V22A-01 INVITED 10:25h
Geochemical and Chronometric Data of Carbonate Veins Provide Insights Into Seawater-Ultramafic Rock Interactions
Ocean Drilling Program Leg 209 recovered drill cores of serpentinized peridotite and hydrothermally altered troctolitic and gabbroic rocks from various locations on rift valley walls along magma-starved ridge segments adjacent to the 15$^o$20N Fracture Zone. We have examined carbonate veins to gain insights into chemical evolution paths of seawater circulating within the lithospheric mantle as it is exhumed and uplifted. Three types of carbonate veins were distinguished in cores from Sites 1270-1275 on the basis of their chemical and isotopic compositions: (1) aragonite veins in serpentinite (T=2-15$^o$C, Sr/Ca=6.6-14.1, $\delta^{13}$C=-2 to +3$\permil$), (2) calcite veins in troctolitic rocks (T=5-120$^o$C, Sr/Ca=0.12-0.79, $\delta^{13}$C=-4 to +3$\permil$, and (3) calcite veins in schistous metaultramafic rock (T=90-220$^o$C, Sr/Ca=0.07-0.14, $\delta^{13}$C=0 to +9 $\permil$). The high-T calcite veins formed within detachment faults, while the aragonite veins formed after exhumation and are related to cracking during uplift. These veins may hence provide insights into two different hydrological systems associated with exposure of ultramafic basement in inside corners of fracture zones. Calcite veins from talcous schists have chemical (low U/Ca, Mg/Ca, high K/Ca) and isotopic (low $^{87}$Sr/$^{86}$Sr and $\delta^{7}$Li, see Rosner et al., this session) signatures of high-temperature fluid-rock interactions. Their C and O isotope compositions preserve a record of thermogenic methane production at high temperatures ($>$300$^o$C), followed by cooling and possibly partial oxidation of methane as fluids migrated up along detachment faults. Aragonite veins are young (2 to 12 kyrs at Site 1274 and 38 to $>$55 kyrs at Site 1271) and reveal interesting co-variations between subseafloor depth and age, temperature, as well as chemical indicators of fluid maturity (e.g., Mg/Ca). These data can be used to estimate that (1) fluid flow rates are on the order of 1 cm/yr, (2) the modern thermal gradient in upper basement is about 100-150$^o$C/km, and (3) and 90-95$%$ of Mg are lost as the temperature of circulating seawater increases by 10-15$^o$C. Our results indicate that subseafloor low-temperature seawater-peridotite interactions constitute a Mg sink that can possibly counteract the Mg loss associated with brucite dissolution and olivine weathering at the seafloor. Our estimates suggest an annual loss of about 40 mol Mg per Watt convective heat loss in these settings.
V22A-02 INVITED 10:40h
Paleomagnetic constrains to the evolution of the mid-Atlantic Ridge from 14° to 16° N.
Mantle rocks outcrop extensively at the sea floor near the Mid-Atlantic Ridge between 14° and 16° N. Tectonic extension occurring at the ends of ridge segments, where igneous crustal accretion is unable to balance sea floor spreading, causes uplift and exhumation of lower crustal and upper mantle materials at the footwall of normal faults. Models of extension along normal faults dipping towards the rift valley require a certain amount of block tilting, in a predictable sense top away from the ridge. Paleomagnetic data from cores can provide information on the amount of tectonic block rotation, provided that orientation of the rotation axis can be constrained. Paleomagnetic results from several sites of Leg 209 yield mean inclinations which are significantly lower that expected for the geomagnetic axial dipole at the site. Assuming that the magnetic remanence has averaged secular variation and that there is negligible error due to magnetic anisotropy, these results imply that syn-extensional block rotations have affected the rocks after they were magnetised at temperatures below 580°C. At sites 1268 and 1270, where structural trends at the sea floor are consistently parallel to the rift, the rotation axis can be considered approximately horizontal and parallel to the spreading axis. For these sites, as much as 90° of tectonic rotation may be required to match observed and expected paleomagnetic inclinations of gabbroic samples. At site 1268, talc-altered serpentinized peridotites yield remanences which are significantly steeper than the gabbros. This suggests, first, a different magnetization timing between the gabbros and altered peridotites and, second, that important tectonic rotations occurred at site 1268 after cooling of the gabbros and prior to peridotite alteration.
V22A-03 INVITED 10:55h
An ocean floor prospecting: Implications from the petrological insights of the abyssal chromitites from ODP Leg 209, MAR 15 20 N FZ
ODP Leg 209 performed to drill mantle peridotites and associated gabbroic rocks (at 8 Sites, 19 holes) along the Mid- Atlantic Ridge (MAR) from 14 to 16 N, both sides of 15 20 N FZ. Several podiform chromitites were recovered at Site 1271, both Hole 1271A and B, south of 15 20 N FZ during the cruise. These chromitites are the first sampled from the Mid-Atlantic Ridge. The primary chromites in the chromitites have moderately high-Cr# (0.52 and 0.48 for Hole 1271A and 1272B, respectively) and chromian spinels in the surrounding chromitites have similar composition. It is considered that podiform chromitite is formed by two kind of melt (melt-mantle interaction, therefore there should be plenty volume of melt in the mantle beneath this area. Abundant of gabbro and dunite were recovered from this area suggest that main volume of melt body were consumed in the upper mantle by melt-wallrock interaction. On the other hand, chromite grains have thick rims of Cr-magnetite or completely replaced by magnetite without chromite core. Cr content elevation in the Cr-magnetite rim occurred with Fe-enrichment. These chemical modifications suggest that the chromitite from Site 1271 were metamorphosed at amphibolite facies because significant Al-missing from chromite cores is taken place above 550 degree Celsius.
V22A-04 INVITED 11:10h
Structure and Composition of Oceanic Crust: What do we know; What do we need to know?
The seismic structure of "normal" crust is broadly consistent with the ophiolite model. P- and S-wave velocities are consistent with a crust of "gabbroic" composition; basalts that comprise the uppermost crust have been widely sampled by dredging, drilling, and submersible surveys, and velocities in the lower crust (layer 3) are in good agreement with the measured properties of hydrothermally altered gabbros containing 5-15%\ alteration products (amphiboles and phyllosilicates). However, diabase dikes have been sampled in only two localities, and the dike section is not well resolved in the seismic structure, possibly because the properties of the dike rocks are controlled by cracks. The nature of the layer 2/3 transition is not well resolved, but lab velocities in gabbros agree with layer 3 velocities, while velocities in dike rocks are lower, indicating that the top of layer 3 is within or at the top of the gabbros. While the seismic structure of normal crust is most consistent with the ophiolite model, there are regions where sampling is dominated by partially serpentinized peridotite. In the vicinity of the 15° 20' Fracture Zone, the ratio of peridotite samples to gabbros recovered by drilling and diving is approximately 3:1. The seismic structure of the crust in this region differs from the structure of "normal" crust in that P-wave velocities increase gradually, with no indication of a transition from layer 2 to layer 3. This structure could indicate either a gradual increase of the gabbro component or a gradual reduction in the degree of serpentinization with depth; profiles of both P- and S-wave velocities are needed to test these hypotheses. Compositional models of the uppermost mantle have not been considered in detail. Low average Mg numbers of crustal rocks and glass suggest that melts rising within the mantle may begin to crystallize as they enter the thermal boundary layer. Thus gabbros may be sequestered in the upper mantle. A comparison of upper mantle P- and S-wave velocities with a Voigt-Reuss-Hill model of the properties of mixed peridotite and gabbro suggests that the uppermost mantle may contain 30 to 60%\ gabbro.
V22A-05 11:25h
Geochemical Study of Mid Atlantic Ridge Peridotites From 15\deg N: Preliminary Results From ODP Sites 1272 and 1274
During ODP Leg 209, eight sites were drilled along the Mid-Atlantic Ridge from 14\deg43 to 15\deg44 N, allowing recovery of ca. 354 meters of residual mantle peridotite intruded by gabbroic rocks (up to 25% of recovered samples). We present here the results of a whole rock major and trace element study of 28 peridotites selected among the less altered samples (those which preserved enough primary features to allow detailed petrographic and mineralogical studies), at Site 1272 and Site 1274 (respectively S and N of the 15\deg20 fault zone). Site 1272 and Site 1274 peridotites are mainly harzburgites (cpx $<$ 5%) with a few dunites. Harzburgites are characterized by highly variable opx contents (10-30 vol.%). Some samples show evidence of trapped melt (up to 3-5%). A few samples at Site 1274 show textural evidence of abundant cpx crystallization, resulting in the formation of lherzolites. All studied peridotites have been modified by alteration ($>$70% serpentinisation), leading to the addition of significant amounts of volatiles (loss on ignition $>$10 wt%). Nevertheless, except for alkali-earth elements and U, major and trace element compositions show no evidence of additional compositional changes during alteration. Site 1272 and Site 1274 peridotites are characterized by high Mg# (100 x molar Mg/[Mg + Fe]) (90.3- 92.2) and low Al$_{2}$O$_{3}$ contents ($<$0.3 wt% in dunites and ranging from 0.6 to 0.9% in harzburgites). Site 1272 and Site 1274 peridotites display flat to light Rare Earth Element (REE) depleted patterns (harzburgites : Ce$_{N}$$<$0.015; Yb$_{N}$$<$0.26 - dunites : Ce$_{N}$$<$0.001; Yb$_{N}$$<$0.04-0.08- $_{N}$: chondrite normalised). The most depleted harzburgites are observed at the southernmost Site 1272 (Ce$_{N}$$<$0.003; Yb$_{N}$: 0.045-0.1). REE allow to distinguish two harzburgites groups at Site 1274. The first one shows patterns similar to those of Site 1272 peridotites yet with slightly higher REE content (Ce$_{N}$: 0.0015-0.015; Yb$_{N}$: 0.14-0.26). The second displays on average lower heavy REE contents (Yb$_{N}$: 0.1-0.18) and more variable light REE contents (Ce$_{N}$: $<$0.002-0.05). These samples are found at the bottom of Hole 1274A where cpx display the highest Na- and Ti contents among the studied peridotite suite. Site 1272 and Site 1274 peridotites composition is similar to that of the most depleted peridotites sampled in ophiolites (e.g., Oman, Cuba) and in oceans (e.g., Izu-Bonin-Marianna). They represent the most refractory peridotites yet sampled at a slow-spreading ridge.
V22A-06 11:40h
Upper Mantle Geochemistry at Peridotites of Site 1274 (ODP Leg 209): Relation to Melt-Rock Reaction and Processes at the Base of the Lithosphere
ODP Leg 209, Site 1274, has penetrated 156 m of upper mantle rocks plus minor gabbro intrusives north of 15\deg20' N Fracture Zone at the Mid Atlantic Ridge. The core has an unusually high amount of dunite (21%) and contains an unusual suite of opx-depleted harzburgites and dunites between 70 and 90 m depth. Between 0 and 70 m depth, harzburgites are chemically highly depleted (Cr\# in spinel 37 to 52, TiO2 in cpx $<$ 0.07% with negative correlation to Cr\#). The chemistry of these harzburgites appears to preserve evidence for progressive melt extraction. A section of 3 m dunite occurring within this interval largely shares the depleted chemical signature, though TiO2, Na2O and REE in cpx are somewhat elevated relative to the local host rocks. In the interval between 70 and 90 m depth, Cr\# are 45-57 and cpx is enriched in TiO2 (0.09 to 0.24%) and Na2O (up to 1%). Microstructural evidence for melt-rock reaction (e.g. cpx rims on opx) is pervasive throughout the core. The pristine preservation of these microstructures argues for efficient freezing during a rapid uplift. The percolating melts appear derived relatively locally because chemical indicators for foreign melts are weak. In the interval between 70 and 90 m, melt infiltration from a more remote source is suggested by the presence of strong reactive tendencies (opx depleted peridotites and dunites) and a more enriched chemical signature in all rocks. Probably, the opx-depleted rocks formed when focused melt channels failed near the lithosphere-asthenosphere boundary. In the upper part of the core, the depleted chemical signature of the dunite is interpreted as derived by melt extraction from the host into the dunite at a late stage, perhaps triggered by a fracture in the overlying lithosphere. Since locally generated melts are only weakly reactive towards opx, the formation of the dunite itself (by dissolution of opx from harzburgite) cannot be explained by locally generated melt. Instead, the early passage of more reactive melts derived from larger depth seems likely. Chemical clues to the former passage of different melts in this dunite are the occurrence of an enriched cpx inclusion in spinel in dunite and preliminary data suggesting that the internal part of the dunite may record the passage of a melt with a more radiogenic Os signatures than the highly unradiogenic harzburgites and marginal dunites. It thus appears that melt infiltration and extraction can occur in close proximity, depending on the local physical properties of the lithosphere-asthenosphere boundary.
V22A-07 11:55h
Melt Extraction Versus Melt Stagnation Along the Ultraslow Spreading Lena Trough, Arctic Ocean
During cruise ARK XX-2 of RV Polarstern in summer 2004, 2400 kg of basement rocks were recovered by 23 dredge hauls along the oblique spreading Lena Trough. This 350 km long linear deep links the orthogonal spreading Gakkel Ridge in the north to the Knipovich Ridge in the south, and was until now barely explored. Lena Trough can be divided into three clearly defined segments: (1) a northern basalt-dominated segment; (2) a central 140 km long basalt-free, peridotitic segment; and (3) a southern basalt-bearing segment. Exclusive basalt recovery in the four northernmost axial Lena Trough dredges suggest that the change in spreading geometry from magmatically robust orthogonal spreading at the western Gakkel Ridge to oblique near-continental spreading at Lena Trough did not lead to an abrupt cessation in partial melting underneath northernmost Lena Trough. However, eruptive magmatism in northern Lena Trough may not have been active for a long time, since an off-axis dredge haul on the western flank at 82.8$\deg$ N contained only peridotites. Central Lena Trough is completely devoid of basalts, and is therefore the longest amagmatic spreading segment globally. Conventionally, the absence of basalt and exclusive exposure of mantle rocks directly translates to amagmatic extension. However, some partial melting may have occurred underneath central Lena Trough. Instead of being extracted from the mantle, melt stagnated and is now present as disseminated plagioclase and crosscutting gabbroic veins, as observed in 20-30% of all abyssal peridotites at elsewhere along the Gakkel Ridge, SWIR and CIR. The textures and mineral compositions of plag-peridotites can only be explained by melt entrapment. In addition, plag-peridotites are virtually absent in the southern- and northernmost dredge hauls and their abundance gradually increases to 100% in central Lena Trough. Melt generation occurred along the entire length of Lena Trough. In northermost Lena Trough, these melts were extracted from the mantle and formed a (thin) basaltic layer. In central Lena Trough, melts were formed in the mantle, but never managed to focus, pool and erupt. The reason for this may be an increase in lithospheric thickness from the Gakkel Ridge to Lena Trough intersection towards the south. In central Lena Trough, the lithosphere may be exceptionally thick resulting from the conductively cooling influences of the extreme obliquity, the ultraslow spreading and the vicinity to continental margin of Greenland and Svalbard.
V22A-08 12:10h
Are MORB and OIB produced by a hybrid flux-melting process instead of `pure' pressure-release melting?
Decompression melting at mid-ocean ridges is the simplest type of terrestrial melting. It is now widely accepted their major element chemistry results from the melting of a peridotitic mantle source (cf. books by Ringwood (1975); Yoder (1976)), yet their trace element isotope geochemistry is heavily influenced by mantle components that arise from recycling non-peridotitic sediments and basalts into the mantle at subduction zones (cf. Dicken (1995) `Isotope Geochemistry'). How can isotopically varying recycled components often dominate the isotope composition of a basalt and its melt inclusions yet have such a small effect on its major element chemistry? This was fairly surprising to the pioneer isotopologists in the early 1980s, as was the idea that the relatively refractory residues of peridotite melting could somehow, after being subducted back into the mantle, be refertilized to remelt again as fertile peridotite. A two-part mechanism of subsolidus stretching and mixing in the mantle (Allegre&Turcotte) followed by diffusive source homogenization during partial melting (Hofmann&Hart) was developed to explain this puzzle. However recent isotopic observations on melt inclusions in particular appear to refute the idea that diffusive homogenization can occur over a km or larger length-scale during typical MORB melting. If this hypothesis does not apply during mantle melting, then how can both isotope and major element constraints be satisfied? One potential path is to return to variants of the ideas of the existence of pervasive metasomatic `ichors', magmatic fluids that infused into, stagnated, and metasomatized a peridotite that much later would undergo pressure-release melting. However, pressures within the deep mantle prevent the formation of silicate melts - thus inhibiting such metasomatic activity. Also, the observation that the recycled enriched sediment and ocean crust components lie above the depleted mantle of a subducting slab that they must metasomatize makes it difficult for this scenario to work. Here we would like to explore a completely different mechanism that might be able to generate the observed major element chemistry of MORB and OIB - melt/wallrock interaction during the ascent of deeper melts generated by pressure-release melting of non-peridotitic compositions. In its essence, this mechanism would imply that typical peridotite melting beneath a hotspot or mid-ocean ridge is actually quite similar to the `flux melting' commonly believed to cause subduction zone volcanism, with the difference that the `flux-agent' at mid-ocean ridges is initially a rising silicate melt that was produced by pressure-release melting of a lower-solidus plum instead of initially being a rising hydrous fluid. If the mantle is a plum-pudding of different recycled components, then it is likely that during ascent small amounts of deeper decompression melting of volatile and incompatible element rich components will take place prior to the main phase of peridotite pressure-release melting. These incompatible and volatile elements will act as `fluxes' that tend to stabilize a melt instead of solid phase, inducing flux-melting of surrounding wallrock as they ascend. Here we will discuss a series of computational thermodynamic melting calculations that illustrate this hypothesis. This mode of mantle melting can reconcile major element and trace element isotope constraints. It is also consistent with the dunite banding in peridotites proposed by Kelemen and others to be the byproduct of basaltic melt-wallrock interactions and furthermore provides a nice `unification' of MORB, OIB, and arc melting -- all are (hybrid) flux-melts.