V23H-01 INVITED
On the Walvis Ridge Hotspot Life-Cycle: Flood Basalt to Robust Aseismic Ridge to Low- Volume Intra-Plate Volcanism
The origin, timing and duration of the emplacement of Large Igneous Provinces (LIPs) are under continuous debate. For example, their relation to hotspot volcanism and age progressive seamount trails is still an unresolved paradigm in Earth sciences, but has important implications for heat and mass transfer in the mantle, the time and length scales over which geochemically distinct mantle domains can exist, and mantle geodynamics in general. The Walvis Ridge hotspot track in the South Atlantic may very well be the best preserved example of continuous intra-plate volcanism, lasting ~130 Myr since its initiation at the once contiguous Parana-Etendeka continental flood basalt province on Gondwana. This volcanic system may record a rare example of the temporal and geochemical evolution of a 'primary mantle plume' that is consistent with the thermal plume model and that thus can provide important insights into its life-cycle consisting of three main stages: (i) the initial formation of the voluminous Parana-Etendeka LIP, (ii) followed by the robust age-progressive Walvis Ridge (and Rio Grande) aseismic ridge(s), and (iii) ending with a low-volume dispersed seamount trail (Guyot Province). The systematic decrease in the volume of volcanism may no doubt reflect the waning of a long-lived plume source. Recent modeling of thermochemical plumes offer mantle convection simulations depicting two major types of mantle plumes features, one consistent with the classic thermal plume model, in which a plume head rises vertically through the mantle, followed by a continuous, straight and singular plume tail. In the second type the plume head stagnates at mid-mantle levels and only secondary plumes (without plume heads) rise to the surface. Although recent seismic tomographic studies have been very successful at imaging plume conduits (tails), there is little convincing evidence for plume-head-type structures in the present mantle. Furthermore, neither of these mantle plume features have been corroborated with field evidence. Field observations of plume shape variability are limited due to the ambiguity that some hotspot tracks might have formed by local extension and fracturing of the oceanic crust rather than from plume volcanism. Moreover, it also is possible that variability in track morphology reflects interactions of a plume with the overlying oceanic lithosphere or deflection of plumes by mantle flow. In the case of the Walvis Ridge the existing sample database has a poor geographic coverage with an age data set below modern-day standards. A systematic sampling campaign, using both dredging and ocean drilling, is required to test and explain the presumed Walvis Ridge hotspot life-cycle.
V23H-02
The PGE Signature of OAE1a: No Evidence of a Bolide Trigger for OJP Volcanism
The large-volume eruption and the short-period of emplacement shown by the geochronological, geological, and geophysical studies of the Ontong Java Plateau (OJP) require some explanation beyond what is already known from plate tectonic processes. Plume impact, bolide impact, and their variants have been proposed to account for the origin of these features. The two end member models (bolide vs. plume) cannot be distinguished from one another based solely on the Os isotopic compositions of the OJP lavas because both meteorites and mantle have unradiogenic Os isotope signature. Additional evidence to test the models can be derived from the platinum group element (PGE) abundances in sedimentary rocks deposited before and during the LIP emplacement. We analyzed the platinum group element (PGE) contents of the Selli Level horizon at Gorgo a Cerbara, central Italy, which was previously analyzed for Os isotope composition, to further test the mantle plume vs. bolide impact models for the origin of the OJP and the Early Cretaceous oceanic anoxic event, OAE1a. We found that Ir concentrations are low compared to other known large impact horizons. Only one black shale bed yielded high Ir concentration of 544 ppt but it is not at the base of the Selli Level interval. In addition chondrite-normalized Os/Ir and Pt/Ir values of 1.4-21 and 7.7-31, respectively, within the Selli Level horizon are nonchondritic, including the bed that yielded the highest concentration of Re and other PGEs. Previous analysis of the Os isotopic compositions of the organic fractions of these sedimentary units showed a major interval of unradiogenic Os isotopic compositions coinciding with the Selli Level (Tejada et al., 2008). The absence of positive correlation between enrichment of PGE and the unradiogenic Os isotopic compositions suggests that the OAE1a event represented by the Selli Level horizon was triggered by a significant amount of mantle input into the marine environment and not by a bolide impact. Highly variable Os/Ir ratios that are uncorrelated with initial 187Os/188Os are very difficult to reconcile with the impact hypothesis. Our PGE data further corroborate our previous results that favor a mantle plume origin for the OJP and that the main phase of plateau volcanism triggered the OAE1a event. Reference: Tejada, M. L. G., Suzuki, K., Kuroda, J., Coccioni, R., Ohkouchi, N., Sakamoto, T., Tatsumi, Y., and Mahoney, J., 2008. Ontong Java Plateau eruption as a trigger for the Early Aptian oceanic anoxic event. Geology, submitted.
V23H-03 INVITED
Fluid mechanics constraints on LIPs origin and formation
Large Igneous Provinces represent the most voluminous igneous events on our planets and have probably ponctuated the Earth history since at least 3 Gyr. Recent fluid mechanics experiments on thermal convection show that the existence and characteristics of LIPs put tight constraints on mantle dynamics, structure and evolution. In particular, a relatively quick emplacement, hot temperature and large volume of LIPs would require a origin of LIPs deep in the lower mantle, a mantle rheology which depends strongly on temperature, and the presence of chemical heterogeneities. In the latter case, the laboratory experiments further suggest that LIPs could be heterogeneous in composition, could be followed by a variety of volcanic activities, and could even trigger subduction.
V23H-04 INVITED
Dating the Deccan: Determining the Timing and Duration of the Indian KTB-Spanning CFB Eruptions
Numerous studies have compared dates of mass extinction events with dates of flood basalt eruption and found significant correlation, thus suggesting a possible cause-and-effect connection. However, to explore the robustness of this relationship requires both detailed understanding of CFBP lava stratigraphies, and the application of high precision dating. Previous dating studies of the Deccan Traps, India, have concentrated upon classic sections exposed along the Western Ghats escarpment. These have provided insight regarding the timing and duration of the CFB, and indicate a geologically rapid eruption near to, and across the KTB. However, whilst these sections provide access to key parts of the Deccan succession, they do not provide a complete chronological record of the eruptions. The well-established stratigraphical architecture of the Deccan CFB has offered an excellent framework for a comprehensive sampling programme. Accordingly, we report argon ages for a suite of Deccan basalts within this well-defined and detailed stratigraphical context. Analyses were performed using the incremental laser heating technique on plagioclase separates, and all ages are reported relative to the GA1550 biotite standard. Data from the stratigraphically lowest (earliest), and highest (final) basalt lavas of the Deccan CFBP are presented, and provide a reliable determination of its timing and duration. Argon analyses of the basal lavas yield late Maastrictian ages (66.7 +/- 0.5 Ma), indicating the onset of volcanism in the north-western region c. 0.5 – 2 Ma before the KTB. Deccan eruption reached an acme of eruption during Chron 29R, with successively younger lava fields building on the southern flank of the evolving Deccan volcanic edifice. The final Deccan eruptions are thus preserved in the south-western Deccan and yield a weighted mean average of (63.7 +/- 0.3 Ma). These latter form the upper part of the widespread and volumetrically important Wai Subgroup which was erupted across the KTB, and which represents the major Deccan eruptive pulse. These bracketing age data demonstrate a duration of at least 2 – 3 Ma for the Deccan CFBP, and are significant because they help place constraints upon the efficacy of Deccan eruptions as a possible cause of late Cretaceous global environmental deterioration.
V23H-05
A Similar Multi-stage Geochemical Evolution for the Manihiki, Hikurangi and Ontong Java Plateaus?
Extensive sampling of the Hikurangi and Manihiki Plateaus, including seamounts on the plateaus, was carried out on the SONNE 168 and 193 cruises. Comparison of age and geochemical data from these plateaus with Ontong Java shows interesting similarities suggesting a similar temporal and geochemical evolution for each of the three plateaus. Each of the plateaus is characterized by a main tholeiitic plateau stage (c. 126-116 Ma) followed by later stage volcanism lasting more than 30 Ma on each of the plateaus. The late-stage volcanism on the Manihiki and Hikurangi Plateaus, including all seamounts, is predominantly alkalic. On Ontong Java the late-stage volcanism is largely tholeiitic, although alkalic rocks do occur, for example as dikes. Large variations in geochemistry occur on all three plateaus with tholeiitic and alkalic rocks generally having distinct compositions. The tholeiitic rocks range from FOZO (or E-MORB)-like compositions, characteristic of the major Kwaimbaita/Kroenke compositional group of lavas on Ontong Java, to EM1-type compositions, characteristic of the Singgalo endmember on Ontong Java. In contrast, the late-stage alkalic rocks range from FOZO- to HIMU-type compositions with 206Pb/204Pb isotope ratios exceeding 20.5. In summary, volcanism on all three plateaus was long-lived, lasting more than 30 Ma after plateau formation, and chemically heterogeneous, involving the same (at least three) components. Each plateau had a similar temporal and geochemical evolution, in particular Manihiki and Hikurangi plateaus which contain abundant late-stage alkalic seamounts. The similar evolution suggests that the plateaus were genetically related and may have formed in a single event and could have originally formed a single plateau. The similarity in composition of the late-stage alkalic volcanism, however, is more difficult to explain, since the plateaus were located thousands of kilometers apart, when this volcanism primarily occurred (c. 80-100 Ma ago). We propose that a second large-scale mantle event is responsible, possibly reflecting a second (FOZO-HIMU) plume head. This plume head may have stalled in the transition zone feeding lower levels of alkalic volcanism over an area of greater than 3000 km in diameter.
V23H-06
Tertiary Magmatism on the Early Cretaceous Ontong Java Plateau
The Ontong Java Plateau (OJP) in the western equatorial Pacific is the largest flood basalt province, or large igneous province (LIP), known in the oceans. Although the bulk of the OJP is believed to have formed in Early Cretaceous time, ca. 122 Ma, ca, 90 Ma basalts have also been recovered from the OJP and obducted OJP sections in the Solomon Islands. Still younger igneous rock is found in the Solomon Islands (Tejada et al., 1996), and the submarine plateau is itself surmounted by atolls, seamounts, and other features of presumed igneous origin, for which age data are lacking. To investigate this apparently younger igneous activity on the OJP itself, we have identified submarine lava flows and/or a volcaniclastic apron around Tauu Atoll on the southwestern OJP using seismic reflection data. Through stratigraphic correlation with Deep Sea Drilling Project and Ocean Drilling Program sites, we interpret the age of the igneous activity that created Tauu Atoll to be Middle Eocene to Miocene. Through similar seismic identification and stratigraphic correlation, we interpret an Oligocene to Miocene age for three hydrothermal vents in the central OJP. On the northwestern margin of the OJP, an unnamed seamount likely represents impingement of the much younger Caroline hotspot with the OJP. More seismic reflection and/or drilling data will be required to date Ontong Java Atoll, one of the largest atolls on the globe, as well as the many other atolls and seamounts surmounting the OJP.
V23H-07 INVITED
Heat and Helium in the Early Iceland Plume
The North Atlantic igneous province preserves a complete magmatic record from its inception at 61 Ma through to present activity in Iceland and provides a unique natural laboratory for the study of LIPs. The earliest magmatism was synchronous across a pre-drift area extending almost 2000 km from Baffin Island (BI), through West Greenland (WG), to Scotland. This start-up phase resulted in the eruption of voluminous picrite lava flows in BI and WG, and these provide information on the temperature and composition of the early Iceland plume. Major-element composition of BI and WG picrites suggests eruption- and mantle potential temperatures of at least 1400°C and 1500°C, respectively. These are comparable to estimates for the Ontong Java Plateau, and at least 100°C higher than those for mid-ocean ridge basalt (MORB). The picrites, however, are very similar to MORB in their 143Nd/144Nd and incompatible trace-element abundances and ratios. Normal (NMORB) and relatively enriched (EMORB) types are present in both areas. Crustal contamination is negligible in most of the picrites, and it is not possible to derive the EMORB type by contamination of NMORB-type magma with any plausible composition of continental crust. BI and WG picrites have the highest 3He/4He (up to 50Ra) yet measured in terrestrial basalts. These high values were previously thought to be restricted to the more depleted (NMORB) types but we now report data showing high 3He/4He also in EMORB types from both BI and WG. This observation appears to contradict models of He-isotope evolution in which primitive 3He is stored in ancient, highly depleted mantle domains. The early Iceland plume appears to have had a composition very similar to the mantle beneath mid-ocean ridges. It differed only in its high temperature and extreme 3He/4He, suggesting that primitive 3He was introduced from a hot reservoir with high 3He/(U+Th). If the Earth's core is the source of the heat and 3He, then the deep mantle must have a similar bulk composition and degree of heterogeneity to the upper mantle.
V23H-08
Mafic mantle sources indicated by the olivine-spinifex basalt-ferropicrite lavas in the accreted Permian oceanic LIP fragments and Miocene low-Ni basalt and adakite lavas in central Japan
Melting of mafic (eclogitic) rocks in the peridotite mantle diapir may be important to generate a large quantity of magma in a short period of time as required for the LIP basaltic magmatism (e.g. Takahashi et al. 1998; EPSL, 162, 63-). Ferropicritic rocks also occur in some LIPs, and Ichiyama et al. (2006; Lithos, 89, 47-) propose a non-peridotitic, Ti- and Fe-rich eclogitic source (recycled oceanic ferrogabbro?) entrained in the peridotitic LIP mantle plume for the origin of ferropicritic rocks, that occur with olivine-spinifex basalt (Ichiyama et al., 2007; Island Arc, 16, 493-) in a Permian LIP fragment that was captured in the Jurassic Tamba accretionary complex in central Japan. Although Ti-poor ferrokomatiitic magma might form through high- degree melting of a primitive chondritic mantle (25wt% MgO and 25wt% Fe+FeO), Ti- and HFSE-rich ferropicritic and meimechitic magmas can not form in this way. On the other hand, Miocene volcanic rocks distributed along the Japan Sea coast of central Japan also represent a product of large-scale arc magmatism that happened coeval to the spreading of the Japan Sea floor. The chemical and isotopic signatures of the magmas are consistent with the secular change of tectonic setting from continental arc (22- 20 Ma) to island arc (15-11 Ma) (Shuto et al. 2006; Lithos, 86, 1-). Some adakites have already been found from these Miocene volcanic rocks by Shutofs group, and mafic rock melting in either subducting slab or lower arc crust has been proposed. We have recently found a wide distribution of low-Ni basalt from Fukui City. The low-Ni basalt contains olivine phenocrysts which are one order of magnitude poorer in Ni (less than 0.02 wt% NiO at Fo87) than those in normal basalt (more than 0.2 wt% NiO at Fo87). The rock is also poor in bulk-rock Ni, rich in K and Ti, and may have formed from an olivine-free pyroxenitic source. Close association of adakite and low-Ni basalt with normal tholeiitic basalt, calc-alkaline andesite-dacite-rhyolite, high-Mg andesite and rare picritic basalt suggests melting of a heterogeneous mantle wedge that was abundantly endorsed with eclogitic and pyroxenitic rocks. Melting pressure greatly differs between the ferropicrite case (5 GPa or more) and the low-Ni basalt-adakite case (2 GPa or less), causing large chemical differences. However, common occurrences of non-peridotite-origin magmas in the LIP and island arc suggest pervasive and voluminous distribution of the mafic materials in the peridotitic mantle and their important role in magma genesis at various tectonic settings.