Laure Dosso

Geochemical morphology of the North Mid-Atlantic Ridge, 10°–24°N: Trace element-isotope complementarity

Abstract The new data presented here from a 10–24°N segment of the North Mid-Atlantic Ridge show that this segment is the most depleted of the 10–70°N ridge section. They also show the existence of: (1) a geochemical gradient from the 14°N anomaly to 17°10′N; (2) a very depleted mantle source (the lowest Sr isotopic ratios found so far in the North Atlantic); and (3) a geochemical limit located at about 17°10′N without any obvious relation with any structural feature. The 15°20′N fracture zone does not show any relationship with respect to this gradient. The basalts located north of 17°10′N have very homogeneous features, which allow their characteristics to be averaged (i.e., 87Sr/86Sr= 0.70238 ± 0.00004, (Nb/Zr)N = 0.28 ± 0.1) and they are defined as normal mid-ocean ridge basalts. The basaltic glasses located south of 17°10′N present a wide spectrum of isotopic compositions and extended rare earth element patterns (from depleted to enriched). Despite this, they have a constant K/Nb of 233 ± 9 (1sM, n = 18) whereas this ratio is 344 ± 29 north of 17°10′N. These observations illustrate the strong coherence of behaviour between K and Nb (Ta) during the petrogenic processes involved in the generation of these mid-ocean ridge basalts and also their fractionation during previous mantle processes. Possible interpretations of mixing processes are discussed and sources at the ridge segment scale are favoured. However, when looking in detail, local heterogeneities are still common and can even be traced back off-axis to 115 my. Placed in the context of the North Atlantic Ridge from 10° to 70°N, the Sr isotopic ratios reveal the Azores superstructure (23–50°N), whereas the trace element ratios (La/Sm-Nb/Zr) trace the second-order structures (33–40°N, 42–48°N) superimposed on the superstructure. This study illustrates the complementarity of information given by certain well chosen trace element ratios on the one hand and by isotopic ratios on the other. Since there is evidence of decoupling between isotopic ratios and/or trace element ratios, it introduces the notion of complementary “chemical memory” as recorded by a given type of trace element ratio or a given type of isotopic ratio

The age and distribution of mantle heterogeneity along the Mid-Atlantic Ridge (31–41°N)

New trace element and isotopic data for basalts from the mid-Atlantic ridge between 31 and 41oN allow a better description of the geochemical gradient south of the Azores triple junction, and the systematics of mantle source heterogeneity. There is a long wavelength enrichment in incompatible trace elements and isotopes associated with the Azores hot spot that extends from the Kurchatov fracture zone near 41oN to the Hayes fracture zone near 33oN. Superimposed on this gradient are local spikes of enrichment, the most prominent being the anomaly near the Oceanographer Fracture Zone (NOFZ). The Oceanographer anomaly spike is reflected modestly in the morphology of the ridge axis, but is not obviously related to a plume. The isotopic data alone are consistent with involvement of subcontinental material, but the samples do not contain the negative Nb‐Ta anomalies which are usually associated with the presence of continental material in the mantle source. Away from the prominent enrichment spikes associated with the Azores and Oceanographer fracture zone, there are systematic relationships in this region between parent=daughter element ratios and isotope ratios. The Pb, Sr and Nd isotope systems all give apparent ages in the range 100 Ma to 300 Ma, with the age increasing with likely parent=daughter fractionation during melting (U=Pb < Rb=Sr < Sm=Nd age). Monte Carlo simulations of an enrichment event in a depleted heterogeneous mantle at 250 Ma produce results that correspond well with the observations for all three isotopic systems. Since this age also corresponds to the pre-opening of the North Atlantic, it raises the possibility that some of the heterogeneity in this region is associated with shallow level mantle heterogeneity resulting from the rifting of Gondwanaland rather than from interaction with mantle plumes. The data may also reflect a mean mixing time for the heterogeneities in the upper mantle source. Sr isotope systematics reveal correlations in a 87 Sr= 86 Sr versus 87 Rb= 86 Sr plot, which are geographically controlled. Data points from 10‐24oN samples and data points from 31‐38oN samples (excluding NOFZ samples) plot on two offset trends of similar slope. Irrespective of the origin of the isotopic variations, these data require end member depleted mantle with distinct isotopic characteristics. Depleted sources with low 87 Rb= 86 Sr (0.005‐0.04) and low (La=Sm)N (<0.5), have 87 Sr= 86 Sr values that vary between 0.70215 and 0.7029. Therefore the depleted mantle source of N-MORB is not a homogeneous reservoir, but shows isotopic variations almost as large as the differences between generic depleted mantle (0.7025) and the enriched Atlantic plumes. Creation of a very heterogeneous depleted mantle in terms of isotopic composition needs to be included as a constraint on models of mantle mixing and convection.

The geochemical structure of the South-East Indian Ridge

Isotopic (Sr, Nd, Pb) and trace element data are reported for samples dredged along the South-East Indian Ridge (SEIR). They confirm the existence of a geochemical province under the Indian Ocean which is distinct from the provinces observed under the North Atlantic and Pacific Oceans. This province is characterized by the association of Sr and Nd isotopic enrichment features with a depleted Pb signature and of magmaphile element depletion features (low(La/Sm)N, (Nb/Zr)N, Rb/Sr with high Sm/Nd ratios). On a regional scale, the data emphasize the strong relationship existing between the physical structure of the ridge system and its chemical characteristics. It is suggested that these chemical characteristics are the result of contribution of different mantle sources corresponding to different structures of the ridge system: -“normal ridge” segments, -the ridge segment located on an immediately north of the St. Paul-Amsterdam massif, -off-axis structures located on the structural link between Broken Ridge and Kerguelen. In a model of ocean differentiation from Bulk Earth, it is suggested that SEIR mantle sources are the result of separate episodes of differentiation, all taking place more recently than the differentiation of the Atlantic and Pacific oceans.

SrNdPb geochemical morphology between 10° and 17°N on the Mid-Atlantic Ridge: A new MORB isotope signature

Basalts dredged along the Mid-Atlantic Ridge axis between 10°N and 17°N have been studied for their trace element characteristics [1]. To give complementary information on mantle source history and magma genesis, these samples have been analysed for their SrNdPb isotopic compositions. There is a good correlation between the structure of the ridge axis which shows a topographic anomaly centered around 14°N and hygromagmaphile element ratios such as Rb/Sr, (Nb/Zr)N or Sm/Nd as well as isotopic ratios plotted as a function of latitude. The samples coming from the 14°N topographic high show new MORB SrNd isotopic characteristics which pictured in a classical mantle array diagram, put their representative points close to HIMU sources of ocean islands such as St. Helena, Tubuaiand Mangaia. The 14°N mantle source presents geochemical characteristics which indicate mantle differentiation processes and a mantle history that are more distinct than so far envisaged from typical MORB data. Pb data indicates that the 14°N mantle source cannot be the result of binary mixing between a depleted mantle and a HIMU-type source. Rather, the enriched endmember could itself be a mixture of Walvis-like and HIMU-like materials. The geochimical observations presented favour the model of an incipient ridge-centered plume, in agreement with [1].

Arago Seamount: The missing hotspot found in the Austral Islands

The Austral archipelago, on the western side of the South Pacific superswell, is composed of several volcanic chains, corresponding to distinct events from 35 Ma to the present, and lies on oceanic crust created between 60 and 85 Ma. In 1982, Turner and Jarrard proposed that the two distinct volcanic stages found on Rurutu Island and dated as 12 Ma and 1 Ma could be due to two different hotspots, but no evidence of any recent aerial or submarine volcanic source has ever been found. In July 1999, expedition ZEPOLYF2 aboard the R/V L’Atalante conducted a geophysical survey of the northern part of the Austral volcanic archipelago. Thirty seamounts were mapped for the first time, including a very shallow one (,27 m below sea level), located at lat 23826.49S, long 150843.89W, ;120 km southeast of Rurutu. A nepheline-rich scoriaceous basalt sample from pillow lavas dredged on the newly mapped seamount’s western flank gave a K-Ar age of 230 6 0.004 ka obtained on pure selected nepheline. We propose that this seamount, already called Arago Seamount after a French Navy ship that discovered its summit in 1993, is the missing hotspot in the CookAustral history. This interpretation adds a new hotspot to the already complicated geologic history of this region. We suggest that several hotspots have been active simultaneously on a region of the seafloor that does not exceed 2000 km in diameter and that each of them had a short lifetime (,20 m.y.). These short-lived and closely spaced hotspots cannot be the result of discrete deep-mantle plumes and are likely due to more local upwelling in the upper mantle strongly influenced by weaknesses in the lithosphere.

Unusually large NbTa depletions in North Chile ridge basalts at 36°50′ to 38°56′S: major element, trace element, and isotopic data

Abstract We present new major and trace element data for 28 basalts and 10 basaltic glasses recovered from 16 locations from the North Chile Ridge (NCR) at 36°50′ to 38°56′S. This part of the Chile Ridge consists of three short ridge segments, which are characterized by deep (3200–4100 m) axial valleys. Chemical compositions of the basaltic glasses vary from primitive to moderately fractionated basalts (MgO = 9.54−7.28 wt%). All rocks are incompatible element depleted, mid-ocean ridge basalts (MORB) with average chondrite-normalized (La/Sm) N ratios of 0.46 ± 0.08. The radiogenic isotopic ratios of 10 representative samples display a narrow range in 87 Sr 86 Sr ratios from 0.70241 to 0.70249. 143 Nd 144 Nd ratios also vary within a small range from 0.51312 to 0.51318, and the 206 Pb 204 Pb ratios range from 18.2 to 18.6, with one exception which has a 206 Pb 204 Pb of 19.1. Overall, isotopic compositions are similar to average depleted MORB from the EPR (East Pacific Rise) and MAR (Mid-Atlantic Ridge), but 207 Pb 204 Pb ratios are significantly higher. Pb isotopic systematics of East Pacific MORB reflect large-scale heterogeneities, which are probably the result of long-lived differences in Th/U ratios in the mantle. Significant differences exist in the inferred primary melt compositions between the NCR basalts and depleted MORB from the South EPR and Galapagos Spreading Centre (GSC). For a given MgO content, basalt glasses from the NCR have systematically higher Na and Ti and lower Ca concentrations than those from the South EPR and the GSC. This has been interpreted as indicating relatively low average degrees of melting, which is possibly the result of cooling the shallow asthenosphere near transform offsets. NCR basalts are, on average, more primitive than basalts from the EPR and GSC, implying the lack of a robust magmatic system in this part of the Chile Ridge. This, together with the characteristic ridge topography and short average segment length, suggests that the magmatic system is short-lived, similar to slow-spreading ridges. This can be attributed to lower upwelling rates and, consequently, low magma supply rates near the transform offsets. The majority of the basalts from the NCR are unusually depleted in Nb and Ta relative to average depleted MORB. Trace element modelling shows that the distinct trace element characteristics of the NCR lavas could be the result of melting a mantle which has experienced a previous melting episode. This episode was possibly related to upwelling and melting of the mantle beneath the Pacific-Farallon ridge more than a million years ago.

Location of Louisville hotspot and origin of Hollister Ridge: geophysical constraints

Abstract The application of a new geometric technique [P. Wessel, L. Kroenke, A geometric technique for relocating hotspots and refining absolute plate motions, Nature 387 (1997) 365–369] recently pointed to a recent change in the Pacific plate absolute motion and suggested that the Louisville hotspot could now be located underneath the Hollister Ridge, south of the Eltanin fault system. However, the pole that was proposed for the last 3 Ma does not fit the trend of most Pacific volcanic alignments, supporting geochemical evidence [I. Vlastelic, L. Dosso, H. Guillou, L. Geli, H. Bougault, J. Etoubleau, J.-L. Joron, Geochemistry of the Hollister Ridge: relation with the Louisville hotspot and the Pacific–Antarctic Ridge, Earth Planet. Sci. Let. 160 (1998) 777–793] that does not favor a genetic relationship between the Louisville hotspot and the Hollister Ridge. We propose a pole near 57°N, 100°W that reconciles kinematic models with a previously proposed location [P. Lonsdale, Geography and history of the Louisville hotspot chain in the Southwest Pacific, J. Geophys. Res 93 (1988) 3078–3104] for the Louisville hotspot (near a Pleistocene volcano dredged at 50.5°S, 139.2°W) and claim that the Hollister Ridge most probably results from intraplate deformation processes.

The role of continental lithosphere metasomes in the production of HIMU-like magmatism on the northeast African and Arabian plates

Intraplate alkaline lavas typically exhibit isotopic characteristics that require a source with long-term isolation from the convecting asthenosphere, such as in the sub-continental lithosphere mantle or a mantle boundary layer. Melting of metasomatically enriched domains, or metasomes, within the lithospheric mantle provides a viable mechanism for generating the geochemical characteristics of intraplate alkaline basalts. The origins and distribution of these metasomes have been attributed to recent enrichment of the lithosphere by a mantle plume or ancient events that occurred during the early evolution of the sub-continental lithosphere mantle. Here, we present a geochemical study of Ethiopian Miocene intraplate alkaline lavas: melts of a lithospheric mantle that was enriched metasomatically during lithospheric stabilization and by recent plume-lithosphere interaction. We find that these lavas have geochemical characteristics consistent with melting of an amphibole-bearing lithospheric-mantle metasome. New Pb and Hf isotope data for these lavas require a HIMU-like source component, similar to other alkaline lavas erupted through the Horn of Africa, Sudan, and Egypt, and adjacent Arabian plate lithospheres. The isotopic characteristics of this component are distinct from the Afar plume mantle source and instead are consistent with the long-term evolution of a lithospheric metasome created during a Neoproterozoic subduction event associated with the Pan-African orogeny. The widespread distribution of easily fusible lithospheric metasomes within the continental lithosphere mantle may facilitate magma generation without the need for substantial lithospheric thinning or elevated mantle potential temperatures. Mantle heterogeneity of this nature has implications for the source origin of HIMU magmas associated with continental lithosphere.

Geochemistry of the Hollister Ridge: relation with the Louisville hotspot and the Pacific–Antarctic Ridge

The Hollister Ridge is located on the western flank of the Pacific–Antarctic Ridge (PAR), between the Udintsev fracture zone (FZ) and the Eltanin fault system. It is a linear aseismic structure, 450 km long, oblique with respect to the PAR. Data show that the most recent activity is located in the central part of the chain, which can be considered as being still volcanically active. Both major/trace element and isotopic data suggest that some interaction occurred between the Pacific–Antarctic Ridge and the Hollister Ridge. The source of the Hollister Ridge samples has its own geochemical characteristics. The geochemical variations observed along the ridge can be explained by mixing between two major end-member components: (1) a PAR depleted source, and (2) a Hollister enriched source. A small contribution (20% maximum) of Louisville plume material is likely to exist in the middle of Hollister Ridge. These data unequivocally reject the possibility that the Hollister Ridge could be the present location of the Louisville hotspot. Ages and geochemistry data support the idea of an influence of intraplate deformation as a probable cause of the origin of the Hollister Ridge.

New structural and geochemical observations from the Pacific‐Antarctic Ridge between 52°45′S and 41°15′S

We investigated the morphology and structure of the Pacific-Antarctic Ridge between 52°45′S and 41°15′S during the Pacantarctic2 cruise using multibeam echosounder together with gravity measurements and dredges. Analysis of the bathymetric, gravity and geochemical data reveal three ridge segments separated by overlapping spreading centers south of the Menard transform fault (MTF) and five segments north of it. Calculation of the cross-sectional area allows quantification of the variation in size of the axial bathymetric high. Together with the calculation of the mantle Bouguer anomaly, these data provide information about variations in the temperature of the underlying mantle or in crustal thickness. Areas with hotter mantle are found north and south of the MTF. Geochemical analyses of samples dredged during the survey show a correlation of high cross-sectional area values and negative mantle Bouguer anomalies in the middle of segments with relatively less depleted basalts.