Earth and Environmental Sciences

We adapt a 2D spectral domain finite difference waveform tomography algorithm previously used in active source seismological imaging to the case of a plane wave propagating
through a 2.5D viscoelastic medium in order to recover P and S wavespeed variations from body waves recorded at teleseismic distances. A transferable efficacy that permits recovery of arbitrarily heterogenous models on moderately sized computers provides the primary motivation for choosing this algorithm. Synthetic waveforms can be generated either by specifying an analytic solution for a background plane wave in a 1D model and
solving for the source distribution that would produce it, or by solving for a scattered field excited by a plane wave source and then adding the background wavefield to it. Because the former approach typically involves a concentration of sources at the free surface, the latter tends to be more stable numerically. We adapt a gradient approach to solve the inverse problem to maintain tractability; calculating the gradient does not require much more computational effort than does the forward problem. The waveform tomography algorithm can be applied in a straightforward way to perform receiver function migration
and travel time inversion.

Primitive chemical characteristics of high-Mg andesites (HMA) suggest equilibration with mantle wedge peridotite, and they may form through either shallow, wet partial melting of the mantle or re-equilibration of slab melts migrating through the wedge. We have re-examined a well-studied example of HMA from near Mt. Shasta, CA, because petrographic evidence for magma mixing has stimulated a recent debate over whether HMA magmas have a mantle origin. We examined naturally quenched, glassy, olivine-hosted (Fo87–94) melt inclusions from this locality and analyzed the samples by FTIR, LA-ICPMS, and electron probe. Compositions (uncorrected for post-entrapment modification) are highly variable and can be divided into high-CaO (>10 wt%) melts only found in Fo > 91 olivines and low-CaO (<10 wt%) melts in Fo 87–94 olivine hosts. There is evidence for extensive post-entrapment modification in many inclusions. High-CaO inclusions experienced 1.4–3.5 wt% FeOT loss through diffusive re-equilibration with the host olivine and 13–28 wt% post-entrapment olivine crystallization. Low-CaO inclusions experienced 1–16 wt% olivine crystallization with <2 wt% FeOT loss experienced by inclusions in Fo > 90 olivines. Restored low-CaO melt inclusions are HMAs (57–61 wt% SiO2; 4.9–10.9 wt% MgO), whereas high-CaO inclusions are primitive basaltic andesites (PBA) (51–56 wt% SiO2; 9.8–15.1 wt% MgO). HMA and PBA inclusions have distinct trace element characteristics. Importantly, both types of inclusions are volatile-rich, with maximum values in HMA and PBA melt inclusions of 3.5 and 5.6 wt% H2O, 830 and 2,900 ppm S, 1,590 and 2,580 ppm Cl, and 500 and 820 ppm CO2, respectively. PBA melts are comparable to experimental hydrous melts in equilibrium with harzburgite. Two-component mixing between PBA and dacitic magma (59:41) is able to produce a primitive HMA composition, but the predicted mixture shows some small but significant major and trace element discrepancies from published whole-rock analyses from the Shasta locality. An alternative model that involves incorporation of xenocrysts (high-Mg olivine from PBA and pyroxenes from dacite) into a primary (mantle-derived) HMA magma can explain the phenocryst and melt inclusion compositions but is difficult to evaluate quantitatively because of the complex crystal populations. Our results suggest that a spectrum of mantle-derived melts, including both PBA and HMA, may be produced beneath the Shasta region. Compositional similarities between Shasta parental melts and boninites imply similar magma generation processes related to the presence of refractory harzburgite in the shallow mantle.

The Cascades arc is a warm-slab subduction zone characterized by the slow (3.5 cm/a) subduction of young (< 10 Ma) oceanic crust. Elevated slab temperatures likely result in shallow dehydration beneath the forearc and reduced slab fluid addition to the mantle wedge beneath the volcanic front. Resulting calc-alkaline magmas should contain low volatile abundances relative to arcs with cooler subducting slabs. We test this hypothesis in Central Oregon, the region of the arc that has experienced the highest mafic output over the last 2 Ma. Naturally quenched basalt and basaltic andesite melt inclusions from tephra erupted from monogenetic vents record pre-eruptive magmatic volatile contents. After accounting for the effects of degassing, our data show that initial volatile contents from seven calc-alkaline cinder cones range from 1.7 to 3.6 wt.% H2O, 1200–2100 ppm S and 500–1200 ppm Cl. Ratios reflecting volatiles recycled from the subducted slab (H2O/Ce ~ 490–1700; S/Nb ~ 110–310; and Cl/Nb ~ 37–190) are elevated above MORB values but are low compared to those from the southern Cascades (Mount Shasta) and other arcs (Mexico, Central America, and Kamchatka). Positive correlations between H2O, Pb/Ce and Ba/Zr suggest that small additions of a slab-derived component to the mantle wedge are responsible for observed volatile contents.Melt inclusion compositions are evolved (Fo82–84 host olivines, 5.0–7.2 wt.% MgO) and fractionated from primitive parental magmas by ~ 16% crystallization of predominantly olivine before being trapped as inclusions. Calc-alkaline basaltic and Sr-rich basaltic melts can be generated by 9–11% partial melting of spinel lherzolite. H2O-rich basaltic andesite melts are highly depleted in incompatible elements and are probably generated from a more depleted mantle source by 12–16% melting. Subarc mantle temperatures inferred from H2O and trace elements are similar to or slightly warmer than temperatures in other arcs. Additionally, compositional proxies for slab surface temperatures in Central Oregon are at the high end of the global arc spectrum, corresponding to temperatures of 850–950 °C, and are consistent with a young, hot incoming plate.

1] We compiled a data set of 100 primitive arc magma compositions from melt inclusion and whole rock analyses to compare volatile contents, slab tracers, and calculated subduction component compositions between 18 subduction zone segments spanning the global range in slab thermal structure. The average primitive magma H 2 O content in our data set is 3.3 AE 1.2 wt.% (1 s.d.) for melts erupted within 50 km of the volcanic front. While there is a wide range of volatile contents in magmas within individual arcs, the highest values occur in magmas erupted from vents along the volcanic front, where the subducting slab is located 104 AE 29 km (avg AE 1 s.d.) beneath the surface. This observation, coupled with positive correlations between H 2 O, Cl, S, and B contents and predictions from geodynamic models, provides strong evidence for the active supply of volatile-rich slab-derived components from the subducting oceanic plate beneath volcanic arcs. We also show that temperature-sensitive ratios (e.g., H 2 O/Ce) for both primitive arc magmas and calculated subduction components are similar and display monotonic behavior with slab thermal parameter. Furthermore, calculated subduction component compositions have higher trace element to H 2 O ratios in arcs with hotter slabs (lower thermal parameter), suggesting that hydrous melts of differing compositions are added beneath different arcs. Finally, we present new volatile outflux estimates for Central Cascades magmatism and then compare these to estimates for the Central American and Kamchatka-Kurile arcs to create a combined data set spanning a large range in slab thermal parameter.

1] We have calculated slab fluid temperatures for 51 volcanoes in 10 subduction zones using the newly developed H 2 O/Ce thermometer. The slab fluid compositions were calculated from arc eruptives, using melt inclusion-based H 2 O contents, and were corrected for background mantle contributions. The temperatures, adjusted to h, the vertical depth to the slab beneath the volcanic arc, range from $730 to 900 C and agree well (within 30 C on average for each arc) with sub-arc slab surface temperatures predicted by recent thermal models. The coherence between slab model and surface observation implies predominantly vertical transport of fluids within the mantle wedge. Slab surface temperatures are well reconciled with the thermal parameter (the product of slab age and vertical descent rate) and h. Arcs with shallow h ($80 to 100 km) yield a larger range in slab surface temperature (up to $200 C between volcanoes) and more variable magma compositions than arcs with greater h ($120 to 180 km). This diversity is consistent with coupling of the subducting slab and mantle wedge, and subsequent rapid slab heating, at $80 km. Slab surface temperatures at or warmer than the H 2 O-saturated solidus suggest that melting at the slab surface is common beneath volcanic arcs. Our results imply that hydrous melts or solute-rich supercritical fluids, and not H 2 O-rich aqueous fluids, are thus the agents of mass transport to the mantle wedge.

Our increased knowledge of the nature and extent of overstepping of metamorphic reactions has challenged our understanding of the tectonic processes associated with metamorphism. The advent of mineral inclusion thermobaRamantry has allowed petrologists to better study metamorphic overstepping as thermobaRamantry relies on mechanical equilibrium and empirical equations of state as opposed to chemical equilibrium. In order to assess the role overstepping plays in subduction zone processes in particular we applied the new epidote in garnet (EpiG) Raman barometer, the quartz in garnet (QuiG) Raman barometer, and textural and mineral assemblage diagram analysis to 5 garnet-bearing blueschist samples from a Southeast-Northwest transect across the blueschist unit in Sifnos, Greece. All the samples are garnet-glaucophane blueschists with variable amounts of epidote/zoisite, quartz and other accessory phases, but their textures and bulk compositions vary. Garnet in samples from the southern and middle sections are typically ~2 cm large and display a honeycomb texture with ample quartz and rutile inclusions. Garnets in the northern most section, however, are generally 2-5 cm in diameter, and poikoblastic with inclusions of quartz and rutile along with occasional zircon and epidote. Preliminary EpiG and QuiG results suggests that in samples SPH99-1, SPH99-2 and SPH99-5, from the southern section of the range, garnets grew at a pressure from 12 to 16 Kbar. EpiG and QuiG analysis of samples SPH99-7 and SPH99-8 from the northern section, however, reveal pressures of 21-22 Kbar. Rigorous thermodynamic modelling is underway and may reveal more differences in the P-T evolution of rocks across this subduction zone. The ~6-8 Kbar difference across the unit may be explained by: 1) the rocks subducting as a coherent unit and then titling nearly 90° during exhumation so that both the deepest rocks (highest P) and shallowest rocks are visible at the surface; 2) tectonic thinning and accretion of a large HP unit as a result of post subduction extension and subsequent unroofing; or 3) variable amounts of overstepping of the garnet isograd which caused garnet to grow at different pressures, despite the possibility of a uniform peak P and T across the region. Further analysis will help identify the correct model.

Some commonly referenced thermal-mechanical models of current subduction zones imply temperatures that are 100–500 °C colder at 30–80-km depth than pressure–temperature conditions determined thermobarometrically from exhumed metamorphic rocks. Accurately inferring subduction zone thermal structure, whether from models or rocks, is crucial for predicting metamorphic reactions and associated fluid release, subarc melting conditions, rheologies, and fault-slip phenomena. Here, we compile surface heat flow data from subduction zones worldwide and show that values are higher than can be explained for a frictionless subduction interface often assumed for modeling. An additional heat source––likely shear heating––is required to explain these forearc heat flow values. A friction coefficient of at least 0.03 and possibly as high as 0.1 in some cases explains these data, and we recommend a provisional average value of 0.05 ± 0.015 for modeling. Even small coefficients of friction can contribu...