V34B-01 16:00h
Mantle Plume Temperature Variations Immediately Following Continental Breakup of the Northern North Atlantic
The amount of melt generated by mantle decompression beneath an oceanic spreading centre and hence the oceanic crustal thickness is controlled in part by the temperature of the mantle. By measuring the thickness of the oceanic crust formed immediately after breakup of the northern North Atlantic we are able to deduce the maximum elevated mantle temperatures caused by the presence of the Iceland Mantle Plume. Crustal thickness variations are caused by temporal variations in the mantle plume temperature: at the present Reykjanes Ridge spreading centre the plume temperature pulses on a 3-5 Myr timescale with temperature variations of c.30K. We show results from the oceanic end of the Hatton-Rockall iSIMM (integrated Seismic Imaging and Modelling of Margins) project, where for this study, 45 OBS (Ocean Bottom Seismometers) were deployed across the oceanic crust of the Iceland Basin adjacent to the rifted margin. The OBS were deployed perpendicular to the margin extending 200km oceanwards from the continent-ocean boundary at Hatton bank and along a 90km strike line on the oceanic crust. OBS spacing was 10km across the oceanic crust with a denser spacing of 4km adjacent to the margin. The profile spans magnetic anomalies 18-24 (39-56 Ma) with the strike line parallel and coincident with magnetic anomaly 20 (44Ma). The seismic source was a 101.4 Ltr (6340 cu inch), low frequency, broadband airgun source designed to optimise seismic penetration at large offsets. Magnetics, bathymetry and gravity data were acquired as well as seismic reflection data using a 3km multichannel streamer (MCS). Processing of MCS reflection data has allowed accurate determination of the sediment velocity structure covering the igneous oceanic layers. Incorporation of sediment structure into a joint reflection and refraction tomographic inversion of the wide-angle OBS data has enabled us to map crustal thickness across the oceanic crust. Results span the first 17Myr of formation of the northern North Atlantic from breakup to mature seafloor spreading. With no apparent decrease in spreading rate thinning of the crust oceanwards suggests a temperature decrease of 70K from immediate post rift formation to normal oceanic crustal thickness 17Myr after breakup. Mantle plume temperature variations during this period would cause rapid changes in uplift of the north-west European margin and probably controls much of the Tertiary sedimentation patterns west of Britain.
http://www.badleys.co.uk/isimm-public/
V34B-02 16:15h
Laboratory Investigations of Plume Head Dispersion at a Segmented Spreading Ridge With a Sloping Rheological Boundary Layer
The dispersion of buoyant plume material in the upper mantle beneath a spreading ridge is examined using laboratory experiments. Specific goals of the experiments included determining the role of a cold rheological boundary layer (lithosphere/asthenosphere boundary) and the effects of a transform offset on along-axis flow of plume material. The rheological boundary layer (RBL) is established by cooling the surface of the fluid with dry ice. Plate driven flow is simulated by dragging mylar sheets in opposite directions across the surface of the fluid. The diverging plates force upwelling of deeper, ambient temperature fluid into the ridge axis, and a sloping RBL similar to lithospheric plates is created. A thermally buoyant plume is introduced into the plate driven flow, midway along a ridge axis segment. The plume signal is observed to spread along axis to a distance no greater than the diameter of the fully evolved plume head. Preferential along-axis channeling of the plume material is not observed. We suggest that thermal erosion of the RBL by the plume head creates an initially dome-shaped indentation in the base of the lithosphere. This pocket grows elongate in the axis-perpendicular directions through time with plate spreading. The volume of the growing pocket is maintained by the flux of plume material through the plume conduit. In cases with a plume located directly beneath a ridge-transform intersection, plume flow was deflected by the cold, highly viscous plane of the transform fault. The axis segment (near or far) to which the plume material is deflected may be spreading rate dependant. In all cases, the presence of a thermal RBL proved to be an important element in dispersion of relatively low viscosity plumes amidst a forced flow of higher viscosity ambient fluid.
V34B-03 16:30h
Plume-Ridge Interaction on the Cocos Plate (ODP Leg 205, Costa Rica): Implication for Fluid Circulation
At subduction zone, the permeability and hydrology of the down-going igneous section play a key role on the behaviour of the seismogenic zone, which produces earthquake and tsunamis. We present, here, evidence of fluid circulation pathways in the igneous section being subducted at the Central American convergent margin (Site 1253 ODP Leg 205, Costa Rica). The geological evolution of the Cocos plate over the last 25 Ma is complicated as a result of plate boundaries re-arrangement (Silver et al., 1998). The Cocos-Nazca spreading centre (CNS) interacted with the Galapagos hotspot, which simultaneously deposited volcanic material on both side of the CNS, on the Cocos and Nazca plates. The oceanic crust of the Cocos and Nazca plates was formed along four spreading centres (Meschede et al., 1998) and the East Pacific Rise (EPR). The EPR oceanic crust has the featureless morphology and low-amplitude magnetic anomalies common to fast spreading ridge (Wilson & Hey, 1995) while the CNS oceanic crust presents a rough topography and high-amplitude magnetic anomalies (Wilson & Hey, 1995). We wish to focus on the ODP 1253 Site drilled in the Cocos plate because it samples the igneous input (rock, heat and fluid) to the Central American subduction zone. Coring at Site 1253 penetrated two separate igneous Units of which the upper one is a sill (Unit 4A) separated from the lower Unit (Unit 4B) by sediment. Both Unit 4A and 4B present similar texture, structure and mineralogy apart from the thin basaltic interval at 513 mbsf, below which Unit 4B becomes more glass-rich and altered. Unit 4B is either a sill complex with multiple intrusions related to the Galapagos volcanic activity or a series of thick slowly cooled lava flows formed at the EPR. Some of these characteristics of Unit 4B are similar to those seen in horizons recovered at Site 1256, which intersected thick-ponded lava flows. Analyses of marine magnetic anomalies indicate that the crust at Site 1253 was formed at EPR 25 Ma ago. However, its thermal gradients and conductivity yield surface heat flow anomalies of 10-40 mW\/m$^{2}$, which is $\sim$70% lower than normal cooling lithosphere of that age (Fisher et al., 2003). Off-axis magmatic and tectonic activity associated with the formation of the Cocos Ridge may have increased fluid circulation pathways within the basement structure. Mobile trace element (Ba, Sr) and Sr isotope variations in conjunction with petrographic observation help identify horizons of fluid/rock interaction, e.g. fluid circulation pathways within Unit 4A and 4B. Based on veins, fractures, and void occurrences, the Unit 4A and 4B were subdivided into two and seven subunits respectively. Along the core-depth profile, mobile element concentrations and $^{87}$Sr/$^{86}$Sr ratios vary mainly in a narrow range (0.703245 and 0.703503) but are still higher than those of EPR or the Galapagos hotspot ($<$0.703). At sub-unit interface, we observe drastic increases in Ba and Sr content and more radiogenic Sr isotope signature (up to 0.705106), especially above the thin basaltic horizon. This reflects exchange of Sr with seawater or hydrothermal fluids during alteration. Further work will determine fluid flux estimate at Site 1253. The origin, nature and structure of the oceanic crust at Site 1253 will be presented in two additional abstracts (see Dreyer and Moe, this session).
V34B-04 16:45h
Plume-ridge Interaction Recorded in the Subducting Crust Offshore Costa Rica: Constraints From Major, Trace Element, and Isotopic Data From ODP Legs 170 and 205
Enriched gabbro sills ($\sim$30m) intruded into post-16Ma sediments and deeper medium-grained gabbros to cryptocrystalline basalts ($>$150m) were cored on ODP Legs 170 and 205 (9\deg39$\prime$N, 86\deg11$\prime$W) on the Cocos plate near Costa Rica. Magnetic anomalies indicate that the basement is EPR lithosphere (Barckhausen et al, 2001). The units studied here may be related to overprinting by the Galapagos hotspot, consistent with the site's paleolocation. The rocks are low- to medium-K (K$_{2}$0 $<$0.44wt$%$) subalkaline tholeiities with SiO$_{2}$ 46-50wt$%$ and Mg$\#$ 0.44-0.60, similar to Leg 206. Major and minor element variations are consistent with previous fractional crystallization of olivine, plagioclase, Fe-Ti oxides, $\pm$augite; discrete alteration is 1-5vol$%$. The small ranges in major and compatible elements largely reflect phase equilibria control, rather than source differences. However variations in incompatible trace element ratios are best interpreted in terms of two distinct magma batches with slightly different mantle sources. On a plot of Sr-Nd isotopes, Leg 205 samples ($^{143}$Nd/$^{144}$Nd= 0.512937-0.513020 $\pm$8, $^{87}$Sr/$^{86}$Sr= 0.703245-0.705106 $\pm$10) overlap the range of Galapagos Island basalts rather than EPR, except when $^{87}$Sr/$^{86}$Sr $>$0.70400. Elevated $^{87}$Sr/$^{86}$Sr is likely a result of alteration, although leachates have not yet been analyzed. The Nd-isotopic ratios cluster near 0.512950 (one value extends to 0.513020) and good correlation between the Sr-Nd isotopes of less-altered samples suggests mixing between mantle domains of variable enrichment. Rocks from Leg 205 record enrichment within EPR-generated lithosphere and may be the most northerly expressions of the Galapagos hotspot. REE patterns for Leg 170 samples form 2 distinct groups, both overlapping ranges of the Galapagos Islands and regional spreading centers (EPR and Cocos-Nazca spreading (CNS) center). The two groups (Grp 1 LREE ~60$\times$ chondrites; Grp 2 ~32$\times$, both with HREE ~16$\times$ chondrites) may correlate with depth, but their differences cannot be explained by fractional crystallization alone. Consideration of HFSE systematics and REE abundances suggest that mixing between heterogeneous sources and differing degrees of partial melting are required to generate these two groups. On a plot of Hf/Ta vs. (La/Sm)$_{N}$, CNS and EPR lavas (Hf/Ta $>$10, (La/Sm)$_{N}$ $\sim$0.5-1.5) are well separated from Galapagos lavas (Hf/Ta $<$5 and (La/Sm)$_{N}$ $\sim$1-2.5). Leg 170 rocks are within the Galapagos field, as are samples from the 14Ma Fisher Seamount SW of the ODP sites. The origin of igneous units from Leg 170, and by association Leg 205, is best explained as 10-20$%$ degree partial melting from a mantle source $<$30$%$ enriched compared to regional depleted mantle. Mantle mixing between the Galapagos hotspot and the nearby EPR and CNS ridges is recorded in off-axis volcanism offshore Costa Rica. Trace element and isotope results suggest that the spatial and geochemical influence of the Galapagos hotspot is more extensive than previously recognized, and may form a significant part of the oceanic section subducting off Costa Rica, a NSF-MARGINS focus site for studies of crustal recycling at convergent margins.
V34B-05 17:00h
Geochemical and isotopic variations along the Gal\'{a}pagos Spreading Center, 90.8-$97.8\deg$W
Basalts dredged during R/V Maurice Ewing Leg EW00-04 to the Gal\'{a}pagos Spreading Center were analyzed for major elements, trace elements, and Pb-Sr-Nd isotopic compositions via XRF, pneumatic nebulization ICP-MS, and ID-TIMS, respectively. Going westward along axis from $90.8\deg$W, basalts change from E- to T- to N-MORB (Cushman et al., G$^{3}$, 5, 2004); Sr and Pb isotope ratios become less radiogenic and Nd isotope ratios more radiogenic. The highest $^{87}$Sr/$^{86}$Sr (0.70302) and $^{206}$Pb/$^{204}$Pb (19.125) and lowest $\epsilon$$_{Nd}$ (+6.5) values are at $\sim$ $92\deg$W. An overall westward along-axis isotopic gradient is evident (cf. Verma et al., Nature, 306, 1983; Schilling et al., G$^{3}$, 4, 2003). However, a pronounced "step" is present in all isotope ratios at the large propagator at $95.5\deg$W; for example, basalts to the east have $\epsilon$$_{Nd}$ less than +8.0, whereas those to the west generally have $\epsilon$$_{Nd}$ greater than +9.0. Several incompatible element ratios display a broadly similar along-axis pattern; e.g., primitive-mantle-normalized (La/Sm)$_{pm}$, (Nb/La)$_{pm}$, and (Nb/Zr)$_{pm}$ all reach a maximum at $92\deg$W, and decrease around $95.5\deg$W. (Th/Hf)$_{pm}$ values decrease by a more than a factor of two across the propagator. One interpretation is that, over time, mantle derived from the Gal\'{a}pagos hotspot has moved westward in "pulses", gradually mixing with ambient N-MORB-source mantle, and that the propagator marks a boundary between mantle from two such pulses. In Nd-Pb and Sr-Pb isotopic space, our data form a linear array extending from a high-$\epsilon$$_{Nd}$, low-$^{87}$Sr/$^{86}$Sr Pacific N-MORB-type composition to a point between values for the Gal\'{a}pagos plume and Wolf-Darwin components proposed by Harpp & White (G$^{3}$, 2, 2001) to be two of the end-members in the mantle source of the Gal\'{a}pagos Islands. Our data may suggest that the hotspot-derived mantle that affects the spreading center represents a specific, favored mixture of these two components; alternatively, only the plume component is involved, but has a slightly different isotopic composition than previously postulated.
V34B-06 17:15h
On Melting, Dehydration and the Geochemistry of Off-Axis Plume-Ridge Interaction
Interaction between off-axis mantle plumes and the mid-ocean ridge system has been proposed to occur via pipeline-like flows along the base of the lithosphere based on of off-axis variations in geophysical properties such as seafloor morphology, gravity and bathymetry. Off-axis variations in geochemistry from places such as the Galapagos and the Easter - Salas y Gomez seamount system provide an important independent constraint on the nature of flow in off-axis plume-ridge systems. We present results from a series of two-dimensional numerical experiments in which synthetic melt compositions are calculated for a system in which a thermally buoyant, off-axis mantle plume interacts with a nearby ridge axis. These experiments incorporate melting and a number of related dynamical feedbacks, including energy loss to latent heating and viscosity increases due to dehydration during melting. Spatial gradients in synthetic melt properties are compared to observed spatial gradients in radiogenic isotopic ratios in rocks from the Easter - Salas y Gomez system in an effort to constrain the dynamics of mantle flow in off-axis plume-ridge systems. Results indicate that the observed gradients between the ridge axis and the plume require significant heating of the ambient mantle material adjacent to the plume. This heating allows ambient mantle to melt off-axis, creating complex structure in off-axis geochemistry, as seen in the observational data. When increases in viscosity due to dehydration during melting are considered a large viscous plug forms on the base of the lithosphere above the plume, deflecting plume material horizontally to the ridge axis at sub-solidus depths. This leads to extremely sharp spatial gradients in the geochemical properties of synthetic melts, at odds with the observational data. This suggests that the effects of dehydration on rheology are minimal for the case of off-axis plume-ridge interaction.
V34B-07 17:30h
Do Processes of Rhyolite Genesis Change as Icelandic Rifts Drift off of the Plume?
The abandoned Snaefellsnes rift zone in western Iceland was the on-land manifestation of the Mid-Atlantic Ridge between 15 and 7 Ma. The rift zone was abandoned at 7 Ma, after it had drifted westward off of the Iceland hotspot, generally interpreted as a mantle plume. The position of the abandoned rift was initially recognized as the axis of a regional syncline analogous to the syncline developed in response to active rifting. Previous paleomagnetic and geochronologic studies have confirmed the position of the abandoned rift axis. Recent seismic tomography shows that the abandoned rift is also characterized by relatively thin crust ($<$20 km, versus up to 46 km above the plume). In the context of supervising Keck Geology Consortium undergraduate research projects in northwestern Iceland in 2003 and 2004, I have studied several silicic centers erupted at different times along the northern Snaefellsnes rift. A compilation of preliminary geochemical data from the Skagi area near the rift reveals several interesting trends that bear on the origin of silicic magmas as activity in the rift was waning. The compositional spectrum of silicic rocks in this area is from dacite (67 wt.% SiO$_{2}$) to rhyolite (75 wt.% SiO$_{2}$). Positive correlation between Na$_{2}$O and SiO$_{2}$ is consistent with either fractionation or decreasing degrees of crustal melting to get from dacite to rhyolite. However, Zr correlates negatively with SiO$_{2}$, consistent with zircon fractionation, but inconsistent with variation in the degree of melting unless zircon is present in the source, unlikely for the meta-basaltic crust of Iceland. Therefore, I suggest these rocks reflect extreme ($>$90%) fractionation of a basaltic parent. A similar argument was advanced by Furman et al. (1992, J. Pet., 1405-1445) for rhyolites at Austerhorn in eastern Iceland. Compelling arguments have been previously advanced that most rhyolites erupted in the modern rifts of Iceland are the products of crustal melting. The same has been argued for some older centers erupted from the Snaefellsnes rift (Kroksfjordur, 9-10 Ma?). I propose the hypothesis that as a rift drifts off of the plume, and becomes more like a normal mid-ocean ridge (thinner crust), rhyolite genesis by crustal melting becomes uncommon, and that what rhyolites are generated are the result of extreme fractionation of a basaltic parent. Ongoing studies will test this hypothesis by more detailed trace element and O-isotope studies and the systematic study of a series of silicic centers erupted at the northern Snaefellsnes rift over its history of activity.
V34B-08 INVITED 17:45h
Hawaiian Hotspot - Spreading Ridge Interaction in the Late Cretaceous: A Fair and Balanced Look at the Evidence
As is so often the case in years divisible by 4, reality turns out to be quite different from reputation. The Hawaiian hotspot, often righteously promoted as the hotspot that the rest should strive to emulate, was not as stable nor as free from interactions with plate boundaries as some supporters suggest. Mounting geochemical and geophysical evidence shows that in its youth the hotspot not only inhaled, but probably snorted and did shots as well. The purpose of this presentation is to summarize what we know about the Late Cretaceous interaction between a spreading ridge and the Hawaiian hotspot from recent work on the Emperor Seamount chain. At the time of this writing, facts are a commodity to be fabricated, deleted, spun, denied, and denied-that-you-denied; but by the time of this presentation, we (hopefully) will be looking toward the future: can the Hawaiian hotspot's checkered past be treated as a bonus rather than a burden? Plate reconstructions of the Late Cretaceous northwest Pacific place a seafloor spreading center very close to, or even directly on top of, the Hawaiian hotspot. The geochemical effects of this hotspot-ridge interaction are now well documented by work on Ocean Drilling Program samples from Detroit Seamount, the next-to-oldest remaining Emperor Seamount. Basalts recovered from ODP Site 883 partway up the east side of Detroit Seamount have trace element and isotopic characteristics more akin to MORB than to Hawaiian Islands basalts. Basalts from ODP Site 884 at the eastern foot of the seamount are highly depleted tholeiites unlike anything else found so far in the Hawaiian-Emperor chain (Keller et al. 2000, {\it Nature}). Their trace element and radiogenic isotope values are essentially indistinguishable from MORB values (Keller et al. 2000), although triple-spike Pb isotope data are distinct from modern EPR MORB data (Regelous et al. 2003, {\it J. Pet.}). These characteristics were the result of the hotspot melting a greater proportion of a depleted mantle component, whether entrained from the surrounding upper mantle (Keller et al. 2000) or intrinsic to the plume (Regelous et al. 2003), while it was close to the spreading center. The Site 884 basalts are surprisingly old (81 Ma; Keller et al. 1995, {\it Leg 145 Sci. Results}) in comparison to recent results for Site 1203 basalts from near Site 883 (76 Ma; Duncan and Keller 2004, {\it G-cubed}). The unique composition and surprisingly old age of the Site 884 basalts could be due to the fact that we have not drilled a similar location deep on the flank of another Hawaiian hotspot island or seamount. However, my preferred explanation is that the temporal and spatial distribution of hotspot volcanism was also influenced by the nearby spreading center. A modern analog for Detroit Seamount may be a complex seamount platform similar to the Gal\`{a}pagos hotspot-ridge system.