V31B-1428 0800h
Ridge-Hotspot Interactions and Submarine Hydrothermal Activity: Do Seafloor Vent-Sites Even Care?
One of the most intriguing aspects of modern hydrothermal exploration has been the divergence from the predicted abundance of hydrothermal activity along the Mid-Atlantic Ridge close to both the Iceland (anomalously low hydrothermal activity) and Azores (anomalously high hydothermal activity) hot-spots. One recent hypothesis has been that the increase in activity SW of the Azores and almost complete absence of activity along the Reykjanes Ridge may result from waxing and waning of hot-spot activity and concomitant swelling and/or fracturing of ocean crust at the ridge-axis which may perturb the abundance of deep-penetrating faults required to sustain high-temperature hydrothermal circulation. A complication of both the Iceland and Azores ridge-hotspot systems, however, is that in each case the ridge-axis shallows close to the hotspot, to depths shallower than those at which phase separation of hydrothermal vent-fluids would be expected. Consequently, variations in vent-discharge along-axis may have aliased earier plume-based studies. To remove that potential problem, new studies have been conducted along the Central Indian Ridge and on the southern Mid-Atlantic Ridge, close to and away from the Rodrigues and Ascension hotspots. In those locations, ridge morphology indicates a thermal influence from hotspot-ridge interactions without any major perturbation of ridge-axis bathymetry itself. Here, we will present an update of our understanding of hydrothermal activity along these systems, based on geochemical analysis of the overlying water column coupled with seafloor imaging, to deduce what the significance (or otherwise) of ridge-hotspot interactions upn seafloor venting may be.
V31B-1429 0800h
The Iceland Deep Drilling Project (IDDP), a 5 km Deep Drillhole on the Mid-Atlantic Ridge at the Icelandic Hotspot.
Beginning in January 2004 one of the major energy companies in Iceland will drill and flow test a 2.7 km deep geothermal well situated on the Reykjanes Peninsula at the southwest tip of Iceland, where the Mid-Atlantic Ridge emerges from the ocean. The Iceland Deep Drilling Project (IDDP) proposes to take over this well in 2006 and deepen it to 3.7 km and carry out a second flow test at that depth. In 2007, after evaluation of the technical and scientific results from this 3.7 km deep well, and subject to funding negotiations now underway, the IDDP will continuously core the well to 5 km depth, where a third flow test will be carried out. A consortium of the three major energy companies in Iceland conceived the IDDP to investigate the deeper levels of hydrothermal systems. Their aim is to determine if utilizing supercritical hydrothermal fluids would improve the economics of power production from existing high-temperature geothermal fields. This will require drilling a series of wells $>$ 4 km deep, to reach temperatures $>$ 450°C over the next decade. Such deep hot wells present both technical challenges and opportunities for important scientific studies. The IDDP has therefore welcomed participation by the international scientific community and with funding from the International Continental Scientific Drilling Program (ICDP) two international workshops have been held, resulting in a science program involving investigators from $>$ 17 different countries. This first IDDP well will allow sampling fluids from 2.5 km, 3.7 km and 5.0 km and produce $>$ 1.3 km of core from a high-temperature hydrothermal system located in an ophiolite-like environment that is actively forming today. The well at Reykjanes is at a superb location for scientific investigations of the deeper levels of a hydrothermal system similar to black smokers, investigations of a high enthalpy geothermal resource and the coupling of magmatic and hydrothermal systems. These investigations will be a very important contribution to global science and have clear connections to the studies of ridge-hotspot interactions by the Integrated Ocean Drilling Program. Indeed Iceland could be considered as a Mission Specific Platform for such studies.
http://www.iddp.is
V31B-1430 0800h
Short-Period Surface Wave Study of the Upper Mantle Beneath the Reykjanes Ridge: Investigation of the Effect of the Icelandic Hotspot on Sub-Ridge Mantle Flow and Melting
Although it is a slow-spreading mid-ocean ridge, the Reykjanes Ridge exhibits some characteristics typical of faster-spreading ridges, presumably due to far reaching effects of the Icelandic hotspot. Two end-member models exist to describe the hotspot's influence. In one model, plume material spreads out in a radial manner under the lithosphere; in the other model, plume material is preferentially channeled down the ridge axis. We analyze short-period Love and Rayleigh wave records for waves that propagated along the Reykjanes Ridge. These waves were generated by regional earthquakes occurring in the North Atlantic to the south of Iceland, and were recorded by the HOTSPOT and ICEMELT arrays and the GSN station BORG, located on Iceland. We show that surface waves traveling along the ridge are focused into a wide low-velocity channel beneath the ridge, at least 125km across, where the low-velocity channel is consistent with high temperatures and a small amount of melt. The waveforms are sensitive to the magnitude and width of this channel. We extract the phase, group, and amplitude information of narrow-pass filtered waveforms over the period range of ~14-80s. Over ~12000 such measurements are included in an inversion for mantle and crustal shear velocity structure; in addition the joint inversion of horizontally polarized Love and vertically polarized Rayleigh wave data provides a measure of mantle anisotropy. We will present tomographic images of the shear velocity structure of the upper mantle and crust and discuss constraints that our results place on thermal structure, the distribution of melt, and mantle flow beneath the Reykjanes Ridge in association with hotspot influence.
V31B-1431 0800h
Faulting and Magmatism on the Reykjanes Peninsula, SW Iceland
As the Mid-Atlantic Ridge comes onshore at the Reykjanes Peninsula (RP), it bends into an $075\deg$ trend marked by a 2-5 km wide seismic belt. The RP is an oblique spreading environment ($30\deg$ obliquity) manifested as 4 or 5 NE-trending fissure swarms. We used 0.5 m/pixel resolution aerial photographs and field mapping to analyze fracture, fault, and dike orientations, geometries, and interactions along the Krisuvik fissure swarm (KFS) in the central RP. This region is within the active rift zone (plate boundary). The plate spreading is marked at the surface by vertical (and hence dilational) normal faults, strike-slip faults, spatter cone chains, hyaloclastite ridges, eruptive fissures, and tension fractures. KFS normal faults are oriented 035-$055\deg$. In basalts, the faults have vertical fault scarps with offsets of a few cm to 10s of m. These offsets can be partially taken up by fault related surface monoclines. In hyaloclastites, the faults occur as steeply dipping shear planes and deformation bands, indicating a behavioral difference between the two rock types. At the km scale, normal faults are made up of numerous segments 100s of meters long with both right- and left-stepping corrugated geometries. These large segments, in turn, are comprised of smaller (10-40 m) segments, most of which have a left-stepping geometry. Similar fault orientations and en echelon geometries occur in the Reykjanes fissure swarm to the west of the KFS. Left-lateral motion along the $075\deg$ rift axis is partially accommodated by 010-$030\deg$ right-lateral strike-slip faults (i.e., bookshelf style faulting) in the central RP. These faults are characterized by a series of elliptical mounds or push-up structures connected by NE-trending en echelon fractures. Strike-slip faults are sometimes connected to the tips of $055\deg$ trending normal faults suggesting a genetic relationship between the two fault types, although the relative timing of activity of each is unclear. Spatter cones in the KFS are surface expressions of underlying dikes that form long, segmented chains trending $\sim$ $040\deg$ (i.e., sub-parallel to the normal fault traces). Segment lengths range from 40-200 m and have both right- and left-stepping geometries. Postglacial lava flows covered pre-existing faults south of the ends of the hyaloclastite ridges. This region is riddled with vertical fractures that are the surface expression of buried faults. These 025-$035\deg$ fractures (i.e., different to older faults) are highly segmented and have a consistent left-stepping geometry, suggesting right-lateral oblique motion. There has been no apparent change in the rift spreading direction in the last 10,000 years so the stress rotation evidenced by changed fracture orientations is likely the result of a local effect. During the emplacement of a dike the magma pressure pushes against the dike walls and creates a locally perturbed stress field. Where dikes are not parallel to normal faults in the subsurface, magmatism may resolve shear stress onto the faults, inducing oblique slip. In response, faults break through the youngest lava flows at the surface as en echelon vertical tension fractures oriented obliquely to the trend of the underlying fault. The fault geometries suggest a likely control on the mechanics of fault evolution that varies as a function of magmatic activity. During amagmatic periods, oblique rifting tectonic stresses dominate, driving the linkage of the en echelon segments that developed during a prior magmatic episode.
V31B-1432 0800h
Late magmatic underplating of oceanic crust at the outer V{\o}ring margin, Norway, Euromargins 2003 OBS experiment
The V{\o}ring margin off mid-Norway is a volcanic passive margin where large volumes of magmatic rocks were emplaced during the breakup of the Norwegian-Greenland Sea during the earliest Eocene. In 2003, an ocean bottom seismometer/hydrophone survey was acquired on the V{\o}ring and Lofoten margins. One profile crosses the V{\o}ring Plateau and a bathymetric high northeast of the East Jan Mayen Fracture Zone (EJMFZ). The P-wave data were modeled by a combined forward ray-tracing and inversion procedure, giving a 2D velocity model. The model shows a rapid transition from continent to oceanic crust (COT) located under the zone of seaward dipping reflectors, similar to earlier results. Maximum igneous crustal thickness was found to be 17.5 km, but the thickness is in general 5 km lower than found by previous studies nearby. From the COT the igneous crustal thickness decreases from 17.5 to 9 km over a distance of 90 km, indicating abating magmatism over 4-5 Ma after continental breakup. West of that, the bathymetric high is supported by increased (12-15 km) crustal thickness. P-wave velocity is well constrained to be 7-7.25 km s$^{-1}$ in the lower crust of the high, indicating a gabbroic composition. The sedimentary pattern over the high seen on recent single or multi-channel reflection seismic profiles, show that the high is not a primary feature, but was created through later uplift.To explain later magmatic underplating of oceanic crust, magmatic and plate kinematic development of the Norway Basin around the extinct spreading axis of the Aegir Ridge is investigated. V-shaped lineations seen in the satellite gravity field indicate NE migrating asthenosphere zones with increased melt production potential, delivering to the Aegir Ridge. Due to low magnetic track coverage, the EJMFZ was reinterpreted from the satellite gravity field. Comparing the new interpretation to published spreading poles, spreading appears more asymmetric than previously believed: East of the Aegir Ridge, spreading followed Greenland vs. Eurasia up to about A13 time. After that, the spreading followed published poles located southeast of Iceland. The latter poles can, however, now be applied to the whole opening history on the western side of the ridge. The revised flowlines show that the V-shaped lineations are reasonably symmetric with respect to spreading direction, even if they are not symmetric with respect to the ridge trend. Estimated asthenospheric transport velocity along the Aegir Ridge is 0.35 cm/a or lower for the Paleogene. It is suggested that partially melted asthenosphere was transported northeastwards when spreading in the Norway Basin ceased, and released through episodic ascent northeast of the EJMFZ. However, there is a need to speed up the transport in the Neogene to cover the distance during the available time, which could be related to the increased activity on the Iceland hot-spot.
V31B-1433 0800h
The Liquid Line of Descent of the Alkali-Olivine Basalts From Floreana Island, Galapagos, Ecuador
Melting experiments have been performed on two alkali-olivine basalts, Fl03-51 and Fl03-75, from Floreana Island of the Galapagos, Ecuador. The experiments were conducted in a 1 atm gas-mixing furnace, and oxygen fugacity was set at the Ni-NiO buffer. The experiments were designed to define the liquid lines of descent for the two magmas, and to test whether high-MgO Floreana magmas are related to low-MgO magmas by crystallization. The samples were selected so as to minimize incorporation of xenolithic debris, which can be pervasive in Floreana lavas. Sample Fl03-51 contains 7.60 percent MgO and crystallizes plagioclase at 1200 C and olivine at 1180 C. The high MgO sample, Fl03-75, has 12.39 percent MgO and has olivine on the liquidus at 1305°C. Compositional variations as a function of temperature indicate that the two liquid lines of descent intersect at about 1180 C. Microprobe analyses of olivine crystals show that sample Fl03-51 more closely resembles the primary magma. Sample Fl03-75 contains an abundance of xenolithic debris, especially olivine, and is not directly related to Fl03-51 by crystallization.
V31B-1434 0800h
Shear Wave Splitting Beneath the Galapagos Archipelago
We report measurements of teleseismic shear wave splitting in the Gal\'{a}pagos Archipelago. The inferred lateral variations in azimuthal anisotropy allow us to examine the dynamics of an evolving hotspot-ridge system. The data are from SKS and SKKS phases, as well as S waves from deep sources, recorded by a relatively dense network of 10 portable broadband seismometers deployed from 1999 to 2003 for the IGUANA (Imaging Gal\'{a}pagos Upwelling and Neotectonics of the Archipelago) experiment and from the GSN broadband station in Santa Cruz (PAYG). We find a delay time between fast and slow shear waves of 0.4 to 0.9 s and fast polarization directions of N85-$90\deg$E beneath five stations at the leading and southern edge of the archipelago. Despite clear seismic signals, we did not find any anisotropy at the six stations located in the interior of the archipelago. For those stations that show shear wave splitting, there is an increase in the delay time toward the expected location of the Gal\'{a}pagos hotspot at the western edge of the archipelago. With the exception of Espa\~{n}ola, fast polarization directions (N85-$90\deg$E) are close to the current direction of absolute plate motion of the overlying Nazca plate (N$91\deg$E). The lack of azimuthal anisotropy in the interior of the archipelago is interpreted as an absence of strongly oriented mantle fabric beneath these stations. The apparent isotropy in this dynamic region, where we expect considerable mantle strain, is surprising. It is not likely that the olivine {\it a}-axis is oriented vertically beneath the interior of the archipelago as the Gal\'{a}pagos plume is thought to lie at the western edge. It is also unlikely that there are two layers of perpendicularly-oriented anisotropy which are solely confined to the center of the archipelago. However, there appears to be some correlation between the region of apparent isotropy and a zone of anomalously low upper mantle velocities imaged beneath Santiago and Marchena from surface waves by Villagomez and others, though the low-velocity region is spatially more confined. This pattern suggests that the presence of melt in the upper mantle may weaken the effects of fabric on shear wave splitting, as suggested by Holtzman and others. An alternative explanation is that the flow field in the near ridge setting is complex, resulting in apparent isotropy. Due to the very young lithospheric age and to the effect of both the neighboring ridge and Gal\'{a}pagos hotspot, the lithosphere thickness is likely thinner than 40 km. Therefore no more than about 40% of the recorded delay time could originate from the lithosphere. We propose that the splitting pattern in the Gal\'{a}pagos Archipelago may be the result of plate drag and frozen lithospheric anisotropy which, beneath the center of the archipelago, is weakened by the presence of melt in the upper mantle.
V31B-1435 0800h
Gabbro fracturing and elemental analysis in the Costa Rica margin (ODP Leg 205): Core-log integrated high-resolution study
Costa Rica margin is an important area for studies of the seismogenic zone and subduction factory. Building on Ocean Drilling Program (ODP) Leg 170 coring and logging while drilling (LWD) at the same sites, Leg 205 drilled three sites to determine the igneous and alteration history of the upper most part of the down going plate, to characterize the hydrological regime above and within the decollement, and successfully installed long-term borehole observatories to monitor downhole pressure and temperature and sample fluids and gases. Among the several highlights from the results, coring and logging of thick gabbro zone at Site 1253 was significant to detail the fracture distribution, structure and nature of oceanic crust through core-log data integration. Even thought logging runs in the sediment section was failed due to the hole conditions, triple combination and FMS-sonic tool strings made successful measurements. Post-cruise research work included quality control on several wireline logs and processing of Formation MicroScanner (FMS), and interpreting these logging data and core sample analysis results in integrated manner. Among several downhole logging results, FMS images are used to characterize structure and fabric in the igneous units. As the hole conditions were good in most of the igneous intervals, the textural and structural variation is clear between the sections and also shows differences within the section itself. Moreover, mineralogic results from the recovered cores are combined with logging data through level-by-level multi-mineral evaluation with an optimized simultaneous equation solver and model combining method. Due to the good logging results and simple mineral composition of the single rock type, the results from the process filled mineral percentage of the no core recovery depths as well. High-resolution core-log integration in this study enhanced the understanding of structural process and nature of the oceanic crust of this particular region.
V31B-1436 0800h
Evidence for a Depleted Mantle Component in Mildly Alkalic High-MgO Basalts From Mt. Marion Dufresne, Kerguelen Archipelago, Southern Indian Ocean
Kerguelen Archipelago flood basalts (29-24 Ma) were erupted onto the Northern Kerguelen Plateau when the Southeast Indian Ridge was 200-400 km northeast of the Kerguelen hotspot. A depleted mantle component is evident in older (28-29 Ma) transitional-tholeiitic basalts on the archipelago, but is absent in younger (24-25 Ma) mildly alkalic lavas. This trend towards more alkalic volcanism primarily reflects decreasing extents of partial melting and increasing depths of melting in the mantle source. A $\sim$700 m vertical section of lavas exposed at Mt. Marion Dufresne, in the southern part of the Plateau Central, contains both transitional-tholeiitic and mildly alkalic basalts (A.I. -0.8 to +2) that are geochemically and isotopically variable. The base of the section is dominated by mildly alkalic, highly fractionated, aphyric to plagioclase-phyric basalts (3 to 5.7 wt% MgO) that trend towards decreasing alkalinity with increasing height. These basalts show a limited range of isotopic compositions ($^{87}$Sr/ $^{86}$Sr= 0.7048-0.7050; $^{143}$Nd/ $^{144}$Nd= 0.5126-0.5127; $^{207}$Pb/ $^{206}$Pb= 15.54-15.57) that are similar to enriched 24-25 Ma mildly alkalic lavas in the southeastern part of the archipelago. These lavas are overlain by lower SiO$_{2}$, high-MgO (6.7-11.4 wt%) olivine-phyric basalts (Fo$_{80-88}$) that straddle the tholeiitic-alkalic divide. These high-MgO basalts record a prominent depleted mantle signature similar to that observed in the older 28-29 Ma basalts and span a range of isotopic compositions from $^{87}$Sr/ $^{86}$Sr= 0.7037-0.7048, $^{143}$Nd/ $^{144}$Nd= 0.5127-0.5129, and $^{207}$Pb/ $^{206}$Pb= 15.48-15.52. This study demonstrates the first occurrence of a depleted mantle component in mildly alkalic basalts from the Kerguelen Archipelago. The superposition of younger lavas with a depleted mantle signature over older, mildly alkalic, enriched lavas is also a situation that has not yet been observed on the Kerguelen Archipelago. This depleted signature is not interpreted to indicate intrinsic plume heterogeneity but is due to mixing between plume-sourced magmas and a depleted asthenospheric component.
V31B-1437 0800h
Evidence for Topographic Swells Beneath Gravity Lineations in the South Pacific.
The Gravity Lineations Intraplate Melting Petrologic and Seismologic Expedition (GLIMPSE) experiment is investigating the origins of the gravity lineations in the South Pacific and their relationship to intraplate volcanism at the Sojourn and Hotu Matua ridge systems. Using the satellite free air gravity anomaly and high quality shipboard bathymetry in the GLIMPSE study region, we have calculated the Mantle Bouguer Anomaly (MBA) and the residual Mantle Bouguer Anomaly (rMBA). In the MBA, we see strong negative anomalies (~20-40 mGals) beneath the Sojourn and Brown Ridges as well as beneath the Hotu Matua volcanic complex, indicating low-density material beneath the ridge systems in the form of thickened crust or lower density mantle material. We subtract the effects of surface loading of an elastic plate by the volcanic ridges to find the residual topography and the residual MBA. The residual seafloor topography shows a systematic ~200 km wavelength undulation in the seafloor with a strike roughly parallel to the spreading direction, producing residual bathymetric highs of 100-300 m beneath the Sojourn and Hotu Matua ridge systems. There are broad (100 km wide) negative anomaly bands in the rMBA of ~10-40 mGals beneath the entire length of the Sojourn Ridge to the East Pacific Rise and discontinuously from west of Hotu Matua to the EPR. Similar features are found in the extension of the Puka Puka ridge into the Rano Rahi seamount field. The swells beneath the intraplate volcanism and the gravity lineations require a contribution from subsurface positively buoyant low-density material. The seismic refraction and Rayleigh wave dispersion experiments corroborate the existence of anomalies at depth. The swells contradict lithospheric cracking models of the formation of the gravity lineations that predict the gravity lineations should be located over bathymetric troughs, and favor dynamic models such as small-scale convection.
V31B-1438 0800h
Crustal Structure Beneath the Gravity Lineations in the South Pacific From Seismic Refraction Data
The GLIMPSE ({\bf G}ravity {\bf L}ineations, {\bf I}ntraplate {\bf M}elting, {\bf P}etrologic and {\bf S}eismic {\bf E}xpedition) Experiment explores the origin of cross-grain gravity lineations and associated volcanic ridges in the central South Pacific. The active-source seismic refraction component of GLIMPSE Experiment investigated the crustal thickness and velocity structure beneath the Sojourn Ridge and the Hotu Matua ridge system (~116.5-115.5{\deg}W, 16.0-12.5{\deg}S). Forward modeling of P-wave first arrival picks along a 450 km refraction line shot to 16 ocean bottom seismometers shows crust thickening beneath the ridges, ranging from the 5.3-5.5 km typical of oceanic crust along this segment of the Southern East Pacific Rise (SEPR) to a maximum of 9.3-9.8 km beneath the Sojourn Ridge. Several small volcanoes crosscut by the refraction line also exhibit root structures, suggesting a consistent pattern of periodic ridge-perpendicular crustal thickness variations with a scale length of 50-75 km. In addition, crustal thickening is consistently asymmetric beneath the ridges, extending further to the north than to the south beneath the associated bathymetric highs. Crustal thickening is primarily associated with variations in layers 2A, 2B and 2C, indicating increased dike emplacement and eruption of extrusives. Small-scale asthenospheric convection and channeling of low-density material from the South Pacific Superswell, models previously invoked to explain the observed free-air gravity pattern in the SEPR region, may be similarly responsible for the periodic changes in crustal thickness. Discrimination between the two models requires further research on upper mantle structure.
V31B-1439 0800h
Petrology and Evolution of a Monoclinally-folded Paleo-rift Lava Sequence, Vatnsdalsfjall, Northern Iceland
The Hvammur-Breiden area of Vatnsdalsfjall in northern Iceland lies 15 km east of the axis of the abandoned Snaefellsnes rift zone that was an active spreading center from 15-7 Ma. It contains an unusual and exceptionally well exposed monoclinal flexure that deforms a thick volcanic sequence produced during the waning stages of rift-related volcanic activity. A sequence of thin basalt flows thickens dramatically into the trough of the monocline , leading upwards to a 150-meter thick basaltic unit called the Hjallin Lens, mapped originally as an intrusion but recently reinterpreted as extrusive (McClanahan, 2004). Flows that are part of the monocline and flows that overlie the monocline can be distinguished petrologically and geochemically, although there is no evidence of erosion or soil development between the two sequences. The monocline sequence contains several silicic pyroclastic units interlayered with basalts of varying textures, whereas the overlying sequence contains only fine-grained intergranular basalts. In plots of major and trace element ratios, the monocline basalts display a high degree of compositional variation which is absent in the overlying sequence. In Harker-type variation diagrams the two sequences show two distinct trends. Both sequences show consistent Zr/Nb ratios less than 10, signifying a plume-influenced MORB source. Enrichment in these and other light incompatible elements increases upwards. but a substantial gap in Zr and Nb concentration occurs at the contact between the two sequences. Two new argon-argon dates, one from the top of the monocline sequence (7.62+/-0.32 Ma) and one from the overlying Hjallin Lens (6.98+/-0.18 Ma; McClanahan, 2004) bracket a period of 0.60 Ma (average) that encompasses both the formation of the monocline and the eruption of the intervening undeformed thin basalts that thicken into its trough.This interval also constrains the age of displacement along a normal fault running just east of the monocline axis. The monocline is interpereted as a shallow structure that formed incrementally during the eruption of the overlying thin basalt flows above an upward-propagating normal fault that ultimately displaced the entire sequence. The flexure is likely the result of local subsidence caused by a change in spreading rate and/or contemporaneous crustal loading due to renewed magmatism, which culminated in the eruption and ponding of the thick basalt flow of the Hjallin Lens.
V31B-1440 0800h
The Elusive Intrusive: A Petrologic, Structural, and Geochemical Analysis of a Basaltic Body in an Abandoned Rift, Vatnsdalsfjall, Northern Iceland
The Hjallin Lens is a thick lenticular basaltic unit in Vatnsdalsfjall near the southern end of the Skagi Peninsula, northern Iceland. The lens was named and first studied by Annells (1968), who interpreted it as a shallow laccolith. However, recent field observations followed by extensive petrographic, geochemical, and geochronolgical analyses support an alternate interpretation: that the lens is actually an extrusive feature formed by the ponding of a thick basalt flow in a topographic low developed above a monoclinal flexure (Ackerly, 2004). The lens extends approximately 3000 meters NNW-SSE and 800 meters ENE-WNW. It reaches its greatest thickness (over 150 meters) near its southern end; its upper contact, however, is not exposed along the ridge crest. The lithology of the lens is virtually uniform throughout: a very hard, fine-grained aphyric basalt that has cooled in well developed columns. It is underlain by a rhyolitic ash flow tuff (previously interpreted as a chilled margin) along a sharply defined lower contact. At the southern end of the lens the ash flow tuff overlies a poorly lithified sediment with well preserved fragments of carbonized and silicified wood. The presence of these relatively undisturbed surficial units immediately beneath the lens is one line of evidence supporting the extrusive origin of its basalt. Geochemically the basalt of the lens is a tholeiite enriched in incompatible elements and with Zr/Nb values less than 10, indicating a plume overprint of N-MORB composition. The striking textural, mineralogical, and geochemical homogeneity of the lens basalt suggests that the entire body must have cooled very rapidly, without the internal differentiation that would likely have accompanied a slower-cooling intrusive basalt of this thickness. New argon-argon dating provides an age of 6.98+/-0.18 Ma for the lens basalt, indicating that it formed during the waning stages of volcanic activity in the Snaefellsnes Rift, which was an active spreading center from about 15-7 Ma.