Composition, Mixing, and Melting of the Mantle III: Posters
Presiding: K Lee, California Institute of Technology; J M Sinton, University of Hawaii
V51A-01 0830h
Evolution of the Icelandic Convective System: Correlation Between Numerical and Real Data
The problem of hotspot existence under Large Igneous Provinces (LIP) is very disputable. However, the results of numerical modeling show that a `hotspot scheme' explains rather reasonably what we have in reality. In this work we take the existence of the hotspot under the Icelandic Convection System with a temperature 1920°C for granted. Convection in the upper mantle was taken as a mechanism for the melting zone development. Temperature distribution with depth in the upper mantle was taken in such a way that in the absence of a hot spot a convection cycle was well-developed. We tried to model the evolution of Iceland and to understand how the plate geometry influences such parameters as the 'asthenolens' existence time, the degree of melt and its chemical composition, the 'asthenonens' location and geometry. Several runs for various volcanic shield thicknesses, starting from oceanic crust, 7 km (when Iceland did not exist), to continental, 30 km, have been completed. In this way we modeled the evolution of Iceland from its origin to the present state. The results show that as the shield thickens from 7 to 30 km: 1) the existence time for the 'asthenolens' decreases from ~ 47 to 40 Ma 2) the maximum melt volume decreases from 4.784×1012 kg/km3 to 2.546×1012 kg/km3 3) the degree of melt decreases from about 16% to 12% 4) the upper boundary of the 'asthenolens' deepens for more than about 10 km. Thus the results of our modeling show that the hotspot under the Icelandic Convection System is dying at present time. The model used will be developed to verify possible melting extents for the lithosphere rocks above the convection zone. Both the data for the melting zone extent and the system existence time are in rather good correlation with geologic and geophysical data. This work was supported by the RFBR (Grants No. 04-05-64276, No. 04-05-64107, and No. 04-05-64332).
V51A-02 0830h
Variations in Post-Glacial Erupted Volume and Lava Composition Along the Western Volcanic Zone, Iceland: Implications for Mantle Heterogeneity and Melting Processes
Recent mapping has defined the outcrop areas, estimated total erupted volumes, ages and major and trace element geochemical variations within the 170 km-long plate boundary zone of western Iceland known as the Western Volcanic Zone (WVZ). Excluding the GrÃmsnes flank field, 35 total eruptions have occurred along the WVZ in post-glacial time. However, 57 % of these eruptions, accounting for 63 % of the total volume were erupted within the first 2-3 ky (25 %) of post-glacial time. Generally similar production variations have been documented everywhere else in Iceland where sufficient data exist, and have been related to the effects of glacial unloading. However, there is debate concerning the extent to which the volume maxima reflect increases in mantle melting or crustal controls on eruptive behavior. The variations in erupted volume along the WVZ approximately correlate with variations in incompatible element ratios that are best explained by melting processes. The lowest Zr/Y ratios coincide with the production maximum in early post-glacial time, suggesting higher total melting in the period immediately following deglaciation. However, production rates have continued to decline and Zr/Y ratios have continued to increase in the period from 9 Ka to present, features not predicted by current models of deglaciation effects. Superposed on the temporal evolution are spatial variations; all lavas from the northern third of the WVZ have lower Zr/Y at all ages than those from farther south. Although a range of complicated melting processes can produce this result without invoking significant mantle heterogeneity, the most likely interpretation, based on preliminary geochemical modeling is that the mantle underlying the northern WVZ is generally more depleted in incompatible elements than that farther south, despite lying slightly closer to the center of the Iceland hotspot. Our results support the general model for melting increases in the period immediately following deglaciation, but for the WVZ we also require significant along-axis mantle heterogeneity and a long-term melting decline that may be unrelated to deglaciation effects. Data from the Reykjanes Peninsula indicate similar long-term chemical trends, suggesting that overall decline in melt production characterizes all of western Iceland, and possibly of the Iceland hotspot generally over the last 9,000 years.
V51A-03 0830h
A Geochemical Comparison of 2 Recent EPR Eruptions North and South of the Clipperton Fracture Zone
A fresh looking lava flow was discovered at 10°44'N north by Alvin divers during the FIELD (Nov 2003) cruise to the "magma starved" segment of the East Pacific Rise (EPR) north of Clipperton Fracture Zone (J. Voight, Chief Scientist). 210Po-210Pb dating demonstrates that an eruption occurred here within a few months of the cruise (van der Zander et al., Fall 2004 AGU abstract). The magmatically starved nature of this eruption locale provides for an interesting contrast with the 1991-92 lava flow at 9°50'N on a magmatically robust segment south of the Clipperton. Results of geochemical analyses reveal interesting contrasts in their petrogenesis. Compositional heterogeneity in the 10°44'N lava flow is much higher than in the 9°50'N flow (3.2 vs. 1.6 using the index in Rubin et al., EPSL, 2001), consistent with the much lower magma supply and great extent of crustal processing to the north. U-Series disequilibria signatures are very different between the two lava flows: (210Pb/226Ra) from the area basalts all show (210Pb)-excesses ranging from about 10 to 50 % excess, whereas 9°50'N sheet flow basalts show equilibrium up to 10 % deficits values. Within flow (210Pb/226Ra) variations in the 10°44'N samples correlates with indices of fractionation such as SiO2/MgO, FeO*/MgO and trace element concentration, and is thus probably related to crustal processes. (226Ra/230Th) from the northern basalts are lower and more homogeneous than the south of Clipperton basalts, which may relate either to melting/melt transport conditions or decay during magma chamber residence. (230Th/238U) on the other hand is higher than for the 9°50'N samples, which can be interpreted as lower upwelling or melting rate at 10°44'N. However, the higher (230Th/238U) of the latter also correlates with the higher Th/U ratios, suggesting a source enrichment effect may also be playing a role on (230Th/238U). Ba/Th and 87Sr/86Sr of the 10°44'N are also elevated, whereas K/Ti ratios overlap in both flows. We discuss these differences and explore the possibility of crustal contamination as a contributor to the geochemical signatures as well as time constraints on crustal and melting processes at the NEPR.
V51A-04 0830h
Upper mantle seismic structure beneath the Azores hotspot from finite frequency tomography
The Azores archipelago represents one of the classic hotspots that interact with the mid-ocean ridges. While geochemical studies have provided clear evidence for chemical anomalies in the mantle from basalts collected on the islands and at the adjacent mid-Atlantic ridge, the seismic structure beneath Azores has never been imaged directly at regional and local scales. In this study, we construct the first high-resolution seismic velocity models of the mantle structure beneath the hotspot from teleseismic body waves recorded by seismic stations on the Azores islands. Unlike in conventional tomography based on ray theory, the three-dimensional sensitivity kernels of body wave travel times are used to account for wave front healing, scattering, and other diffraction effects of realistic seismic waves. The results from ~150 P wave relative travel times show a pronounced low velocity anomaly in the shallow mantle (~100 km depth) along the Azores islands. The low velocity anomaly becomes elongated in the NE-SW direction at greater depth (100-200 km). The model shows a columnar low-velocity anomaly that extends from 200 km to 400 km depth northeast of Terceira. These results are consistent with geochemical observations and provide seismic evidence for the proposed mantle plume-ridge interaction model, in which the plume conduit bends toward Azores triple junctions, supplying melt along a lateral sublithospheric channel.
V51A-05 0830h
Olivine-Melt Equilibrium and the Oxygen Fugacities of Lavas from Hawaii.
Oxygen fugacities of magmas provide information about the redox states of their mantle source regions. Therefore, oxygen fugacities for magmas erupted in different tectonic environments can be used to map variations in the redox state of the underlying mantle. Previous data have been interpreted to indicate that the mantle source of MORB is reduced relative to that of OIB, implying that plumes are oxidized relative to MORB mantle. However, oxygen fugacities calculated from analyzed Fe2O3/FeO for lavas from Hawaii (ΔQFM=0.363 to -1.812, ave= -0.705) and Loihi (ΔQFM=0.293 to -2.121, ave= -0.994) fall in the range of those for MORB (ΔQFM=0.402 to -2.732, ave=-1.304). This suggests that the mantle source of Hawaiian magmas is similar to that of MORB. However, there is some dispute about the validity of oxygen fugacities calculated from analyzed Fe2O3/FeO of lavas. Therefore, we have also calculated oxygen fugacities for lavas from Hawaii, and for MORB samples using olivine-melt equilibrium. For Hawaiian samples we obtain ΔQFM=0.586 to -2.059, ave=-0.779, values similar to those obtained from analyzed Fe2O3/FeO providing confidence that the values accurately represent magmatic oxygen fugacities. For MORB samples (mostly from the EPR and MAR) we obtain ΔQFM=0.190 -2.594 ave=-0.713, values that are also similar to those obtained from analyzed Fe2O3/FeO. These results therefore indicate that the redox state of the mantle source of Hawaiian magmas is virtually identical to that of the MORB source. This confirms results obtained from Icelandic lavas, and strongly suggests that the redox states of the mantle sources of OIB and MORB are essentially identical at the depths of magma segregation. It follows that there is no evidence that the redox state of the lower mantle (or deep mantle) differs from that of the upper mantle if OIB are derived from deep mantle plumes. Unless the redox states of plumes are affected by exchange with surrounding the mantle during ascent, this may provide evidence that the lower mantle is relatively oxidized.
V51A-06 0830h
Time Series Periodicity for Lava Compositions of Shield Volcanoes Above Hot Spots
Thoileitic sheets for the Siberian Platform trapps and basalts of Hawaii and Iceland have similar time series for lava profiles in the following respects: 1. periodicity 2. interruption of distribution functions both at the suit boundaries and within continuous discharge of active volcanoes 3. presence or absence of time trends 4. substantial difference in distribution function types for profiles resulted from fissure conduit located at 30-40 km distance from each other. Wavelet analysis shows the functions for tholeiitic shield volcanoes are similar for the time intervals form 20 to 20000 years. Spectral characteristics of petrogenous and trace components as well as isotope ratios show coherent and incoherent values for frequency spectra. The main tendencies for evolution of compositions in completed sequences are similar. The spatial zoning depends on structural and geodynamic conditions of a lava sheet formation and its size. The trapps of the Siberian Platform can be regarded as unique structures from this viewpoint. The most various are time and spatial series for the Hawaiian Island. Two types of time series evolution can be distinguished: `Icelandic' and `Hawaiian'. It is interesting to note, that for the Western Siberian Slab the former type is typical, and for the Siberian Platform - the latter. The reasons for this phenomenon are still not clear. This work was supported by the RFBR (Grants No. 04-05-64276, No. 04-05-64107, and No. 04-05-64332).
V51A-07 0830h
Oxygen Fugacities of Lavas from the Galapagos Islands and the Galapagos Spreading Center
Oxygen fugacity is one of the most important intensive variables that controls the phase relations and compositions of precipitating minerals. It is accepted that oxygen fugacity reflects the redox state of the mantle source region. Therefore oxygen fugacities can be used to probe the mantle redox state and determine mantle heterogeneity. Plume-related magmas and mid-ocean ridge basalts (MORB) have been shown to fall within the same range (ΔFMQ = -2.732 to 0.402) suggesting source regions with similar redox states. The Galapagos Islands are a result of plume-related volcanoes and located on thin, young crust created by the nearby Galapagos Spreading Center (GSC). This study focuses on the differences between the redox state of the Galapagos plume related magmas and the nearby GSC. The Galapagos Islands show unsually high iron concentrations, which should be reflected in the fO2 values of lavas. Values of fO2 calculated from Fe2O3/FeO ratios for GSC lavas ranged from ΔFMQ = -1.801 to 0.402. Values calculated from coexisting olivine and melts show excellent agreement with a range of ΔFMQ = -1.818 to 0.069. There is reasonable correlation between oxygen fugacity and Mg, consistent with the idea of crystallization controlling oxygen fugacity. Oxygen fugacities were calculated from coexisting olivine and melt samples from Roca Redonda, Fernandina, and Volcan Darwin on the Galapagos Archipelago. Melt samples were based on groundmass analyses rather than glasses. The results ranged from ΔFMQ = -1.962 to -0.059, similar to values for MORB from the GSC, despite the unusually high Fe8.0 and low Si8.0 contents. The latter have been used to suggest that the island magmas are generated at greater depths than MORBs. This study supports other work suggesting that OIB from deeper mantle sources have similar oxygen fugacities to MORB from upper mantle sources. This suggests that both sources have similar redox states, or that differences are not observed in oxygen fugacities of magmas originating from these sources. In the case of the Galapagos Islands this is surprising, given other evidence for differences in melting temperature and depth and source region composition.
V51A-08 0830h
First Principles Calculations of Point Defects in Major Mantle Minerals
First-principles calculations within the local density and pseudopotential approximations are performed to investigate the effects of pressure on the energetics and structural behavior of charged vacancy defects in MgO and MgSiO3 perovskite. The simulations are performed for a supercell containing a couple of hundreds of atoms with their positions being fully optimized. In particular, the formation and migration energies of cation and anion vacancies are shown to substantially increase over the pressure regime of Earth's mantle. Our results thus suggest that pressure should suppress intrinsic diffusion mediated by ionic vacancies in these two major mantle minerals over the mantle pressure regime. The calculated three-dimensional datasets for atomic displacements and electron charge density are explored in details using an interactive visualization system. Although the atomic and electronic structures are mostly distorted in the close vicinity of the defects (i.e., in the region covering up to the nearest and next-nearest atoms), the effects are not negligible at farther distances.
V51A-09 0830h
Electrical Conductivity and Tomographic Imaging of Olivine-FeS Partial-Melts
The presence, distribution, and composition of melt affect the physical properties of polycrystalline ultramafic rock and are important to our interpretation of the Earth's lower crust and upper mantle, and to our understanding of planetary core formation via liquid-metal segregation. A key issue in models of planetary core formation is the interconnectness of molten iron-sulfides in contact with silicates at high temperature and pressure. Olivine-FeS partial-melts are also considered to be possible explanations for anomalously high conductivity regions beneath mountain ranges such as the Pyrenees and Andes. The interconnectivity and tortuosity of the melt phase, in combination with the properties of the individual melt and crystal phases, have bearing on the extractability of the melt, and on the rheology, and electrical conductivity of the bulk material. We have begun an integrated study of the electrical conductivity-texture-permeability relationships of olivine-sulfide partial-melt samples. Olivine-sulfide partial-melts containing 0, 1, 3, 6, and 10% by weight non-wetting compositions (Fe64S36) and wetting compositions (Fe34S19Ni47+O2) in a San Carlos olivine matrix (Fo91) have been synthesized in a piston cylinder apparatus at 1250 C and 1 to 2 GPa. Electrical conductivity measurements of the partial-melt and the individual melt and crystalline phases have been performed in a 1-atmosphere gas-mixing furnace up to 1400 C. Additional measurements in solid medium-pressure apparatus (D-DIA, piston cylinder) have begun. Samples are characterized using X-ray microtomographic (XRCT) performed at the Advanced Light Source with spatial resolution approaching 2 microns. Determination of the 3-D structure and interconnectedness of the melt phase, combined with the electrical conductivity measurements have been used to estimate the permeability of the mixtures at various experimental conditions. Results indicate sulfur fugacity is an important parameter controlling the wettability and interconnectivity of the melt phase. This work was performed under the auspices of the U.S. Department of Energy by the University of California Lawrence Livermore National Laboratory under contract W-7405-ENG-48 and supported specifically by Laboratory Directed Research and Development funding
V51A-10 0830h
Lattice dynamics and thermodynamics of hexagonal-close-packed iron at high temperatures and pressures
We performed linear-response Linear-Muffin-Tin-Orbital (LMTO) calculations to understand and predict the lattice dynamical and thermal properties of hexagonal-close-packed iron at high temperatures and pressures. The phonon dispersion and phonon density of states have been calculated at different volumes and various c/a axial ratios, which show good agreements with available experimental data. We also calculated the thermal conductivity and electrical resistivity at different pressure. We derived the Hemlmholtz free energy functionals based on the LMTO calculations, and have further applied to establish the thermal equation of state, bulk modulus K0, dK0/dT, and thermal expansion coefficients at high pressures and temperatures. The variations of c/a ratios with temperature and pressure have been predicted. We also used the particle-in-cell approach to examine the thermal properties based on tight-binding total energy calculations, and made a detailed comparison with lattice dynamics calculations and experiment. The influence of anharmonic effects has been examined. This work was supported by US Department of Energy ASCI/ASAP subcontract to Caltech, Grant DOE W-7405-ENG-48 (to REC).
V51A-11 0830h
Density Measurement of Liquid FeS Under High Pressure and High Temperature
Sulfur is considered one of the possible light elements in the core which might be responsible for the density deficit. We studied the liquid state of sulfur in iron due to sulfur's lack of amount in the mantle; easiness to alloy with iron; and the predicted 5% ~ 10% amount of this light element in the core (Ahrens, 1979; Sherman, 1997). Restricted by the modern development of the multianvil high pressure experimental equipments, the experiments are limited at a lower pressure range (<30GPa) comparing with the outer core pressure condition. Therefore, extrapolation of data derived at low pressure range to the condition of the outer core (130-330GPa) has to be applied and may produce results which are way far from the true numbers. However, at the point while the techniques are limited, studying the physical properties of the molten FeS at relatively low pressure still provides us a better picture of the physical behaviors of the liquid outer core comparing with data derived from solid state FeS experiments. The lack of melt density data at low pressure provides another motivation for us to study the physical properties of melt. The radiography (shadowgraphy) system on Beam Line X17B2, NSLS at the Brookhaven National Laboratory is an add-on system attached to the in situ x-ray beam line setup. It includes a YAG fluorescent screen, an optical mirror, focusing-magnification lenses, and a CCD camera and/or a video camera. Before the melting temperature, the radiograph system yields a maximum 1% difference in density comparing with the data collected by the traditional x-ray diffraction method. We have successfully examined liquid FeS samples by applying this technique at the NSLS. With a sapphire (Al2O3) sphere surrounded by FeS powder. The image of the sphere was clearly shown due to the absorption coefficient difference between these two materials. The density fitting method developed by our group has produced convincing data. The preliminary results of the density measurements of molten FeS show that the derived liquid density variation for the same sample remains under 1%. This study has collected in situ high pressure and high temperature x-ray diffraction data of the FeS sample up to 4GPa and 1400°C. Combined with the derived density data, the equation of state of the liquid FeS can be constructed.
V51A-12 0830h
Hydrocarbons Encapsulated in Diamonds From China and India
We examined a large number of diamonds from a kimberlite pipe located in Fuxian, China, and alluvial diamonds from Panna, India. We selected 6-10 diamonds from each locality based on certain characteristics: they are white, brilliant, mostly devoid of mineral inclusions, fracture-free, many contain microscopic bubbles, some display etched circular patterns. These diamonds were examined under ultraviolet (UV) light using a fluorescence microscope, then, investigated using a Nicolet 6700 FT-IR spectrometer. Several diamonds emit blue fluorescence when excited with UV light, while others appear dim because they are not fluorescent. It is the latter that render the included bubbles clearly visible, glowing as yellow and blue spherules within the dim diamond host. These fluorescent bubbles are probably filled with hydrocarbon fluids of variable compositions. FT-IR spectra of diamond typically show absorption due to intrinsic diamond lattice vibrations. We found in most of our diamonds used in this study an additional, outstanding group of absorption bands located just below the wavenumber 3000. Peak positions in this region correlate well with symmetric and asymmetric stretching of methylene and methyl groups, attributable to H bonded to C atoms. Comparing them with standard spectral shapes, we found a good match with an alkane molecule composed of saturated aliphatic hydrocarbons. Our observations provide evidence that hydrocarbons might be important components in the deep mantle, but, to transport them up to Earth's surface would require strong capsules which, perhaps, only diamond could provide.
V51A-13 0830h
Mantle Plume Temperature Variations Immediately Following Continental Breakup of the Northern North Atlantic
The amount of melt generated by mantle decompression beneath an oceanic spreading centre and hence the oceanic crustal thickness is controlled in part by the temperature of the mantle. By measuring the thickness of the oceanic crust formed immediately after breakup of the northern North Atlantic during the early Tertiary, we are able to deduce the maximum elevated mantle temperatures caused by the presence of the Iceland mantle plume. Crustal thickness variations are caused by temporal variations in the mantle plume temperature: at the present Reykjanes Ridge spreading centre the plume temperature pulses on a 3-5 Myr timescale with temperature variations of c.30 K. We show results from two long-offset profiles acquired over oceanic crust; firstly a 170km line perpendicular to the Faroes rifted continetal margin where oceanic spreading developed close to the Iceland mantle plume; and secondly, a 200km line perpendicular to the Hatton rifted continental margin where oceanic spreading developed 800km south of the plume. Each survey recorded long-offset refractions and reflections on OBS (Ocean Bottom Seismometers); 25 instruments, with a spacing of 2-3 km, were used for the Faroes line; and 45 instruments, with a spacing of 4-10 km were used for the Hatton-Rockall line. Accurate information for sediment velocity and thickness was acquired for the Faroes profile using a 12 km long streamer; whilst adequate sediment information was determined for the Hatton-Rockall profile using a 2.4 km streamer. By incorporating sediment structure into a joint reflection and refraction tomographic inversion of the wide-angle OBS data, we have been able to map crustal thickness across the oceanic crust in both regions. Crustal sections across the Faroes and Hatton lines cover the first 14 Myr and 17 Myr respectively, corresponding to the time interval from continental breakup through to mature seafloor spreading. With no apparent decrease in spreading rate observed thinning of the crust oceanwards suggests a mantle temperature decrease of 50 K for the Faroes profile and 70 K for the Hatton profile. For both profiles early oceanic formation seems to have been dominated by a transient high temperature anomaly, while for later spreading there is more dependance on the distance of each profile from the plume. Mantle plume temperature variations during this period would have caused rapid changes in uplift of the north-west European margin and probably controlled Tertiary sedimentation patterns west of Britain. The iSIMM Scientific Team comprises NJ Kusznir, RS White, AM Roberts, PAF Christie, R Spitzer, N Hurst, ZC Lunnon, CJ Parkin, AW Roberts, LK Smith, V Tymms, J Eccles and D Healy. The iSIMM project is supported by Liverpool and Cambridge Universities, Schlumberger Cambridge Research, Badley Technology Limited, WesternGeco, Amerada Hess, Anadarko, BP, ConocoPhillips, ENI-UK, Statoil, Shell, the NERC and DTI.
V51A-14 0830h
U-series disequilibria in volcanic rocks from the Futuna spreading centre (North Fiji Basin)
The North Fiji basin is the largest active back arc basin of the SW Pacific. It is 12 m.y. old and characterised by a regional upper mantle thermal anomaly (Garel, 2001). The 200Km long Futuna spreading centre (Pelletier et al., 2001) is located west of Futuna and Alofi islands at the border between Lau and North Fijian Basins. This ridge is composed of an axial valley spotted with numerous seamounts. It starts from the north of the Fiji platform (15° 40'S) and ends in the north Fiji transform zone, northwest of the Futuna and Alofi islands (13° 35'S). New U series measurements were performed by TIMS on fresh dredged glassy samples. As the topography of this spreading centre exhibits a peculiar structure, the geochemical signature is also unusual. U and Th concentrations range between 0,09<U<0,51 and 0,31<Th<2,16, with a corresponding large variation in Th/U ratio 3,4<Th/U<5,4. Such variations in Th/U ratios likely reflect the mixing between different long-lived sources. (230Th/238U) ratios vary from 1.15 to 1.42. Three samples have 230Th excesses >30%, six have a 230Th excess within 25 ~ 5% and three have 230Th excess of about 15%. These 230Th excesses are typical of MORB or OIB source melting and not a back arc volcanic processes. It is also symptomatic of a melting in presence of garnet. In the equiline diagram, samples fall on a linear correlation indicating a mixing between two end-members with different Th/Usources ratios. The trend observed in this diagram requires the involvement of an enriched component, likely an EMII component, as defined by the trace elements and Sr-Nd-Pb isotope compositions (Benoit et al., in prep). This trend falls within the OIB field and plots close to Samoan samples analysed by Newman et al. (1984). This trend demonstrates the influence of a melt may be even more enriched than the samples of Savai'i volcano. This EMII melt mixes, within the upper mantle, with the Indian ocean like component as it was already established that it is present below Lau and North Fiji Basins spreading centres (Loock, 1992; Turner et al., 1992; Hickey-Vargas et al, 1995; Zhang et al., 1999).