The ultramafic ridge of the Galicia margin, in the Ocean-Continent Transition zone, is made of mantle and magmatic rocks. The main objective of this work is to determine petrological and geochemical features of mantle rocks sampled in this area in order to study the subcontinental or suboceanic nature of the mantle and the different processes which have affected this part of the mantle before and during continental break-up.
Three main types of peridotites can be distinguished: 1) Spinel peridotites with Al-rich spinel (Cr*=0.02 to 0.2) and Na-rich clinopyroxene (Na2O=2.2 to 1.5%) representative of a rather "fertile" mantle. However, these CPX have high <epsilon>Nd values (+11.6 to +14.9) and LREE depleted patterns (La/Sm=0.01 to 0.33). These features are consistent with a long-term isolation of this part of the mantle from the convective asthenosphere (several hundreds million years), and thus corresponds probably to the subcontinental lithospheric mantle. 2) Spinel peridotites with Cr-rich spinel (Cr*=0.3 to 0.34) and Na-poor CPX (Na2O=1.5 to 1.3%). The clinopyroxenes have slightly depleted LREE patterns (La/Sm=0.4) but low <epsilon>Nd value (+4.4). These geochemical features involve a multi-stage history for these mantle rocks, and also support a subcontinental origin. 3) Spinel and plagioclase peridotites with pargasitic amphibole in epitaxic association with CPX. These rocks show important high-temperature deformation. Major element compositions of the CPX have been affected by the formation of plagioclase at the expense of spinel during decompression. Laser ablation ICP-MS analyses of amphiboles reveal very heterogeneous trace element compositions at the thin section scale and led us to consider a modal metasomatic event at the origin of these minerals. Furthermore, the similar <epsilon>Nd value (+5.9) of these amphiboles with magmatic amphibole in associated dykelets (<epsilon>Nd = +5.6, Charpentier et al., 1998) 122 My old (Féraud et al., 1988) suggests a recent event related to the North-Atlantic opening.In conclusion, the very heterogeneous petrological and geochemical features for the Galicia margin mantle rocks are not consistent with a suboceanic origin but are rather similar to what is found in many orogenic lherzolite massifs such as Beni-Bousera, Ronda or the Pyrénées massifs. The Galicia margin mantle will be discussed in terms of a part of the American-Iberian subcontinental lithospheric mantle rather than the first North-Atlantic oceanic lithosphere.
Charpentier S, Kornprobst J, Chazot G, Cornen G, Boillot G, CR Acad. Sci. Paris, 326, 757-762, (1998).
Féraud G, Girardeau J, Beslier MO, Boillot G, CR Acad. Sci. Paris, 307, 49-55, (1988).
Remnants of former Tethyan ocean-continent transitions (OCT) are preserved in the North Pennine Tasna nappe and in the South Pennine-Austroalpine Platta and Err nappes of the Eastern Alps. At these OCT, mantle rocks were exhumed along low-angle detachment faults and exposed at the sea floor like at the Iberia Abyssal Plain off Portugal (ODP Legs 149 and 173). In all these areas, no volcanic activity is observed before continental break-up.
Low-angle detachment faults in the three OCT are late, shallow crustal structures with a top-to-the-ocean sense of shear. They form a break-away in the continental crust and exhume uplifted mantle rocks. Tectono-sedimentary breccias and continent-derived blocks of variable size (extensional allochthons) overlie continental and exhumed mantle rocks along the detachment faults and are buried by post-rift sediments.
The exhumed mantle rocks in all three OCT are variably serpentinized harzburgites, lherzolites and pyroxenites and commonly preserve an old spinel foliation. Mylonitic shear zones associated with the exhumation of these rocks are rare and are usually affected by intense hydration. Near the sea floor calcite replaces highly strained serpentine minerals. Fractures and tectono-sedimentary breccias preserve cement fabrics and internal sediments documenting exposure at the sea floor. Mantle rocks and breccias are stratigraphically overlain by basalts and post-rift sediments.
Along the different OCT small gabbroic bodies within serpentinized peridotites are observed. Basaltic flows clearly postdate mantle exhumation at the sea floor. In the Platta nappe the volume of the basalts increases away from the OCT over 10 to 20 kilometers from zero to some hundreds of meters thickness, whereas the volume of gabbros does not change and remains small (< 5%). Thus, the first oceanic crust preserved in the Platta nappe does not show a classical ophiolite sequence, but an anomalously thin oceanic crust, similar to that observed along slow spreading ridges.
The similar architecture and tectonic evolution of the three OCT indicate that thinning of continental crust and exhumation of mantle rocks resulting in the formation of a so-called "transitional crust" is best explained with crustal-scale detachment faulting with a break-away in the continental crust. The transition to sea-floor spreading and magmatic activity seems, as indicated in the Platta nappe, to be gradual rather than restricted to a narrow, well defined zone.
The Gorringe Bank (SW Portugal) corresponds to an upper mantle peridotite ridge enclosing a 500 m thick - 50 km long laccolith-like body of gabbro. It also shows rare tholiitic dikes and pillow lavas resting locally directly over the peridotites. The whole is cut by late alkaline dikes. Strain analysis of oriented gabbros sampled during the Gorringe diving cruise demonstrates that the Gorringe complex underwent heterogeneous near-horizontal ductile shearing, preferentially developed along the gabbro-peridotite boundary, ie, at the brittle-ductile interface between upper mantle and crustal rocks. Shearing is responsible for mylonitic fabrics acquired in a simple shear regime, at decreasing temperature and pressure. These rocks alter experienced brittle deformation as expressed by steeply dipping conjugate normal faults related to pure shear. U-Pb datings on zircons were performed to obtain precise ages for the gabbros and late alkaline lavas and Hf isotopes measured to constrain their mantle sources characteristics 14 size fractions define ages of 140 and 139 (+/2) Ma for two gabbros which bear initial epsilon Hf (<epsilon>Hfi) values and + 21.4 and +19.2. A late alkaline gabbro, dated at 79 Ma, has an <epsilon>Hfi at 7.5. This shows that the basaltic magmas of Gorringe were generated from a MORB-type mantle whereas the akaline intrusions reflect magma extraction from an only slightly LILE depleted mantle domain. These new data confirm that formation of the Gorringe primary structures occurred during the first stages of oceanic spreading, at the end of the continental rifting. Its dynamics of formation likely reflects that of a slows-preading ridge.
The External Liguride Units of the Northern Apennines display thick successions of Upper Cretaceous age where large slide blocks of lower and upper continental crust are associated with Jurassic ophiolites. This association has been considered (Marroni et al., 1998) as representative of the continent-ocean transition in the southern margin of the Ligurian Tethys, being originally interposed between the westernmost oceanic realm now testified by the Internal Liguride Units and the easternmost thinned Adria continental margin (Tuscan Units).The ophiolites consist of (i) undepleted spinel-lherzolites, showing partial re-equilibration under plagioclase-facies conditions, which were interpreted as exhumed subcontinental mantle; (ii) rare gabbroic rocks crystallised from N-MORB magmas; (iii) basalts with N- to T-MORB affinity, locally covered by radiolarian cherts. The slide-blocks of lower continental crust include mafic and felsic granulites, locally preserving primary contacts. Mafic granulites derive from gabbroic protoliths, most likely crystallized from crustally-contaminated tholeiitic liquids (Montanini et al., 1998). The upper crustal rocks (mostly peraluminous two-mica leucogranites and biotite-bearing granodiorites) locally preserve primary contacts with ophiolitic rocks, being intruded by basaltic dykes or covered by basaltic flows and radiolarian cherts. Gabbro-derived and felsic granulites show a common retrograde metamorphic evolution, from granulite- to amphibolite-, greenschist- and subgreenschist-facies conditions. This evolution started at about 290 Ma, when the igneous protoliths of the mafic granulites were emplaced in the lower continental crust, whose remnants are represented by the felsic granulites. Retrogression to amphibolite-facies conditions was accompanied by development of ductile shear zones, which were overprinted, at lower temperature, by cataclastic deformations. The local formation of rodingitic veins in mafic granulites, presumably related to a serpentinization process in adjacent ultramafic rocks, suggests that lower crust was exhumed in association with the subcontinental mantle. The granitoids show evidence of polyphase brittle deformation under subgreenschist-facies conditions which predated the basalt emplacement. Available geochronological data indicate that the uplift of lower crust to shallow levels occurred in late Triassic-middle Jurassic times. In the same period, the granitoids underwent polyphase brittle deformation under subgreenschist-facies conditions, i.e. extensive brittle faulting at shallow crustal levels gave rise to extensional allochtons formed by stretched slices of granitoids. The ocean-continent transition represented by the External Liguride domain mainly consisted of ultramafic rocks of subcontinental mantle origin, probably covered by slices of lower and upper continental crust. In the oceanward portion of the transition, the ultramafic rocks were intruded by minor gabbroic bodies and basaltic dykes with MOR affinity. Basalts also occur at the top of the upper continental crust slices and were covered by Late Jurassic-Late Cretaceous pelagic deposits. Both mantle and continental crust rocks, as well as the MOR-type gabbros, were characterized by localized shear zones (represented by mylonites and cataclasites), developed at different times and structural levels as a consequence of the extensional processes leading to continental break-up. The petrological and structural data presented here fits well with the available data on present-day continental margins derived from passive lithosphere stretching.
Marroni M, Molli G, Montanini A & Tribuzio R, Tectonophysics, 292, 43-66, (1998).
Montanini A, Tribuzio R & Castorina F, Miner. Mag, 62A, 1017-1018, (1998).
(*) Shipboard Scientific Party: E. Boeuf, C. Buchanan, F. Chatin, J. Girardeau, G. Jacovetti, A. Moreau, M. Munschy, C. Partouche, S. Thomas, and U. Roberts.
During the Margau cruise (May 1998, R/V Marion-Dufresne), the transition between the thinned continental crust of the SW australian passive margin and the oceanic crust of the Australian-Antarctic Basin was explored between 108°E and 120°E by swath-bathymetry, imagery, magnetic and gravity data (Royer et al., 1999) and dredging of crystalline basement rocks.
The 11 dredge sites are located on basement structural highs in the three E-W oriented morpho-structural units observed along the margin, which are from north to south: (1) the continental slope, (2) a flat and sedimented 150 km-wide magnetic quiet zone, and (3) a rough 100 km-wide zone in the prolongation of the Diamantina zone. The varied and locally well-preserved lithologies which were dredged in these 3 zones are respectively: (1) continental high-grade sheared metamorphic rocks; (2) partly serpentinized spinel- and plagioclase-bearing peridotites; (3) a peridotite/gabbro/basalt assemblage which displays a retrograde tectono-metamorphic evolution, including a mylonitic shear deformation at the northern zone boundary.
Preliminary studies of the data suggest, by analogy with the comparable west Iberia margin, that (1) the continental breakup zone is located at the foot of the continental slope, (2) mantle rocks were tectonically exhumed in the breakup zone, (3) the flat and sedimented zone and the prolongation of the Diamantina zone represent a wide ocean-continent transition (OCT), which formation occurred after the continental breakup and before the beginning of oceanic accretion and was controled by amagmatic extension and/or ultraslow spreading. The granulitic continental rocks may represent either old granulites of the Gondwana craton remobilized on the margin during rifting, or hopefully are part of the syn-rift lower continental crust, usually burried under a thick sedimentary cover in rifted areas.
Implications for the models of formation of passive margins from continental rifting to incipient oceanization will be discussed.
Royer J-Y, Beslier M-O, Hill PJ, and Margau Shipboard Scientific Party, J. Conf. Abs., 4, (1999).
Small mantle diapirs structurally defined and mapped in the southern part of the Oman ophiolite are associated with a thick dunite transition zone and copious wehrlitic intrusions in the overlying crust. The criterion of special abundance of dunites and wehrlites has been used as a first indicator of focalised melt delivery to the crust related to mantle diapirism. A second indicator of melt activity is the thickness of the dunite transition zone which has been shown to sharply decrease away from mapped diapirs (Jousselin et al., 1998). Both indicators converge but the second allows a finer definition of the presumed upwelling areas. Throughout this 500 km long ophiolite, three broad areas may represent the loci of mantle upwelling below the paleo-ridge of origin. A first area, 60 km long, is located in the United Arab Emirates part of the ophiolite ; another, 30 km long, is located in northern Fizh massif. The last one coincides with a 120 km long propagator which opened into a very young lithosphere in the central and southern part of the ophiolite (Nicolas and Boudier,1995). In contrast with the situation of the two northern areas, where the ophiolite represents a piece of lithosphere which has been drifted away from the oceanic ridge of origin, the spreading axis of this propagator and thus the underlying mantle diapirs have been incorporated in the ophiolite during its oceanic detachment. Several individual diapirs, 10-20 km across, which have been mapped within the propagating segment only represent distinct heads at Moho level of a much larger mantle upwelling. Using the criterion of transition zone thickness three diapirs are now defined within the whole segment. They are 20-60 km across and encompass from one to three small structural diapirs. In a preceding study based on 32 cross sections through the crust, it was shown that the gabbro section of the crust was significantly thinner above the small structural diapirs than away (Nicolas et al, 1996). We extend this study with 18 new and more complete sections and find the same result for the wider upwelling domains now identified; the gabbro section above these domains is on average 2.5 km thick versus 3.7 km thick away from the upwelling areas. The sheeted dike section may also be thinner above upwellings, being on average 1 km thick versus 1.2 km away from these areas. Adding to these figures the 0.6 km estimation for the Geotimes or V1 lavas directly associated with the ophiolite the total thickness of the crust is only 4.1 km above the presumed mantle upwelling and 5.5 km away from these areas.
Jousselin D, Nicolas A &Boudier F, JGR, 103, 18153-18170, (1998).
Nicolas A & Boudier F, JGR, 100, 6179-6197, (1995).
Nicolas A, Boudier F & Ildefonse B, JGR, 101, 17941-17950, (1996).
The Oman ophiolite is believed to represent a fragment of oceanic lithosphere formed at a fast spreading rate. We are making detailed field-based, structural and geochemical studies of gabbros from the lower crustal section in order to constrain the processes of magma migration and storage. We have identified a fossil melt lens at the top of the plutonic section: a thin horizon containing gabbros of basaltic composition, which overlies a much larger pile of gabbros that represents a 'cumulate' crystal residue. Whole-rock geochemical studies of these lower gabbros show that they normally contain extremely low abundances of incompatible trace elements, and hence that little interstitial liquid was trapped in any of these 'cumulate' gabbros. This is true even directly below the melt lens, implying efficient melt extraction from the crystal mush, and a relatively solid floor to the melt lens, compatible with the deductions from the East Pacific Rise seismic experiments for some parts of the ridge axis.
In rare instances we find that equivalent cumulate gabbros in some crustal sections in Oman contain much higher concentrations of incompatible elements - and, by implication, residual interstitial liquid - in the uppermost sub-melt lens gabbros. This suggests that the processes of magma delivery and/or storage varied along the axis and/or with time and that some regions were characterised by a more extensive melt-rich crystal mush zone, perhaps equivalent instead to the crystal-rich regions of the East Pacific Rise melt lens. We attempt to constrain the variability in magma plumbing beneath the Oman palaeo-ridge and assess the implications for eruption-replenishment cycles beneath fast-spreading ridges.
The gabbros of the Oman ophiolite crystallised from the magma chamber of a fast paleo-spreading ridge. We describe here their textures and fabrics within wiew to understand the crystallisation, magmatic deformation and liquid phase circulation. The gabbros textures are described at different depth through two cross-sections, from two distinct massifs. The plagioclase shape has been carefully analysed showing that the length of plagioclase mesured parallel to the magmatic foliation remains identical troughout the cross sections, although their width increases with depth. Statistical analysis of the alignment in the foliation of the (010) preferentially developed face has been carried out by considering the albite twins, showing a similar alignment whatever the depth. We compare these data to the plagioclase lattice preferred orientation. An important result is that the deformation acquired is sufficient to generate a strong fabric. The main difference between upper gabbros, crystallised during a time lapse shorter than lower gabbros, and lower gabbro is in the fraction of post deformation crystal growth; i.e. upper gabbros where it was small, the fine-grained post deformation mineral represent a former melt. This allows to estimate melt fraction and to show the wetting angles point to an anisotropic permeability of the chamber.
The Bela ophiolite complex of Pakistan contains a complete ophiolite-accretionary wedge-trench sequence emplaced onto the Indian continental margin during the northward drift of India-Seychelles over the active Réunion hotspot. A structurally higher ophiolite overlies an accretionary prism, which is thrust over a foreland basin. Sedimentary rocks in the accretionary wedge indicate Aptian-Albian pillow lavas, initially deep water conditions, and increasing influence from the continent until the Eocene. Hornblende crystals in gabbroic rocks associated with tectonic slivers of peridotite in the accretionary complex yielded concordant 39Ar/40Ar ages of 148.4 ± 1.6 and 130.3 ± 2.1 Ma. Red to whitish granitic to granodioritic rocks within the accretionary sequence yielded unprecise biotite and hornblende ages of ca. 650-800 Ma supporting Gansser's (1979) suggestion that these rocks represent Indian basement fragments. The ophiolite emplacement was predated and accompanied by Fe-tholeiitic and alkaline magmatism (Sarwar, 1992) related to the Réunion hotspot and continuous incorporation of these rocks into the accretionary wedge. A hornblende concentrate from an alkaline rock yielded an 39Ar/40Ar plateau age of 70.9 ± 0.7 Ma. 39Ar/40Ar dating also showed that the ophiolite formed around 70 Ma. Shear-sense determinations in peridotite mylonites in the ophiolite footwall and imbrication structures in the underlying accretionary wedge indicate a west-southwest emplacement. Intraoceanic subduction initiated between 70 and 65 Ma, obduction onto the Indian passive margin occurred during Réunion hotspot activity (Basu et al., 1993) and formation of the Deccan traps at ca. 66 Ma, and final thrusting onto the continental margin ended in the Eocene (ca. 50 Ma). The ophiolite emplacement occurred in response to the counterclockwise separation of Madagascar and India-Seychelles which caused shortening and consumption of oceanic lithosphere between the African-Arabian and the Indian-Seychelles plates. The Proterozoic basement found within the accretionary complex represent probably isolated continental fragments from the break-up of East and West Pangea.
Gansser A, Geodynamics of Pakistan (Farah A & DeJong KA, eds), Geological Survey of Pakistan Quetta, 193-214, (1979).
Sarwar G, Tectonophysics, 207, 359-381, (1992).
Basu AR, Renne PR, Dasgupta DK, Teichmann R & Poreda RJ, Science, 261, 902-906, (1993).
Petrogenetic studies of ophiolite complexes generally concentrate on the magmatic oceanic crust. Peridotites are neglected because they are altered and have extreme chemical compositions. The Tibetan ophiolites provide an unique opportunity to study mantle processes because they are fresh and well exposed. We have used ID ICP-MS and NdO techniques to measure REE and Nd isotopes in highly depleted peridotites. We present major, trace element and Nd isotope data of 15 harzburgites from Luobusa ophiolite massif, Tibet. The aim of this study is to obtain a better understanding of mantle processes that lead to the formation of oceanic crust. The Luobusa peridotite massif is situated in the Yarlung Zangbo Suture Zone and is assumed to have the same formation and emplacement ages as the Xigaze ophiolite (respectively 110 and 50 Ma). The massif is 43 km by 3-5 km. It has a mantle sequence mainly composed of diopside-bearing harzburgites with abundant dunite lenses and podiform chromitites that are residual after extraction of tholeitic (MORB type) magmas (Zhou et al., 1996). The major element composition of Luobusa peridotites is on average more depleted than alpine peridotites. CaO and MgO contents vary between 0.6 and 2 wt% and 42.1 and 45.1 wt% respectively. There is an excellent correlation between major element contents (e.g., Al2O3 or CaO) and HREE. Similar relationships have been reported elsewhere and are consistent with HREE depletion being associated with the loss of Al- and Ca-rich phases during melting. Luobusa peridotites have REE patterns with subparallel HREE and constant MREE depletion ((Gd/Yb)N ~ 0.4). LREE are significantly more variable and vary from LREE depleted, with minimum values of La at 0.004 x chondrite, to those with a marked inflection at Sm such that (Ce/Sm)N ratios are = 0.9 ("spoon-shaped" REE patterns). The 143Nd/144Nd ratios vary between 0.5128 and 0.5143 and Sm/Nd from 0.9 to 5.4. These data define an errorchron of ~ 90 Ma with an unrealistic initial ratio <epsilon>Nd -10. 143Nd/144Nd also correlates positively with (Ce/Sm)N. The degree of LREE depletion in the peridotites is controlled by stratigraphic position: shallow peridotites being characterized by the highest (Ce/Sm)N. Together with the "spoon-shaped" REE patterns, these data imply that either melt was trapped in the residua during the extraction process or that the residua have been subsequently LREE enriched by melt addition. These observations suggest that fluid infiltration played an important role in the formation of the Luobusa peridotites. The detailed REE and Sr/Nd isotope analysis of minerals are underway and will constrain: the nature and the provenance of the fluid involved; the succession of processes that occurred in the melting column and the geodynamic environment in which the massif was formed.
Zhou MF, Robinson PT, Malpas J & Li Z, Journal of Petrology, 37, 3-21, (1996).
Along the Romanche FZ, a 950 km-offset transform zone in the Equatorial Atlantic, mantle-derived peridotites have a different composition from that of other peridotites of the mid-ocean ridge systems. Some of the peridotites along the Romanche FZ have fertile composition, suggesting very low degrees of melting. Other peridotites are characterised by anomalous modal and mineral composition, due to the abundance of plagioclase, clinopyroxene, and Ti-enriched spinel. Such regional-scale mantle heterogeneity has been explained in terms of mantle rock impregnation by basaltic melts. This process is consistent with a relatively cold oceanic lithosphere in which melt is not able to segregate and join the overlying oceanic crust, but freezes and is trapped at depth. Textural and petrologic data of mantle peridotites sampled in the central and western parts of the Romanche FZ during the oceanographic expedition PRIMAR-96 carried out in 1996 with the Russian R/V Gelendzhik show that these peridotites are impregnated by basaltic melts at various extent, confirming the results obtained by previous work. However, with respect to previous sampling, these new data show that impregnated mantle peridotites were emplaced also in the western Romanche RTI, where only fertile plagioclase-free peridotites were described in the past. This finding suggests that magma percolation is more widespread than previously thought. Textural features of peridotites suggest that melt percolated the mantle rocks, with diverse melt/rock ratios, primarly in the ductile part of the lithosphere and less frequently in the brittle domain. Electron and ion microprobe analyses show that the primary mineralogy of peridotites has been modified where melt products are present as interstitial minerals or in mineral aggregates crystallized during the ascent of melt by diffusive porous flow in the ductile part of the lithosphere. Minor or no chemical interactions occurred in peridotites containing veins cross-cutting the mantle foliation, produced by melt migration within focused channels at relatively shallow parts of the brittle lithosphere. Chemical changes primarily involve enrichment in Ti and Cr in spinel and of Ti and Zr (as oxides and HFSE) and REE in residual Cpx. LREE of Cpx are less fractionated in more impregnated peridotites, and a rough direct correlation is inferred between the (Ce/Yb)N ratio and the degree (%) of mantle rock impregnation. In less impregnated peridotites, Opx has a REE pattern similar to that of its related Cpx. By contrast, in more impregnated peridotites, the Opx REE pattern is different for not showing the Cpx's steep-and-flat trend. Melt products in impregnated peridotites have chemical features consistent with tholeiitic melts but reflect different degrees of melt fractionation.
Mantle tectonites and ultrabasic to basic plutonic rocks from the liguro-piemontais Montgenèvre Ophiolitic Complex and from the Equatorial Mid-Atlantic Ridge Vema and Romanche fracture zones are considered in this comparative study. Both environments are assumed to be expressions of slow rate accretion systems. The above localities present peridotitic tectonites, and the members of a strongly differenciated cumulate gabbroic sequence, with dominant high-level type gabbros.
Microprobe analyses of the primary and secondary phases have been performed in order to better constrain the effective part of alpine metamorphic imprint and thus to allow a more valid comparison of the concerned Moho transition zone lithologies.
The plagioclases of gabbros from Vema plot in labradorite field, they are commonly slightly to strongly deformed and thought to have been pervasively affected by hydrothermal fluids. Therefore they are not representative of primary compositions. Associated neoblastic secondary plagioclases have the same compositions. In ophiolitic equivalent lithologies, compositions are strongly albitic with again no significant variation between primary altered crystals and their associated neoblasts, suggesting in this case a more drastic pervasive albitisation. The style of deformations in these rocks does not seem to be different from the oceanic one but this observation needs to be improved by a U-stage study of intracrystalline deformations. In addition, fluid inclusions study should be attempted to better distinguish between magmatic, hydrothermal and metamorphic processes.
The diverse varieties of amphiboles have been extensively investigated by microprobe analyses. Either in oceanic or ophiolitic gabbros the distinction between primary and secondary amphiboles remains difficult. Nevertheless, in ferrogabbro types, the frequently observed brownish to dark brown amphiboles are considered to be primary, magmatic or late-magmatic; they are characterized by an edenitic composition, close to pargasitic, along a trend between pargasite and tremolite in the Na+K vs Si diagram. In the other gabbros, frequently of Ca-Mg types, colorless or greenish calcic amphiboles are mostly observed; they plot in the ferro-actinolite to actinolite (rarely tremolitic) and ferro- to magnesio-hornblende fields (Leake et al., 1997). Most of the latter varieties are interpreted as secondary in origin, this being often clearly evidenced by textural intergrowth relationship with relics of clinopyroxene or earlier developed amphibole. These data do not provide significant discrimination features between oceanic and ophiolitic amphiboles.
The microprobe analyses on relics of primary pyroxenes, such as in recrystallised neoblasts indicate diopsidic clinopyroxene in both environments. All these results plot very close, pointing out no alpine imprint.
There is no good evidence from the microprobe data to discriminate an alpine metamorphic imprint except maybe the severe albitisation of ophiolitic plagioclases. But at this point of the study it is not yet quite clear that this latest phenomenon is strictly alpine.
Leake BE, et al, American Mineralogist, 82, 1019-1037, (1997).
Gabbros associated with rifted continental margins may document magmatic underplating of continental crust, passively exhumed during much later rifting or, alternatively, initiation of partial melting during rifting and incipient sea-floor spreading. Both types appear to occur along present-day ocean continent transitions (OCT), e.g. along the Iberian margin off Portugal. In the eastern Alps of Switzerland the two types can be distinguished based on age, field relations and thermobarometric data.
In the Alps, OCTs bordering the southern Piemont segment of the Alpine Tethys (Platta nappe) and the northern Valais ocean (Tasna nappe of the Engadine window are partly preserved. Gabbroic intrusions occur along both OCTs, however, their ages, mode of emplacement and geodynamic significance are highly different.
In the Tasna OCT, continental basement rocks are separated from serpentinized lherzolites by a low-angle fault. Both continental and mantle rocks are stratigraphically overlain by Lower Cretaceous black shales sealing a previously exhumed low-angle normal fault that exposed serpentinized peridotites at the sea-floor during Mesozoic rifting. Lenses of mylonitic clinopyroxene- and amphibole-gabbro occur along the tectonic contact between the serpentinites and the continental basement and between the exhumed serpentinites and the overlying Cretaceous shales. Thermobarometric data from an amphibole-gabbro show that deformation occurred under decreasing pressure and temperature. U-Pb zircon ages of 254 Ma have been determined on a mylonitized gabbro (Froitzheim and Rubatto, pers. comm.) as well as on crosscutting undeformed tonalite dikes, interpreted as intrusion ages. Emplacement and deformation of this gabbro is therefore related to Permian extension and magmatic underplating and not to mantle exhumation during Mesozoic rifting, as previously assumed.
In the Platta nappe, gabbroic intrusions associated with serpentinized mantle rocks are intruded by basaltic dikes and occur as clasts in pillow breccias within a matrix of radiolarite of late Middle to Late Jurassic age. One particular body, 200 m long and 20 m across is composed of Mg-gabbro, Fe-gabbro and Fe-Ti-gabbro, and is cut by dioritic pegmatoid veins. Field evidence and mineral chemistry of the clinopyroxene suggest that this gabbros originated from the differentiation of the same parental magma. A diffuse high-temperature foliation indicates that the gabbros intruded into a tectonically active setting. Single zircon U-Pb dating of the Fe-Ti-gabbro and of one dioritic vein yield an age of 163 Ma. Age, crystallization sequence and thermobarometric data suggest that the gabbro intruded at a shallow crustal level after break up of the continental crust and may document the initiation of sea-floor spreading.
Located at the French-Italian border, the Pelvas d'Abries massif (2929 m) is an ophiolitic mass of the Piemont-Ligurian Schistes Lustrés zone. The unit is represented by a klippe of cpx-gabbros, minor Fe-Ti gabbros, and peculiar ultrabasic intercalations concentrated in the southeast part of the massif. Their relationship to the gabbros are still unclear. Ultramafic cumulate layers, ultrabasic intrusions, or peridotitic tectonic slices are possibilities. Contact relations between ultramafic rocks and gabbros are variable. The largest intercalations display sharp contacts. In the lower part of the eastern flank of the massif, the ultramafite appears to intrude the gabbro resulting in a magmatic melange: fragments of gabbro are enclosed in an ultramafic host, ultramafic apophyses cut gabbro, and evidence of assimilation of gabbroic enclaves is observed. The contact zones between the two lithologies may also appear to have been tectonised (shear zone, cataclasites, mylonites). Field and microscopic observations demonstrate that the high level-type gabbros are sometimes affected by oceanic flaserisation. Incipient magmatic layering is rarely observed. Mylonite zones occasionnaly crosscut the gabbroic mass. The north-northwest contact zone with the country rocks exhibits spectacular evidence of intensive tectonisation shown by slikenslides, mylonitised gabbro and the developement of a pervasive recrystallisation foliation. The primary non deform typical cumulate textures of the ultramafic rock contrast with those of the gabbros. The olivine is the dominant cumulus phase, with rare or absent orthopyroxene. Intercumulus plagioclase, often including chromiferous idiomorphic to subiodiomorphic spinels, crystallised as minor to large impregnation constituent. Except spinel, the primary minerals are completely serpentinized and chloritized. This strong textural contrast between gabbros and ultramafites is consistent with the intrusive character of the latter. The injection of the ultramafic magma seems to postdate oceanic deformation of the gabbros. Moreover, part of the peridotite appears to have intruded gabbro at an early magmatic stage (magmatic melange), whereas others were emplaced in a colder environment which explains the brittle behavior of the gabbros (sharp contacts). The Alpine metamorphism affected in different ways these two lithologies. The gabbros show typical recrystrallisation to blueschist or greenschist facies assemblages superimposed on oceanic imprints. The ultramafics, due to their chemical composition not have recorded any significant imprint of alpine metamorphism and have preserved their oceanic alteration (serpentinisation, chloritisation).
There are several tectonic sheets of ophiolites in accretionary wedge of Late Jurassic-Early Cretaceous Uda -Murgal island arc. All ophiolites are heavily broken and reduced to serpentinite melange. The Main Melange Zone consists of peridotites, amphibolites, garnet amphibolites, greenshists,island arc volcanic and sedimentary rocks, and oceanic basalts and cherts. Two type of mantle peridotites (oceanic and suprasubduction) are combined in melange. South Melange Zone consists of harzburgite, gabbro, and sheeted dyke blocks. Geochemistry and mineral composition of harzburgites correspond to highly depleted peridotites from suprasubduction geodynamic setting. Both melanges are located between Triassic-Early Cretaceous oceanic basalt-chert assemblages (to south) and Jurassic-Early Cretaceous ensimatic island arc complexes (to north). Supported by INTAS (grant 96-1880).
A diploma thesis at NTNU, involving mapping as well as structural, geochemical and petrological studies in Bymarka, Trondheim in the Central Norwegian Caledonides, has revealed an Early Ordovician dismembered ophiolite fragment with occurrences of oceanic plagiogranite intruding greenstones.
The greenstones consist of sheeted dikes, deformed pillow lavas and a few minor, coarse-grained greenstones believed to be gabbroic intrusions, representing oceanic layer 2, with layer 1 and 3 missing. Geochemically the greenstones show similarities to MORB, and regional considerations suggest a marginal basin setting. Petrographically the greenstones are dominated by chlorite, epidote, amphibole and biotite, and the geochemistry shows Na-enrichment, probably due to sea water interaction.
The plagiogranites occur as one large, coarse-grained intrusive body (one of Goldscmidt's original trondhjemites), several hundred meters thick, and as thinner fine- to medium-grained dikes, ranging from a few centimeters to several tens of meters in thickness. A chilled margin at the greenstone contact is accompanied by a narrow, contact-metamorphic zone in the greenstone. The plagiogranites are dominated by quartz, plagioclase, albite, muscovite and epidote, classifying as quartz-trondhjemite. The muscovite and epidote probably formed during autometasomatism.
The plagiogranite body has geochemical features distinctly different from plagiogranites as defined by Coleman and Peterman (1975), with an average K2O content of 2 wt.% and enriched REE-pattern, being more similar to Red Sea granophyre. The chilled margin and thinner dikes, however, have geochemical features typical for plagiogranites, although the REE-pattern is still enriched. The investigation shows this to be a secondary feature, either due to hydrothermal leaching in the hot oceanic crust, or later greenschist-facies regional metamorphism. The geochemical conclusions should therefore be based on samples from interior parts of the larger intrusion.
The model for formation of the plagiogranite is as follows: Partial melting of upper mantle lherzolite caused blobs of basaltic melt to rise buoyantly through the asthenosphere and pool at relatively shallow depth below the ridge, creating a small magma chamber. Due to discontinuous volcanic activity (most likely in a slow spreading ridge), the magma chamber may not have received another blob of melt for a long time, if ever, and the magma evolved into trondhjemitic compositions through fractional crystallization. The magma then intruded overlying diabase dikes and pillow lava, maybe as a result of renewed volcanic activity, that had cooled sufficiently to develop a narrow contact aureole. Karson et al.'s (1987) investigations of cyclic tectonic and volcanic activity on spreading ridges support such a model.
Intrusion into relatively cold basaltic rocks probably inhibited strong hydrothermal alteration of the plagiogranite, explaining the geochemical characteristics. The investigation supports the role of sea water in the evolution of plagiogranite inferred by Coleman and Donato (1979).
Coleman RG & Donato MM, Trondhjemites, Dacites, and Related Rocks. Elsevier, (1979).
Coleman RG & Peterman ZE, J Geophys. Res, 80, 1099-1108, (1975).
Karson JA, Thompson G, Humphris SE, Edmond JM, Bryan WB, Brown JR, Winters AT, Pockalny RA, Casey JF, Campbell AC, Klinkhammer G, Palmer MR, Kinzler RJ & Sulanowska MM, Nature, 328, 681-685, (1987).
Dunite-clinopyroxenite units are typical of ophiolite crustal sections (transition zone) and non-ophiolitic differentiated plutons (early fractionates). Here we describe the petrogenesis and internal structure of a range of such units from the Urals chain. The differentiated plutons (DP) of the Uralian Platinum Belt show evidence of marked fractional crystallization that developed in magma chambers at different depths. The largest massif of this belt is Kytlym, Middle Urals. The dunite-clinopyroxenite complexes of this massif are not associated with any residual peridotites. They are early products of fractional crystallization and show an enrichment in incompatible elements with decreasing Mg#. Later fractionates from the same magma include: olivine gabbros, gabbro-norites, hornblende gabbros, and plagiogranites. The primary magma was especially enriched in alkalis and Ti, which suggests a relatively fertile source, probably located in the sub-lithospheric mantle.Compared to the rocks of the Uralian Platinum Belt, dunite-clinopyroxenite complexes in ophiolite sections characterized by dominantly harzburgite mantle peridotites (DH-type; e.g. Voykar massif, Polar Urals) represent the early products of similar, but poorly developed, fractional crystallization. Later fractionates from the same magma include: gabbro-norites and hornblende gabbro-diorites. The dunite-clinopyroxenite complexes in ophiolite sections characterized by dominantly lherzolite mantle peridotites (DL-type; e.g. Nurali and Kraka massifs, South Urals) show a very limited evidence for fractional crystallization. In the Nurali massif the olivine-clinopyroxene ultramafics form a 200-m layered section. The structural position of this section and the chemical data suggest: (i) small, replenished, magma chambers that crystallized as sequences of olivine-clinopyroxene layers above the contact with the mantle peridotites; (ii) gradual focusing of the magma source within the uppermost part of the mantle section, now represented by narrow dunite-harzburgite zones; (iii) late injections of most primitive magma portions which produced high-Mg olivine-enstatite layers at the top of the dunite-clinopyroxenite section. For both DH- and DL-type complexes less fertile magma sources than for the DP type are suggested. Further information can be gained about the petrogenesis and history of the dunite-clinopyroxenite complexes by considering differences in their deformation style and intensity. The DP units display the most complicated tectonic and magmatic evolution, compared with the other two types. Early fractionates underwent increasing deformations during crystallization, related to hot ductile flow. Subsequently, as the P-T conditions of the stress were reduced, the deformation graded into semi-brittle and brittle. In contrast, the DL-type dunite-clinopyroxenite complexes show relict magmatic textures, and poorly developed ductile deformation. Finally, in the DH-type dunite-clinopyroxenite complexes magmatism is more complicated, and both ductile and semi-brittle deformation are present.
We studied mantle peridotites dredged in the central and western parts of the Romanche FZ during the Italian-Russian oceanographic expedition PRIMAR-96 carried out in 1996 with the Russian R/V Gelendzhik. The studied rocks are serpentinized Sp-bearing harzburgites and lherzolites with coarse- to medium-grained porphyroclastic upper mantle texture. Relics of protogranular texture have been observed in some samples. All samples contain Pl, or its pseudomorphic products, occurring in different amounts and with different habits. The modal composition of the Pl-peridotites is characterized by high Pl contents ranging between 0.4 and 36.8%, with an average of 10.2% that is higher than in the average abyssal peridotite from the world ocean ridge system containing 0.5% Pl. Our peridotite samples rich in Pl have also high modal Cpx (up to 30.6%). We consider the Pl-peridotites as hybrid rocks derived by upper mantle tectonites that were impregnated by percolating melts with gabbroic composition. This evidence confirms the results obtained by previous works. However, new data show that impregnated mantle peridotites were emplaced also in the western Romanche RTI, were only fertile plagioclase-free peridotites were sampled in the past. The products of trapped melt observed in our samples are texturally and mineralogically diverse. We have recognized three main types of magmatic intrusion: type 1consists of isolated interstitial Pl and Pl-Cpx-impregnation lenses; type 2 is represented by Pl-Cpx-Opx-Ol aggregates with xenomorphic granular texture, and type 3 by Pl-Opx -Cpx veins. Pl-impregnation lenses of type 1 are mostly concordant with the Sp foliation and must therefore have crystallized during early stages of magma circulation, when peridotites were still being deformed by mantle flow or soon after. By contrast, undeformed veins of type 3 that crosscut the mantle foliation may be interpreted as cracks or fractures opened by elastic deformation of the surrounding rock, and then filled by minerals crystallized from the melt. In this case, the host peridotite was probably well beyond its solidus and mantle flow had already stopped. Electron and Ion microprobe analyses suggest that the percolating melts interacted with the Romanche peridotites during mantle upwelling (see Tartarotti et al., this volume) mostly modifying the chemical composition of Sp and Cpx. Equilibrium between mantle rocks and transient melts was likely attained when high melt-fractions migrated through the peridotite by diffuse porous flow that likely account for type 1 and type 2 impregnation products. Sp and Cpx compositions were modified in equilibrium with MORB-type melt. By contrast, melt focused in open fractures have induced into the host peridotite limited chemical reactions.
Chromite pods form in the upwelling mantle beneath mid-ocean ridges and deform predominantly by fracture in the outward flow. Stress required for fracturing is loaded in the chromite inclusions by ductile flow of the dunite matrix. Results from finite element modelling of a deforming system composed of a rigid elastoplastic (rate-independent) inclusion in a linear viscoelastic (rate-dependent) matrix illustrate how a fractured inclusion can tell us about the strain rate of the deforming system, if the rheologies of the two phases are well-constrained. Chromite strength is estimated from fracture studies of similar spinels, translated to Drucker-Prager plasticity. The matrix is represented by a linear estimate of dry olivine at 1200°C, and additionally by a power law rheology. A series of experiments differing only in constant imposed strain rates yield a minimum strain rate/background matrix stress capable of fracturing the inclusion. Unknowns of the matrix rheology include temperature and intergranular melt volume. A positive first-order test of this gauge is that the models predict high matrix stresses, and high stress microstructures (well-defined subgrains) are observed in olivine grains immediately adjacent to fractured chromite from Oman. To rigorously test this gauge, it must be able to simulate analogous physical experiments.
When the strain rate gauge is "calibrated", it will be applied directly to specific chromite pods in mantle sections of ophiolites, ideally to the Maqsad diapir, Sumail Massif, Oman. Building from microstructures to whole pod models, combined with larger scale fabric studies and kinematic analysis, we aim to estimate flow velocities of surrounding peridotite, to understand the mechanics of crust-mantle interactions and the dynamics of mantle flow beneath ridges.
In order to interpret shallow seismic refraction measurements carried out in serpentinized peridotites from the Oman ophiolite in two crossed directions (Warsi, 1998), complemented by Lab measurements on oriented specimens (Kern, 1998), we have studied the geometry of the fractures networks from the scale of the field, all these fractures are serpentinized, to that of minerals aggregate, where the fracturation is caracterised by serpentine veining and cracks network. The investigated rocks are harzburgites with a high-T plastic flow fabric. The serpentinization (50%) develops a network of chrysotile veins and lizardite mesh, including also magnesite and iron oxide trails. This is a common rate of serpentinization in peridotites from ophiolites and from the ocean floor. U-stage measurements of 3D organization of the serpentine veining at the scale of minerals show a remarkable geometrical correlation with serpentinized fractures measured at the scale of field exposures. The serpentine network exhibits two preferred orientations: one perpendicular to olivine mineral lineation, the second parallel to (010) olivine preferred orientation. The microcraks (U-stage measurements), cross-cutting the serpentine veining, exhibit two preferred orientations: one horizontal and a second in the NW-SE sector, possibly related with the recent stress field. It is concluded that (1) the serpentine network is homothetic from the scale of the aggregate to that of the field exposure, (2) the serpentine network is controled by the olivine crystallography: serpentine veins are related to the olivine crystallographic orientation in the peridotite and (3) the microcracks could be controled by post-ophiolite emplacement stress field. The P-waves velocities present low values, between 6.05 and 5.95 km/s for lab measurements, between 5.3 and 4.2 km/s for field measurements. Further experiments should be conducted in relation with the serpentine framework in order to access to the nominal seismic anisotropy of partially serpentinized peridotites.
Western Ecuador is made of magmatic oceanic terranes accreted in the Late Cretaceous (80 Ma) and Paleocene (58 Ma). They form the oceanic basement of coastal Ecuador and occur as slices caught in the Late Cretaceous and Paleocene sutures.
The oceanic rocks exposed in the sutures consist of paired assemblages. The assemblage to the East is composed of ultramafic-mafic cumulates, pillow basalts, and dolerites intruded by shallow gabbroic stocks. The western assemblage consists of massive flows of picrites and basalts interbedded with bedded tuffs.
Peridotites of the eastern assemblage contain cumulus olivine + cpx ± opx with local late interstitial plagioclase. They are markedly depleted in LREE and their REE concentrations are very low (1 time the chondritic abundances). Associated gabbros are layered and contain cpx ± opx, or are isotropic and contain amphibole. These rocks have higher REE levels (10 times chondritic abundances) and are less depleted in LREE. Relative to N-MORB, all the cumulate rocks are depleted in Ta, Nb, Zr and Hf. An internal Sm/Nd isochron for these rocks gave an age of 123 ±13 Ma. Dolerites are olivine-free and are composed of oxide grains and plagioclase laths enclosed in clinopyroxene. They have almost flat REE patterns with REE concentrations similar to those of the gabbros. Relative to N-MORB, these rocks are depleted in Zr and Hf but enriched in Nb and Ta. All these rocks have a large range of isotope compositions [+10 < <epsilon>Nd(T=123 Ma) < +4.5; 17.89 < 206Pb/204Pb < 18.58 and 15.21 < 207Pb/204Pb< 15.51], suggesting derivation from a heterogeneous mantle source, or mixing of N-MORB and OIB-type sources.
The picrites and Mg-rich basalts of the western assemblage contain clinopyroxene with or without pseudomorphs after olivine phenocrysts. The picrites are slightly depleted in LREE and their REE concentrations do not exceed 3 times chondritic abundances. In contrast, the Mg-rich basalts are LREE-enriched and have higher REE abundances (20 times the chondritic abundances). Their <epsilon>Nd(i) range range between +10.24 to + 8.20. Their 206Pb/204Pb [19.24 < 206Pb/204Pb < 19.59] ratios are higher than those of the rocks of the eastern assemblage and are similar to those of the Upper Cretaceous Duarte Complex Mg-rich basalts from Hispaniola. This indicates that they derive from an enriched source, the composition of which includes the HIMU component.
The eastern assemblage likely represent a crustal section of an Early Cretaceous (123 Ma) oceanic plateau formed by the activity of near-ridge or ridge-centered hot spot while the picrites and Mg-rich basalts could represent remnants of the younger Late Cretaceous (~90 Ma?) Caribbean plateau.
We present here a relatively simple Isotope Dilution ICP-MS method for accurate and precise measurement of geological samples containing between 1 and 10 ppb of Rare Earth Elements (REE). The driving force behind establishing this method is the desire to obtain large numbers of high quality REE data of mantle-derived peridotites. The REE concentrations of such rocks can be so low (especially Light REE) that concentrations in diluted solution for ICP-MS analysis are close to, or below, detection limit (~5 ppt). Therefore the precision and reproducibility of the data measured by traditional ICP-MS method are compromised in many samples (errors are typically of the order of 120%, 2sd). Although the results can give a general idea of the REE distribution pattern, they should not be used for accurate quantitative modeling of mantle melting processes.
We have developed a simple pre-concentration/matrix removal technique combining isotope dilution (ID) to negate the effects of variable column yields. Once dissolved and spiked, we use a one-column chromatography technique to separate REE from the rest of the sample matrix and pre-concentrate to levels that give sufficient signal for precise isotope ratio measurements. Total blank procedures are in the 10 pg range. Measurements are made on an ICP-MS using a direct injection or micro-concentric desolvating nebuliser depending on the LREE enrichment of the samples. Instrumental analytical time is on the order of 10 minutes compared to the 6 hrs that is commonly required to obtain similar data by thermal ionization mass spectrometry. The data are corrected off-line for mass-interferences and blanks/background and then used for simple ID calculation of concentrations. We have achieved reproducibilities of better than 5% (2sd) on mantle rocks with concentrations as low as 1 ppb.
This basic technique can be modified for REE analyses in single crystals or microfossils, e.g. conodonts. Moreover it is not restricted to REE and can be applied to other elements that are also important in characterizing mantle signature. In addition Sr, Nd and Pb isotopes can be measured on the same dissolution if an aliquot is taken prior to spiking. To illustrate the technique we report new precise REE data for USGS international standards (PCC1 and DTS1) and highly depleted Tibetan peridotites.
The southernmost massif of the Ordovician Bay of Islands Ophiolite (Newfoundland, Canada), the Lewis Hills massif, exhibits an unusual structure that contrasts with the classic layered structure of oceanic lithosphere as defined by the Penrose Conference (Geotimes 1972). The eastern part of the Lewis Hills massif consists of a paleohorizontal sequence of mantle harzburgites and dunites, the latter containing lenses of layered ultramafic to mafic rocks. This eastern sequence is magmatically welded against a near vertical high-temperature shear zone of the central part of the massif. The shear zone is about 4-5 km in width and comprises granulite- to amphibolitefacies metagabbros that were intruded syn- and posttectonically by a diversity of mafic to ultramafic rocks. The western part of the Lewis Hills is composed of mainly undeformed hornblende gabbros, including basaltic dikes and small intrusions of trondhjemites.
Recently, strikingly different isotope signatures (western part: <epsilon>Nd(t) = -1.5 to +2.0; eastern part: <epsilon>Nd(t) = +7.4) and an age difference of ~ 20 m.y. were discovered for both parts of the Lewis Hills massif (Kurth et al., in press). Further, the syn- to posttectonically intruded rocks of the shear zone span a wide range from N-MORB type, depleted to refractory geochemical characteristics when compared to N-MOR basalts (Elthon et al., 1986; Sassen et al., 1997). Also, negative Nb- and Ta-anomalies of basaltic intrusions support their island arc affinity.
In the light of the structural and isotope data the shear zone of the Lewis Hills represents a fracture zone contact between an island-arc type lithosphere and a marginal-basin type lithosphere (Cawood and Suhr, 1992; Kurth et al., in press). Within this tectonic configuration where the spreading ridge of the eastern part of the Lewis Hills massif abuts against the split island arc of the western part we suggest a new model for magma genesis along the arc - spreading ridge intersection.
Along the interface of the island-arc and marginal-basin lithosphere the uprising asthenosphere of the spreading ridge will be guided along the arc lithospheric wall. Because of the diverging motions of the lithospheric plates, components of the upwelling asthenosphere are expected to rise to the arc-ridge intersection. During upwelling, the hot asthenosphere will heat, and may incorporate, portions of the arc lithosphere. As a result, the asthenospheric portion becomes heterogeneous by the entrainment of variable amounts of arc lithosphere. Decompession melting of this heterogeneous source will create a diversity of magma types within a spatially narrow area as can be seen within the shear zone of the Lewis Hills massif.
Cawood PA & Suhr G, Tectonics, 11, 884-897, (1992).
Elthon D, Karson JA, Casey JF, Sullivan J & Siroky FX, Earth Planet Sci. Lett, 78, 89-103, (1986).
Penrose Field Conference, Geotimes, 17, 24-25, (1972).
Kurth M, Sassen A, Suhr G & Mezger K, Geology, (in press).
Sassen A, Suhr G, Seck, HA, Galer, SJG & Hofmann A, Terra Nova Abst. Suppl No1, 9, 333, (1997).
During the last stage of solidification of a magmatic cumulate, a certain amount of liquid becomes isolated from the main melt reservoir. This quantity of melt is called -trapped melt-. During further cooling such melts are forced to crystallize within the cumulate framework either as overgrowth on pre-existing minerals and/or as newly-formed intercumulus minerals. If the isolation from the melt reservoir is perfect, the trapped melt crystallizes in a closed system, which has some consequences regarding the final whole rock as well as the equilibrated mineral chemistry. The entire melt must become part of the final cumulate rock. Thus, the solid rock is especially enriched in those elements which are excluded from the original cumulus phases. The trapped melt has been well-documented in the case of continental intrusions (Barnes, 1986, Paster et al., 1974; Grant Cawthorn, 1996), and oceanic rocks including ophiolites (Meyer et al., 1989; Rampone et al., 1997, Ross & Elthon, 1997).
This work presents microprobe and Laser-Ablation-ICP-MS REE, Zr, Ti and Na data on minerals from a traverse through a mafic sill-intrusion occurring in the lowermost ultramafic crust of the Bay of Islands Ophiolite. Compositional trends in clinopyroxenes (cpx) and plagioclases are inconsistent with fractional crystallization. For example, (1) Ti and Na contents decrease with decreasing Mg-number, (2) and plagioclase Na contents in dunites are higher than those in associated gabbro, wehrlite and troctolite, and (3) dunitic and troctolitic cpx have higher La/Sm and Zr/Nd and lower Sm/Yb than gabbroic cpx. Quantification of the effects of trapped intercumulus liquid during solidification can explain all of these chemical effects, and yields reasonable amounts of trapped melt in equilibrated minerals (0.5% - 5%). The enrichment factor of an element in the minerals is strongly dependent on the rock mode and whether the rock matrix is buffered. Buffered rock matrices are those with a high bulk partition coefficient (D-bulk) for a given element (e.g. Ti in cpx-rich gabbros), and show only slight enrichment in constituent minerals compared to the original composition. Unbuffered rock matrices have a low D-bulk for incompatible elements (e.g. dunites), and exhibit highly enrichment in the most incompatible elements in their minerals.
The fractionation history and the parental melt composition of an intrusion or the plutonic section of ophiolites are often derived from mineral and whole rock compositions of cumulates. Because the final composition of a mineral/rock can be modified by postcumulus processes, such as trapped melt, the evaluation of this effects become important, especially when incompatible elements (e.g. REE) are used to reconstruct the parental melt of the cumulate.
Barnes S J, Contrib. Mineral. Petrol., 93, 524-531, (1986).
Paster T P, Schauwecker D S, Haskin L A, Geochi. Cosmochi. Acta, 38, 1549-1577, (1974).
Grant Cawthorn R, Contrib. Mineral. Petrol., 123, 109-115, (1996).
Meyer P S, Dick H J B, Thompson G, Contrib. Mineral. Petrol., 103, 44-63, (1989).
Rampone E, Piccardo G B, Vannucci R, Bottazzi P, Geochim. Cosmochim. Acta, 61, 4557-4569, (1997).
Ross K, Elthon D, Proc. ODP, Sci. Res, 153, 333-350, (1997).
The gabbroic rocks from Northern Apennine ophiolites were formed by low-pressure fractional crystallization of N-MORB type liquids, possibly in a slow-spreading ridge system. In the gabbroic rocks, accessory titanian pargasite occurs as the last crystallizing mineral, in interstices between plagioclase and clinopyroxene and as rim around interstitial Fe-Ti-oxides. In order to unravel the mechanisms leading to the crystallization of titanian pargasite, we have carried out a microanalytical study of amphiboles from Mg- to Fe-rich gabbroic rocks. Major elements have been determined by electron microprobe, and halogen and trace elements have been analyzed by ion microprobe. Further, a few separates of amphibole from Fe-rich rocks have been analyzed for Sr isotopic compositions.
The Mg# value of amphibole ranges from 0.78 to 0.70 and from 0.61 to 0.53 in Mg- and Fe-rich rocks, respectively. Amphibole has Ti varying from 0.3 to 0.5 apfu and relatively high F/Cl values. As a whole, Al and Ca contents in amphibole decrease with decreasing Mg#, whereas Mn increases. Calculations of structural formulae evidence for a significant oxy-component.
The REE pattern of amphibole is characterized by LREE depletion, nearly flat HREE and variable negative Eu anomaly. Total REE contents are highly variable. Normalization of incompatible trace elements to chondrite reveals that Ba, K and Sr are markedly depleted relative to LREE. Nb, Zr and Ti are slightly depleted to slightly enriched relative to neighbouring REE.
The incompatible trace element pattern is overall similar to that of poikilitic titanian pargasite from continental crust gabbroic rocks derived from contaminated tholeiitic liquids, only differing in a more marked LILE depletion. The initial Sr isotopic compositions of titanian pargasites from Fe-rich gabbroic rocks are consistent with those of fresh Mg-rich gabbroic rocks and fall within the range of present-day N-MORB.
Geothermometric calculations on the basis of amphibole-plagioclase equilibrium yield temperature estimates which are consistent with a development under igneous conditions. The igneous origin is further supported by the high REE contents in titanian pargasite, which cannot be ascribed to a metamorphic crystallization in the presence of seawater-derived fluids. It is argued that titanian pargasite derived from a magmatic liquid with relatively high H2O contents (5-8 wt.%) and a slight LREE enrichment.
The origin of titanian pargasite could be explained with the infiltration in the gabbroic cumulate pile of a highly-evolved (trondhjemite-type) liquid, which contained relatively high amounts of H2O. Mixing relations between the cumulate pile and such exotic liquid could produce an interstitial liquid, which evolved in-situ through fractional crystallization, thus finally giving rise to the titanian pargasite. In an alternative hypothesis, the interstitial liquid derived from the interaction of the cumulate pile with a percolating exsolved igneous fluid, which concentrated relatively high amounts of REE.
We present depth-images, produced by prestack depth migration, of Cretaceous oceanic crust in the eastern Central Atlantic. These images, from profiles that are close to flow-lines, show evidence for the scale and geometry of normal faulting in the oceanic crust formed at a half spreading rate of less than 10 mm/yr. They also document the changing style of faulting with position within a spreading segment. Whereas most of the data studied are located away from large fracture zones FZs), part of a profile runs along the inside corner of one. This line exhibits larger-scale basement topography, with basement highs rising up to 2 km above the intervening basins. In contrast, the rest of the data are characterized by basement topography less than 1 km in amplitude. The different topography of the IC crust is also associated with differences in the style of faulting. Away from major FZs, fault-related extension is at most 20% and may be as low as 7%, most faults are relatively steep (> 35 degrees), are associated with moderate basement relief (0.2-1 km relief) and may penetrate to deep crustal levels. These could be related to the lifting of the lithosphere out of the median valley to the flanking mountains. Also observed away from FZs are gently-dipping to subhorizontal upper crustal detachment faults, overlain by steep faults bounding small rotated blocks. These sub-horizontal detachments are not exposed as slip surfaces over significant portions of top basement. However, the more rugged IC crust is characterized by gently dipping (<25 degrees), deep crustal rooted detachment faults that project to the ridgeward flank of large dome-shaped basement highs (1-2 km relief) and seem to continue as the ridge-facing flank of these highs. These detachments are up to 20 km long and have a heave of c. 10 km, sufficient to have accommodated up to 50% extension and to have exhumed deep crustal and perhaps even mantle rocks, as found at oceanic core complexes. We suggest that detachment faulting is important both at the IC ends of spreading segments, where the detachments are exposed as oceanic core complexes, and within the center of spreading segments where the detachments are everywhere covered by a carapace of faulted and fractured upper plate extrusives. We speculate that both types of detachment may be formed by a rolling hinge model, but that the different exposure of the detachments may be related to the variation in the flexural response of the lithosphere along the spreading segment and the implications this has for the development of fault slices.
The Mirdita ophiolite in Albania occupies a N-S corridor which has escaped most Alpine and Cenozoic deformations, possibly due to a thick ophiolitic basement. The sheeted dike complex is NS and steeply dipping, indicating that the ridge of origin was oriented parallel to the NS corridor and that the ophiolite has not been largely tilted although differential motion between individual massifs cannot be excluded. Classically, the western lherzolitic massifs are opposed to the eastern "harzburgitic" massifs. Detailed structural mapping has revealed that the deeper mantle section was harzburgitic in both situations and that the major differences were restricted to the uppermost mantle and lower crust section. They are typically "ophiolitic" in the eastern massifs, being composed of a thick dunitic transition zone rich in basaltic impregnations and chromite deposits with above, a lower crust of layered gabbros. In contrast, in the western massifs, the uppermost mantle is composed of highly strained to mylonitic lherzolites which derive from more depleted harzburgites by impregnation and tectonic dispersion of melt during deformation occuring at 1000°-800°C. Layered gabbros are locally absent and the crust can be reduced to diabase dikes or sills and extrusives. The diabase intrusions are locally sheared together with the peridotites and transformed to amphibolites. The contrast between the eastern and western situations is ascribed to successive episodes of magmatic and amagmatic spreading in a slow spreading environment. The low-T, high strain deformation of the western massifs is localized in the dome-shaped envelope of these massifs. This structure, and even the present-day topography of these massifs, evoke the "turtleback" domes described along the MAR and explained by mantle denudation [Tucholke et al., 1998]
Tucholke BE, Lin J & Kleinrock MC, J. of Geophys. Res, 103, 9857-9866, (1988).
Recently, Cretaceous remnants of oceanic plateau have been characterized in the Caribbean and northern South Andes. This implies new interpretations in the geodynamic evolution of these areas and the key role of oceanic plateaus in initiation of subduction and/or continental growth. The Greater Antilles that extends from Cuba and Hispaniola up to Venezuela consists of an intra-oceanic Cretaceous arc which is inferred to have locally an oceanic plateau basement. The island of Cuba occupies the western segment of the Cretaceous island-arc. Its central part consists of: i) a late Jurassic ophiolitic belt; ii) a Cretaceous volcano-plutonic arc-assemblage; iii) an amphibolitic complex; and iv) the metamorphic Escambray complex. The amphibolitic complex is inferred to the basement of the Cretaceous arc which is composed of three magmatic suites. The first one consists of Lower Cretaceous intermediate to rhyolitic pyroclastic rocks with calc-alcaline affinity, overlain by Albian limestones. The second one is constituted by Upper Cretaceous tholeiitic basalts and andesites. 78 Ma granites form the third assemblage. Meta-ankaramites and meta-dolerites have been recognized in the amphibolitic complex. The ankaramites display abundant and large clinopyroxene phenocrysts while plagioclase forms microliths in the groundmass. Preserved clinopyroxene will be dated by the Pb/Pb method. The meta-dolerites consist of plagioclase laths cemented by clinopyroxene altered to green actinolitic hornblende. Ankaramites are characterized by an enrichment in LREE relative HREE, and are relative to Primitive Mantle enriched in Nb, Pb and Th. The Cuban Cretaceous arc is similar to that exposed in Hispaniola. Moreover, the meta-ankaramites are geochemically similar to those belonging to the Upper Cretaceous (86 Ma Ar/Ar age) Duarte Complex of Hispaniola. Thus, these data suggest that the basement of the Cuban Late Cretaceous arc is built on a basement which could be a fragment of an oceanic plateau.
Voluminous crustal growth took place in Central Asia during Neoproterozoic and Palaeozoic times, leading to the formation of the Central Asian Mobile Belt (CAMB). Sengör et al. (1993) postulated that from 5.3 million km2 continental crust about half of it was juvenile and directly extracted from the mantle by subduction-accretion processes. We present geochronological-, isotopic- and trace element data of the Agardagh ophiolite and the Tannuola island arc which are part of the CAMB. Our goal is to reveal Neo-proterozoic to Palaeozoic crust formation processes and regional mantle geochemistry.
The volcanic rocks of the ophiolite are enriched in incompatible trace elements (primitive mantle normalized (Ba/La)n~2.1, (La/Yb)n=4.5-9.7, (La/Sm)n~2) and absolute rare-earth element (REE) concentrations are higher than in bulk continental crust. Eu-anomalies are absent or slightly negative. In the majority of the volcanic rocks, Pb is enriched relative to N-MORB with (Ce/Pb)n~1.3 (compared to 2.7 for N-MORB), which may reflect the involvement of a crustal component during magma genesis or melt ascent. This is supported by negative Nb anomalies. The gabbroic rocks (cumulate and isotropic gabbros) of the ophiolite have distinct characteristics. Trace element abundances are low, the LREE are depleted relative to the HREE ((La/Yb)n=0.22-0.92) and Eu-anomalies are variable (Eu*=0.64-3). Calculated melts in equilibrium with the gabbroic rocks have significantly lower trace element concentrations than are observed in the basalts, and even melts calculated by fractional melting do not match the basalt compositions.
Single zircon 207Pb/206Pb dating yielded an age of 569 Ma for the ophiolite and a minimum life span for the island-arc between 512 and 447 Ma. Initial <epsilon>Nd-values for the ophiolite and the island-arc vary between 4.9 and 7.1 and thus are below the late Proterozoic depleted mantle value. Pb isotopic compositions of the volcanic rocks plot within the MORB field of Zindler & Hart (1986), the µ -value calculated for the source (8.2) is consistent either with crustal assimilation or a back-arc type setting.
Our results indicate that the ophiolite is an oceanic mantle/crust segment developed within a shallow marginal or back-arc basin. It is shown that a simple genetic link between the gabbroic and the volcanic rocks of the ophiolite does not exist. We propose that, after subduction-induced melting, the rising magma was fractionated within a magma chamber to form the gabbroic rocks. After gabbro formation, the rising magma was contaminated with overlying continental crust prior to eruption.
Sengör AMC, Natal'in BA & Burtman S, Nature, 364, 299-307, (1993).
Zindler A & Hart S, Ann. Rev. Earth Planet. Sci, 14, 493-571, (1986).
The late Proterozoic Fawakhir ophiolitic sequence, covering about 300 km2, is composed mainly of serpentinites, metagabbros, boninitic metagabbros and metavolcanics. The serpentinites are characterized by high Ni and Cr contents reflecting their development from a depleted-mantle peridotites. Major and trace element characteristics show that the ophiolitic metagabbro and metavolcanic rocks are of tholeiitic to calc-alkaline affinity. REE abundances in the metagabbros and metavolcanics are characterized by flat to fractionated patterns. The Fawakhir ophiolitic sequence shows a spectrum of compositions ranging from arc-type lavas, P-type MORB and boninitic rocks suggesting that these rock units developed in a back-arc environment. A three-stage rising diapir model has been proposed for the evolution of the Fawakhir mafic rocks. The first stage involves different melting proportions of a harzburgitic mantle source (about 20-35%) in a mantle melting column at a relatively deep depth subsequent by pyroxene, plagioclase and olivine fractionation to produce the gabbroic rocks. Pyroxene removal from the more depleted liquid, resulted from the melt extraction of the first stage, could generate the Fawakhir volcanics (Stage II). The boninites can be attributed to plagioclase and amphibole fractionation of a more refractory mantle source relative to the other rocks at shallow depth (Stage III).
Index of EUG 10 Volume
Further EUG 10 Information
Index of the Journal of Conference Abstracts
Cambridge Publications Home Page
Last Updated on Wednesday, March 17, 1999.
© 1997 Cambridge Publications