V53C-01 13:40h
The Role of Recycled Oceanic Crust in Mantle Plumes -Revisited
The role of recycled material in mantle plumes is difficult to quantify on the basis of incompatible trace elements and isotopes because of the great variability of subducted material. Another approach is to use major elements and compatible trace elements because these are more uniform in the mantle and are strongly controlled by the phase petrology of melting. Subducted crustal lithologies invariably differ from mantle peridotite, and this introduces olivine-free lithologies such as pyroxenites and eclogites into the mantle. Our massive study of olivine phenocrysts and trapped melt inclusions shows unusually high Ni and Si contents in many recent primary Hawaiian magmas. Similar compositions are found in the Canary Islands, W. Greenland, and the Siberian flood basalts. These magmas are not in equilibrium with an olivine bearing source under thick lithosphere (more than 100 km) typical of these localities, because an olivine-pyroxene assemblage would buffer both Ni and Si at lower levels. In contrast, magmas from plumes located under thin lithosphere, such as Iceland or Azores show no significant Si and Ni excess, and they could be in equilibrium with a shallow, olivine-bearing source. High-Si magmas can be produced by melting of eclogite, but this does not yield high Ni contents. Therefore, the eclogite-derived melt must acquire high Ni by converting surrounding peridotite to a solid pyroxenite, which ultimately melts a shallower level. Because unreacted peridotite may also begin to melt at shallow depths, this results in mixed melts derived from (secondary) pyroxenite and peridotite. In settings of thick lithosphere, the amount of peridotite-derived melt will be relatively small. Therefore, the recycled component represented by pyroxenite-derived melt may dominate. In settings of shallow melting, the peridotite will melt more extensively, and the signal from the recycled component will be diluted. Quantitative modeling shows that over half of the Hawaiian magma volume formed during the last 1 Myr came from secondary pyroxenite representing the recycled oceanic crust. The results are consistent with a plume with potential temperature of 1600 deg.C containing about 20 percent of recycled oceanic crust in the central part. These results are also consistent with estimates of volcano volumes, magma volume flux, and seismological observations. In the context of this model, the recent increase in Hawaiian magma flux is produced by an unusually high proportion of recycled crustal material in this part of the plume
V53C-02 13:55h
The Metasomatic Alternative for the Origin of OIB: a Model which Reconciles Experimental Petrology and Geochemistry
Variation of trace element and isotopic ratios in OIB is commonly ascribed to the recycling of ancient oceanic crust associated with crustal or pelagic sediment assimilation. However this model based on geochemical arguments is in opposition with experimental petrology data. Partial melts of oceanic-crust lithologies produce silica saturated liquids whereas many oceanic island rocks are characterized by silica undersaturated compositions. Experimental data indicate that only partial melting of peridotite in presence of carbonate [1] or of pyroxenite [2] produce liquids which are close to the nephelinite or basanite major element compositions observed in oceanic islands. In this way, recycling of subducted oceanic basal lithosphere enriched by metasomatic veins seems to represent a convincing alternative for the source of OIB [3]. However, this hypothesis does not explain the formation of isotopic heterogeneity observed in OIB. Chemical variations observed in Cantal basalt (France), interpreted as the result of a lithospheric metasomatic mechanism [4], allow us to constrain the chemical evolution of a metasomatic agent within basal lithosphere. These data demonstrate that - 1) fractionation of trace element ratios (U/Pb, Th/Pb, Rb/Sr, Sm/Nd, Nb/La .) necessary to generate, after subduction and isolation, the EM and HIMU components can be explained by metasomatic process and - 2) partial melting of veins-plus-enclosing lithospheric mantle produce basalt with composition perfectly similar to major and trace elements composition observe in OIB. This suggest that isotopic and trace element variations observed in basalts from individual oceanic islands may more likely be the result of melting heterogeneous, metasomatised, subducted oceanic lithosphere rather than a mixture of chemically distinct mantle reservoirs. End-member isotopic compositions would correspond to the extreme trace element fractionation generated by metasomatic process within the lithosphere. The new interpretation of OIB sources proposed here requires a re-evaluation of the processes that control chemical evolution of Earth's mantle reservoirs and plumes geochemical tracers. Recycled metasomatised lherzolite may be the major mantle component sampled by OIBs; recycled oceanic crust and sediment may be less common in OIB sources than is commonly assumed. The metasomatic hypothesis is entirely consistent with isotopic heterogeneity observed in the source of oceanic basalts and experimental data which indicate that the most plausible source of OIB material is pyroxenites. [1] K. Hirose (1997) Geophys. Res. Lett. 24, 2837 [2] M.M. Hirschmann et al. (2003) Geology, 31, 481. [3] Y. Nui, M. O'Hara. (2003) J. Geophys. Res.,108, B4, 2209. [4] S. Pilet et al. (2004) Geology, 32, 2, 113.
V53C-03 14:10h
A Fossil Mantle Plume under the Emeishan Flood Basalts: Integration of Geology, Geophysics and Geochemistry
The plume hypothesis is now challenged because some fundamental aspects predicted by the modeling of plumes are found to be lacking in classic regions like Iceland and Yellowstone. Instead of invoking a bottom-up process, some researchers favor a top-down hypothesis for the formation of large igneous provinces (LIPs), in which shallow lithospheric processes may fuel melt production. Seismic investigations and tomographic models help trace mantle plumes in modern, active hotspots, but are of limited benefit in identifying ancient plumes, mainly because geophysics provides us with a snapshot of the present-day Earth's structure. Consequently the geological footprint associated with thermal anomalies are the clues to tracing ancient plumes. According to some theoretical models, pre-volcanic lithospheric uplift is the most important criteria used to identify the presence of plumes. The lack of such evidence, on the other hand, is an argument against the involvement of plumes in the formation of LIPs. Recent examination of the middle-late Permian sedimentology in southwest China reveals kilometer-scale lithospheric doming prior to the Emeishan flood volcanism (He et al., 2003). This, and correlations between diverse, independent parameters involving crustal doming, paleo-geography, sea level change, mantle melting mechanism and crust-mantle structure, provide evidence for a fossil mantle plume under the Emeishan LIP. Specifically, the consequences of plume-lithosphere interaction include: (a) pre-volcanic uplift including thinning of marine carbonates, a marine to sub-aerial transition, local provenance of clastic sediments, and a marked erosional unconformity, evident as palaeokarstic surfaces on the marine carbonates; (b) a domal structure (700 km radius); (c) variations in the thickness of volcanic rocks across the domal structure; (d) variations in flood basalt geochemistry from the center to the edge of the domal structure that are interpreted as high temperature melts in the center and lower temperature melts at the edge; (e) gradual decrease in crustal thickness from the center to the margin of the dome; and (f) the presence of high velocity lower crust (20-30km) immediately beneath the domal structure which is consistent with significant melt production and possible underplating/intrusion into the lower crust.
V53C-04 14:25h
The Mantle Plume Hypothesis Pro and Con: Evidence from Earth's Most Voluminous Large Igneous Provinces
Mantle plumes are upwellings of large volumes of mantle material in focused conduits, the leading ends of which are referred to as plume heads. Large igneous provinces (LIPs) are suspected to form from magmatism resulting from plume head decompression melting, but, evidence for this theory for the origins of LIPs is mixed. We have now reached the point of having either to modify the theory to fit characteristics of individual LIPs or to abandon the theory and search for a more unifying explanation. A case study of the two biggest LIPs on Earth - the Ontong Java Plateau (OJP) in the western equatorial Pacific Ocean, and the Kerguelen Plateau / Broken Ridge (KPBR) in the southern Indian Ocean - allows us to examine key predictions of mantle plume theory, including: (1) subaerial eruption of large portions of oceanic LIPs (2) large extents of partial melting in the plume head, resulting in tholeiitic basalt-type magmas, coupled with (3) rapid formation of the LIP, (4) post-formation subsidence comparable to normal oceanic lithosphere, and (5) the presence of a hotspot track and/or an active hotspot. The KPBR formed largely above sea level over a protracted time period ($\sim$120 Ma - present) in the growing Indian Ocean basin. Early Cretaceous melts were derived from a heterogeneous source, complicated by subsequent local assimilation of continental crust. Most lavas recovered from the plateau are tholeiitic, but alkalic and evolved volcanics occur in several, widespread locations. Subsidence of the plateau has followed predictions for normal oceanic lithosphere. A prominent hotspot track, the Ninetyeast Ridge, connects Broken Ridge with Early Cretaceous continental basalts on the eastern margin of India. The Kerguelen hotspot is still active today, creating Heard and MacDonald Islands on the central plateau. The OJP was constructed well below sea level on existing Pacific lithosphere. Nearly the entire volume of magma is believed to have been created instantaneously, at $\sim$120 Ma, from large degrees of partial melting ($\sim$30%) of a homogeneous source. Melting extents were high, so either a large temperature anomaly, or a major decompression event is required because volatile contents were low. It has subsided relatively little, or erratically, since its formation. No known hotspot track is associated with the OJP, nor is any active hotspot. Some of these observations agree with expectations of a mantle plume head origin, but several are contrary to predictions. Alternative mechanisms for the formation of LIPs, including extraterrestrial (i.e. bolide impact) or tectonic causes, are also problematic. A few of the obvious concerns are (1) large bolide impact events might not occur often enough to account for the number of known LIPs, (2) tectonic origins for LIPs ignore the presence of LIPs on planetary bodies where evidence for plate tectonics is nil or scant. A common mechanism for the formation of LIPs is highly desirable, yet, at present, all existing hypotheses appear in some way deficient.
V53C-05 14:40h
High Volatile Content and Shallow Melting at the end of the Siberian Flood Basalts: Experimental Results
Constraints on the depth and temperature of melting for flood basalt lavas are critical for evaluating melting models. Obtaining primary melting conditions through straightforward experimental petrology is unfeasible for most flood basalts because of significant secondary processing. The final lavas in the Siberian flood basalts (SFB), however, are candidates to be nearly primary magmas, having experienced only olivine addition or subtraction following mantle melting. These ultramafic lavas, predominantly meimechites, form a 1400 m stack at the top of the SFB section in the Maymecha region. One-atmosphere, piston-cylinder, and multi-anvil experiments have been performed on a synthetic analog of a meimechite olivine melt inclusion to determine its pressure and temperature of original mantle melting. Although the major element compositions of meimechite melt inclusions have the same trends as do the bulk rocks, the alkali contents of the melt inclusions are systematically higher, suggesting that the lavas have lost alkalis in a post-eruption serpentinization event. Meimechites are enriched in incompatible elements, particularly the LREE (Basu et al., 1995; Arndt, 2003), and are hydrous, evidenced by groundmass phlogopite. Analysis of meimechite major element trends indicates that compositions with approximately 25 wt% MgO represent liquid compositions. The experimental composition has 25.5 wt% MgO, 8.3 wt% CaO, and Mg\# 77. The experimental composition with 2% water is multiply saturated on its liquidus at 3.0 GPa and about 1600\deg C with olivine, sub-calcic augite, garnet, and Ti spinel. This multiple saturation is at an anomalously high temperature, interpreted to imply an even larger volatile concentration in the source region, which would likely lower the temperature of melting by as much as 100 to 150\deg without largely changing pressure. Pre-eruptive water and carbon dioxide content estimates are being obtained in current research analyzing melt inclusions. The multiple saturation point is interpreted as the conditions of batch melting in the mantle, or the mid-point of melting during an adiabatic ascent. The experimental meimechite therefore is inferred to have originated from mantle melting at about 100 km depth, surprisingly shallow for melting in a plume under an ancient continental lithosphere. Alternative models for melting are also required to explain the strong geologic evidence for subsidence during the first kilometer of eruption (Federenko and Czamanske, 1997). We suggest that the lower lithosphere delaminated, pulling topography down and allowing shallow melting. As it sank and heated, the lithosphere would have dewatered, providing volatiles for the meimechite source. The volatile input from the lithosphere can also explain the unusual meimechite trace element compositions. We also note that orthopyroxene instability in the source region, implied by its absence at multiple saturation, is further evidence for hydrous metasomatism of the mantle. This model is not contradictory to a plume model, but requires an upwelling weak enough (with relatively little buoyancy) to allow the delaminating material to cause subsidence in the lithosphere.
V53C-06 15:05h
Hundreds of Rimmed Circular Structures on Venus Formed From Impacts Before 3.9 Ga, Not From Young Plumes
Venus displays hundreds of circular structures, with topographic rims 10-2,000 km in diameter, that have the morphology, cookie-cutter superposition, and log frequency/log size distribution required of, and unique to, impact craters and basins. They nevertheless are assumed to be endogenic by specialists. Many have interior central or ring uplifts or broad, low volcanic constructs. Many are multiring. Old uplands are saturated with the structures, which there are variably eroded, whereas lowland structures are partly to entirely buried. The largest (Artemis, Heng-O, and Quetzalpetlatl, rim diameters 2000, 900, and 800 km) are among the youngest. Analogy with dated large, and similarly relatively late, Imbrium impact basin on the Moon requires ages greater than 3.85 or 3.90 Ga. Venus preserves much of its surface of late-stage main planetary accretion. Early investigators of Venusian radar imagery accepted the possible impact origin and great age of the circular structures, but in the late 1980s impact was replaced, almost without analysis, by plume conjectures. Almost all specialists now assume that Venus has a thermal structure and heat loss comparable to that of Earth, and that its only impact structures are mostly-pristine small to midsize (maximum rim diameter, 270 km) craters with an assumed age of less than 0.5 or 1.0 Ga. (Ages as old as 3.9 Ga are advocated here for these young craters.) The older circular structures are rationalized as produced by mantle plumes and upwellings that deformed crust and upper mantle from beneath, with or without extrusion of subordinate lava, and that magmatically and tectonically resurfaced Venus in a brief period before the late impacts. Extrapolation of plume conjecture to Venus from Earth has little merit. Terrestrial plume speculation is based on assumptions whose predictions have been consistently falsified. Not only do plumes probably not exist on Earth, but even the least-constrained attributions of geologic and tectonic features to them do not include circular structures that in any way resemble those of Venus. Conversely, Venusian conjectures neither address nor account for circularity and superpositions. The hot-Venus assumption behind young-surface speculation also is dubious. The lack of a magnetic field on Venus (its core is likely solid), the positive correlation of its topography and geoid (outer Venus is much stiffer than Earth), and its origin close to the Sun (less potassium, so much less early radiogenic heat), and other factors indicate Venus to be much colder internally than Earth below the depth of influence of greenhouse atmosphere. The most eroded and breached, or buried, of the quasi-pristine craters are discriminated only arbitrarily from the best-preserved of the ancient, and mostly larger, circular structures. From those in turn, there are all gradations back to the deepest-eroded, or the most-buried, structures of the old family. Broad, low volcanic constructs (unlike any terrestrial volcanoes) inside impact basins likely are products of cogenetic impact melts. Other large, low volcanoes also are circular, are isolated, and may be of impact melts that buried their basins. Broad tessera-surfaced plateaus are of layered rocks, display deformation and topography indicative of outward gravitational spreading, and may have formed from ancient impact-melt lakes. Venusian lowlands are floored not by young lava plains but by ancient sediments, probably including deposits in a transient ocean, derived from uplands. The plains are speckled with mud volcanoes (not lava cones) that, like minor deformation of the sediments, are due to top-down heating by the evolving atmosphere.
V53C-07 15:20h
Large topographic rises, coronae, large flow fields and large volcanoes on Venus: Evidence for mantle plumes?
Voluminous volcanic deposits at topographic rises, coronae, large flow fields and large volcanoes have led these features to be linked to mantle plumes. Topographic rises have broad, swell-like topography (typically $\sim$1800 km across, 1.5 km high), large positive gravity anomalies, and associated volcanism. Coronae are circular to irregular features (typically $\sim$300 km across, 1.0 km high), defined by their fracture annulus and associated with uplift and volcanism, followed by subsidence. Large volcanic flow fields ($>$30,000 km2) and large volcanoes ($>$100 km in diameter) are sites of voluminous outpourings of lava. We interpret the variations in styles of volcanism, surface deformation, topography and gravity signatures to indicate differences in the nature of the underlying thermal upwellings that formed these features. Most topographic rises are likely to be formed by primary or deep-seated plumes, while coronae and probably most volcanoes result from shallower upwellings or secondary plumes. Extension clearly plays a critical role in the formation of large flow fields, coronae and some large volcanoes. We do not interpret large flow fields to be related to plumes. There are a similar number of primary plumes on Earth and Venus, but Venus has a much larger number of secondary plumes. The increase number of secondary plumes on Venus may result from the lack of slab cooling at the core-mantle boundary, its lack of a low viscosity zone or its stronger lithosphere.