scattering from the diamonds. Up to 10 elements can be simultaneously measured at the ppm level. This new apparatus offers new opportunities for in situ geochemistry at HP-HT. [1] Sanchez-Valle et al, American Mineralogist, 88, 978-985, 2003 [2] Schmidt et al, American Mineralogist, 88, 288-292, 2003 [3] Pasternak et al, Review of Scientific Instruments, 79, 2008
V31B-2145
Investigation of F, S, and Cl Standards by ion Probe and Electron Microprobe
Microanalytical measurements of volatile elements, such as H, C, F, S, Cl, in melt inclusions provide critical information about their pre-eruptive abundances. Such information can shed a new light on processes of arc magma genesis. In recent years, there have been significant advances in analyzing tens of ppm-level abundance H and C by ion probe, while high quality FTIR calibrations are also available for these elements. On the contrary, F, S, and Cl are commonly analyzed by electron microprobe for samples above hundreds of ppm-level abundances. We have examined a series of silicate glasses with volatile values published by multiple laboratories, using the high mass resolution Cameca IMS 1280 ion probe at WHOI (USA) for cross calibration for F, S, Cl. Our preliminary results show linear correlations, spanning 2 to 4 orders of magnitude, between Si-normalized volatile ion probe intensities and the published concentrations. This exercise reveals, however, significant electron microprobe – ion probe disparities for some glasses with F and Cl abundances below 100 ppm. The ion probe measurements are systematically lower than those of electron probe. The same tendency can be found among the data available for MPI-DING glasses. We also note that the StHs6/80-G glass we have analyzed resulted in significantly lower F, S, and Cl abundance that the published values. While it should be stressed that F, S, and Cl abundances of the MPI-DING series are only consultative values, our observation suggests that cross calibration studies are necessary for all silicate glass "standards".
V31B-2146
Constraining melt generation and migration at convergent plate boundaries with trace element transfer models
We investigate the transfer of stable trace elements (Ba, La, Ce, Nd, Sm and Yb) and radioactive 238U and 230Th from the solid to the melt both by melting and by diffusive exchanges at convergent margins. The aim of this study is not to exactly reproduce trace element concentration and U-series variations for specific subduction zones, but rather to relate trace element and U isotope variations produced by simplified trace element transfer models to melt transport mechanisms in the wedge. Wet melting at convergent margins differs significantly from the dry decompression melting occurring at mid-ocean ridges: the majority of melt production in the wedge is confined to a sharp front located where hydrous fluids rising from the subducting plate first encounter mantle at high enough temperatures to melt, whereas decompression melting at mid-ocean ridges occurs over a larger range of depths. At convergent margins, solid flow determines the rate at which fertile material is brought to the melting front, thereby controlling melt production rates. Variations in melt production rates, pressure and temperature within the melting region significantly influence the mode of trace element transfer. Melts from different parts of the melting region consequently have highly heterogeneous trace element concentrations. These melts may migrate toward the surface via buoyant porous flow, channelized flow or diapiric ascent. Comparing trace element and U-series distributions from the models with distributions observed in arc lavas suggests several possibilities for the origin of arc lavas within the melting region: 1) arc lavas are derived from melts produced in the region of inflow of fertile material near the leading edge of the melting region, or 2) slab-derived fluids add a significant amount of highly to moderately incompatible elements into the melting region. Furthermore, U-series results suggest that melts need to migrate toward the surface with limited chemical exchanges with surrounding mantle material. The quantitative constraints of these possibilities are evaluated in the present study.
V31B-2147
Deformation Experiment of Antigorite Serpentinite with a Solid Medium Deformation Apparatus
Serpentinite is believed to be widespread in and along oceanic plates undergoing subduction. Deformation of serpentinite probably plays an important role in subduction and exhumation process of slab. However, rheological behavior of serpentinite is poorly understood, especially at more than 1GPa. We conducted constant strain rate experiments of antigorite serpentinite to understand the rheological behavior of it using modified Griggs type apparatuses. Experimental conditions were 1 GPa and 3GPa confining pressures, 300 C to 700 C, and 2.0 to 2.4E-5/s of strain rate. The serpentinite sample was collected from Nagasaki metamorphic rock in Japan. The sample was cored parallel to the foliation plane for the starting materials. The sizes of starting materials were ca. 7.0 mm diameter ~ ca. 7.0 mm length or ca. 3.0 mm diameter ~ ca. 8.0 mm length. The mechanical data indicate the following two characteristics: 1) the flow stress at same temperature increases with an increasing confining pressure, 2) temperature dependence flow stress also increases with an increasing confining pressure. The second character is probably attributed to change of deformation mechanism. We are now observing the microstructures of recovered samples with TEM to clarify the deformation mechanism. In the sample from 3GPa and 600 C conditions, many stacking faults developed compared with the starting sample, which suggests that formation process of stacking faults controls the plastic deformation of antigorite at high pressure.
V31B-2148
Talc and "talc"-bearing dehydrating serpentinite rheology within subduction zones
Fluids released by subducting slabs hydrate peridotites within the mantle wedge and produce weak phyllosilicates in significant quantities (1). Depending on the original chemistry of peridotites and on the silica content of the fluids, either talc or the high-pressure (HP) variety of serpentine antigorite, or both, can form down to 180 km depth. A talc-like phase is also produced transiently during serpentine dehydration (2). The extent to which such weak materials at the slab-mantle wedge interface can influence the dynamics of subduction zones from human (seismicity - post-seismic ground deformations) to geological (convection) timescales is still unknown. Promising deformation experiments on antigorite have shown that its low strength makes it a potential actor for governing silent earthquakes that release elastic energy in subduction zones and for localizing deformation at the slab interface, thereby modifying the mantle wedge convection, heat fluxes and seismic anisotropy (3). The strength of talc at low pressure is also exceptionally low compared to other silicates (4) and at sub-surface conditions the weakness of talc is thought to govern deformation in major faults such as San Andreas (5). At low pressures (< 500 MPa), talc behaves cataclastically and hardly achieves distributed deformation (4). However, higher pressures should promote crystal plasticity i.e. different deformation mechanisms in talc. Investigating the rheology of talc, and talc-bearing assemblage during serpentine dehydration, at P-T conditions corresponding to subducting slabs, is necessary in order to better understand localization of deformation in the slab-mantle wedge interface as well as the nature of this interface. We conducted deformation experiments using the D-DIA (6) apparatus at GSE-CARS (APS sector 13), with in situ strain and stress measurements using synchrotron X-ray imaging and diffraction, respectively. Stress-strain data were obtained on talc at P-T conditions between 2 and 5 GPa, and T between 350 and 700C, at strain rates between 10-4 and 10-5 s-1. Dehydration of serpentine was conducted around 4 GPa and 10-5 s-1. The results will be presented together with their implications, in particular the influence of the talc-like phase within a dehydrating serpentine, and more generally for subduction zones dynamics. 1 Ulmer and Trommsdorff GSA spec. publ. (1999). 2 Perrillat et al EPSL (2005) 3 Hilairet et al Science (2007) 4 Escartin et al EPSL (2008) 5 Moore and Rymer Nature (2007) 6 Wang et al RSI (2003)
V31B-2149
Plastic and dehydration instabilities of antigorite serpentinite
We conducted a constant displacement rate test of antigorite serpentinite by a triaxial solid medium deformation apparatus installed at Hiroshima University. Experimental conditions were P = ca. 1.0 GPa, T = 450 C to 800 C, which cover from stability to dehydration conditions of antigorite, and strain rate of digit of 10-5 /sec. Samples were cylindrical shape cored from serpentinite and their sizes were 7 mm x 7 mm and 5 mm x 5 mm in diameter by length. Faults were observed in recovered samples from all temperature conditions. Mechanical data of faulted samples showed stress drop of several 10 MPa during experiments. The followings summarize microstructural observations of the recovered samples, and propose generation processes of fault at stability and dehydration conditions. 1) Stability field (plastic instability): Antigorite grains develop a lattice preferred orientation (LPO) along a fault, which characterized by (001) cleavage face parallel to fault plane. This fact suggests the fault is generated by the following process. i) Antigorite grains on the plane applied by maximum shear stress are preferentially deformed by plastic manner, and then develop the LPO. ii) The arrangement of cleavage face gradually reduces the strength of this plane. iii) Eventually, embrittlement occurs at the critical point when the rock strength along this weak plane becomes smaller than shear stress. 2) Dehydration condition (dehydration instability): Fine grained dehydration phases less than 1 micrometer in size such as olivine and talc are detected as a thin vein along the fault. Moreover, talc and olivine, or antigorite ca. 1 – 5 micrometers in size with angular shape are observed within the fault as a fault gauge. These facts suggest the following generation process of fault. i) Dehydration reaction of antigorite begins on the plane applied by maximum shear stress and expands along this plane. ii) The dehydration gradually reduces the strength of this plane. iii) Eventually, embrittlement occurs at the critical point when the rock strength along this weak plane becomes smaller than shear stress.
V31B-2150
Elasticity of Glaucophane and Applications to Seismic Properties of High-pressure Low- temperature Oceanic Rocks in Subduction Zones
Blueschists and eclogites, high-pressure and low-temperature metamorphosed rocks are often associated with serpentinites in subduction zones. Because their seismic velocities are lower than in the surrounding mantle, these rocks are detected through tomography or as wave guides (possibly a mixture of metastable gabbro and eclogites)[1]. The average properties of rocks are known, but subduction rocks are subjected to deformations, which create anisotropy. To model anisotropy, the single-crystal elastic properties of each components of blueschists and eclogites are required. carefully selected from the Sesia-Lanzo Zone (Aosta Valley, Western Alps). We therefore conducted detailed Brillouin spectroscopic measurements to obtain the elastic properties of glaucophane. Over 100 velocities were obtained on three oriented faces close to [100],[010],[001]. Brillouin spectroscopy under ambient conditions gave a bulk modulus of K0=95.8(2.3)GPa and a shear modulus of G0=63.7(0.9)GPa. These results agree with those obtained from Equation of State on glaucophane Ks=96.0 GPa [2]. The elastic constants give C33=254.5(5.8)GPa > C11=122.3(1.9) GPa, as compared with C33=191.6 GPa > C11=116 GPa obtained by [3] on Hornblende. The anisotropy of Vp (6234-9230 m/s) and Vs (3939-5244 m/s) is significant and velocities are faster in the [001] direction. Electron BackScattering Diffraction and Careware software will be used to obtain pole figures and preferred orientations of natural samples and combined with our new set of elastic data for glaucophane. The influence of lattice preferred orientations [4,5] on seismic anisotropy in blueschists and eclogites will be discussed. [1] Abers et al., GRL, 23, 1171-1174, 1996. [2] Comodi et al., Eur. J. Mineral., 3, 485-499,1991. [3] Bass, AGU, 45-59, 1995. [4] Kern et al., JGR, 102, 3051-3065,1997. [5] Kitamura, Geophys. J. Int.,165,1058- 1065,2006.
V31B-2151
Phase relations in the peridotite-carbonate-chloride system at 7.0-16.5 GPa and the role of chlorides in mantle dynamics
Melting phase relations have been determined in a model chloride-carbonate-bearing peridotite (CaO-MgO- Al2O3-SiO2-Na2O-K2O-CO2-Cl) at 7.0-16.5 GPa and 1200-1800°C. The typical assemblage coexisting with partial melts is forsterite/wadsleyite - enstatite/clinoenstatite - garnet. In contrast to carbonated peridotite, clinopyroxene was absent in chloride-carbonate system. Also, carbonatite melt formed at low-degree of partial melting are enriched by CaO. We observed Ca- and Mg-bearing carbonatite melts in all experiments. The solidus temperature was adjusted by the stability of magnesite at 1300°C at 10.5 GPa and 1400°C at 16.5 GPa. Halides are stable 100-200°C above this 'apparent' solidus. This fact indicates that melt composition in chloride-carbonate peridotite can be effectively controlled by the presence of water. In the anhydrous environments a low-fraction of 'dry' Cl- and alkali-poor Ca-Mg-rich carbonatite melts can be formed and migrate from the source, whereas under the hydrous conditions, chloride can be dissolved into water to form alkali-rich chloride-carbonate brine-like melts. At higher temperatures (1400-1600°C) two immiscible carbonatite and chloride-carbonate liquids may coexist. The composition of carbonatite and chloride-carbonate melts is consistent with that of melt inclusions in fibrous/cloudy diamonds worldwide. In addition, the Na-rich composition of chloride-carbonate melts is in agreement with Na-enrichment in melt inclusions from peridotitic diamonds. Typical trends of melt evolution upon cooling and formation the melts included in diamonds may involve formation of immiscible Si-poor carbonatite and Si-poor chloride-carbonate melts from homogenous Cl-bearing kimberlite or carbonatite. Our results also applicable for carbonate and chloride stability and dynamics during deep subduction.
V31B-2152
Local Os isotope mantle heterogeneities preserved during melt transport processes in supra-subduction mantle at Voikar Ophiolite (Polar Urals)
We report a detail study of Highly Siderophile Elements (HSE) and Re-Os isotopes of lithologies from the Voykar Ophiolite (Polar Urals) which formed in a suprasubduction tectonic environment. Our data suggest that refractory harzburgite exhausted by melting more than 2 Ga ago was then reactivated by suprasubduction melts at about 0.6 Ga. These melts have formed dunite reaction channels and pyroxenite veins and locally redistributed the HSE in the surrounding harzburgite. In dunite channels and pyroxenite veins numerous sulfide globules transported by silicate melts have been observed. These sulfide globules make up the HSE budget of the pyroxenites and they are characterized by low concentration of Os, Ir and Ru compared to Pt, Pd and Re. High Re abundance in the pyroxenite (up to 2 ppb) suggest contribution from subduction slab. HSE composition of dunites results from reaction between percolating melt and host peridotite. Close to the contact, residual peridotites become enriched in Pt, Pd and Re. But primary HSE features of the harzburgite, such as Pd, Pt, Re depletion relative to Ir, Os and Ru appear to be preserved in a distance of about 50 cm away from the contact with dunite channel. The more important that even ancient Os isotope signatures (2 Ga model ages) are preserved in the harzburgite at 50 cm from the contact with dunite. Also relatively small blocks of harzburgite relicts within network of replacive dunites preserve similar ancient Os isotope signatures. These suggest that local heterogeneity in HSE concentrations and Os isotopes in the convecting mantle could survive melting and focused melt transport events.