T13B-1355 1340h
Morphology and Tectonic Evolution of Endeavor Deep
Endeavor Deep is located on the Nazca/Juan Fernandez plate boundary near the triple junction of the Pacific, Nazca and Antarctic plates. The deep is the tip of the northward propagating East Ridge, which defines the eastern side of the microplate and is presently exposing ~3 Myr old oceanic crust created at the ultra-fast spreading (~150 km/myr) East Pacific Rise. Recently collected high-resolution EM300 bathymetry, deep-tow DSL120 sidescan, surface-towed magnetics, and near-bottom JASON II observations provide important details about the tectonic character and origin of Endeavor Deep. These data define a 70 km-long, 40 km-wide, and 3 km-deep rift which shoals and narrows toward the rift tip to the NW and is deeper and wider away from the rift tip toward the SE. The southern wall of the rift is uplifted and has a characteristic flexural profile. The northern wall is also uplifted, however, the classic flexural profile is complicated by the presence of a large EW-trending massif, which appears to be a rift-truncated compressional ridge emplaced during a phase of NS-oriented compression. Along both rift walls, a series of terraces suggest a series of down-dropped blocks associated with ongoing extension. Along the rift floor, a relatively flat, featureless bottom in the NW evolves into hummocky terrane in the central part of the basin that is characterized by volcanic features reminiscent of 1-2 km diameter pancakes in plan-view. Farther to the SE, tectonic lineations and pillow ridges oriented parallel to the trend of the rift valley dominate the basin floor. Magnetic profiles across this portion of the survey area indicate a well-formed central magnetic anomaly with a width equivalent to a spreading rate of 20 km/Myr, which is predicted by tectonic reconstructions of the plate boundary. Overall, these observations define a four-phase evolution of Endeavor Deep: 1) initial crustal formation at the ultra-fast spreading East Pacific Rise ~3 Ma, 2) regional compression and thrust faulting associated a shear couple along the Nazca/Juan Fernandez plate boundary ~1-3 Ma, 3) initial amagmatic rifting that began about 1 Ma, and 4) initial volcanism evolving to super slow seafloor spreading that began prior to 0.78 Ma.
T13B-1356 1340h
Magnetic Properties of Ocean Crust from the Walls of Endeavor Deep: Implications for the Source Layers of Marine Magnetic Anomalies
The presence of lineated marine magnetic anomalies in the ocean basins is the foundation for plate tectonics and global plate reconstructions. The relative importance of the source layers of these anomalies, however, is still poorly constrained. Identifying the source layer(s) of the magnetic anomalies is important for present-day models of crustal accretion that require crustal rotations of up to 20-30 degrees. These rotations may result in anomalously skewed seafloor magnetic anomalies, which could significantly alter global plate reconstructions. To address the question of the magnetic sources at ultra-fast spreading rates, we have measured the magnetic properties of 135 precisely located rock samples collected with JASON II from the walls of Endeavor Deep. Endeavor Deep is the tip of a propagating rift on the Nazca/Juan Fernandez plate boundary. It exposes a 70 km-long, 40 km-wide, and 3 km-deep section of ~3 Myr old upper oceanic crust (layers 2A and 2B) created at the ultra-fast spreading (~150 km/myr) East Pacific Rise. Measurements of natural remanent magnetization, magnetic susceptibility, and median destructive field define three "stratigraphic" units and are characterized relative to similar measurements by Pariso and Johnson (1991) on samples from DSDP Hole 504B. The upper 200 m of extrusives were likely altered by low temperature oxidation, and carry a relatively low magnetization (0.4 A/m). The underlying 300 m of extrusives carry a relatively high magnetization (2.1 A/m), which likely contributes significantly to the amplitude of the magnetic anomaly signal. The transition zone from extrusive layer 2A to intrusive layer 2B (pillow basalts to dykes) and possibly the upper portion of layer 2B carries an unexpectedly high magnetization (1.6 A/m) and likely contributes to the magnetic source layer as well. A dramatic increase in susceptibility is present below the base of layer 2A that has not been previously observed.
T13B-1357 1340h
Structure of the Upper Crust Exposed at Endeavor Deep: Implications for Crustal Accretion at Ultra-Fast Spreading Rates
Endeavor Deep lies at the tip of the propagating spreading center defining the Juan Fernandez/Nazca plate boundary. This 3 km-deep, amagmatic basin,which rifted into ultra-fast spread 3Myr old Nazca Plate crust, was recently surveyed and sampled with Simrad EM300, DSL-120 and ROV Jason II. Over 140 structural orientations measured from Jason II video of the south rift wall show that flows in extrusive layer 2a strike north-south and dip shallowly to the west, while dikes in intrusive layer 2b strike east-west and dip steeply to the south. Using a general model for crustal accretion in which dikes are emplaced vertically and extrusives horizontally, a rotational history is determined for the 3 Myr old crust exposed in the walls of the deep. Multiple rotations are necessary with two-fold intent, first to return structure to the original off-axis orientation prior to tectonic reorganization; second to account for rotations involved in the process of accretion. Tectonic events are first addressed with a 10-25 degree rotation about a horizontal rift parallel axis to account for flexural uplift. Next a 65-degree rotation applied about a vertical axis to return magnetic lineation 2a to its proposed paleo-off axis orientation. After rotation, intrusive and extrusive populations are strike parallel (N5E). In this orientation, dikes average 65-degree dip away from and extrusives 25-degree dip towards the ridge axis. This generally conforms to observations at Hess deep, Blanco FZ, and ODP hole 801C. The second goal of rotation is to account for off axis adjustments during crustal accretion; a 25-degree rotation about a ridge parallel, horizontal axis returns the average dike inclination to vertical and the extrusive dip to horizontal.
T13B-1358 1340h
Fresh Abyssal Peridotites -- not an Oxymoron!
The study of the Earths mantle and the generation of its crust has been hampered so far by the lack of unaltered mantle rocks from mid-ocean ridges. Whereas ophiolites, orogenic peridotites and xenoliths sometimes provide this freshness, their often unknown tectonic setting sets limits for a one-to-one transfer of the results to processes on mid-ocean ridges. Abyssal peridotites are usually characterised by a high degree of serpentinization, usually in the range of 70% up to 100%. Even this alteration still allows the analysis of relict mineral phases and thus the investigation of major processes affecting the samples, but their study until now limited by this restriction. Completely fresh abyssal peridotites would allow bulk chemical studies and mineral separation studies with far-reaching implications for mantle evolution. So far however, unaltered abyssal peridotites have remained almost a contradiction in terms and as rare as hens teeth. A few samples exist, mostly mylonites. During the 2001 AMORE expedition to the ultra-slow spreading Gakkel Ridge, two dredge hauls recovered several kilograms of very fresh residual and plagioclase-bearing peridotites with a degree of alteration between 20% and 0%. Even including those samples, all of the truly fresh mantle peridotites dredged or drilled from the global system of mid-ocean ridges could fit into a single five-gallon sample bucket. Nevertheless, the investigation of these few fresh rocks has already been a key part in studies of radiogenic isotopes, the effects of alteration, and melt migration. More and better samples are urgently needed. In summer 2004, the Polarstern ARK XX-2 returned to the amagmatic 3$\deg$E region of Gakkel Ridge for a more detailed sampling and mapping of this region. Gakkel Ridge is an ultraslow spreading orthogonal mid-ocean ridge, where ultramafic exposure is common, particularly in a zone from 3$\deg$E to 12$\deg$E, a distance of over 100km, where exclusively ultramafic lithologies were dredged. A previous fresh dredge haul PS59-235 at 84$\deg$ 37.5'N and 4$\deg$ 13.9'E, was nearly exactly repeated, numbered as PS66-238. It recovered 275 kg of absolutely fresh peridotite. They are stunningly fresh: most are still olivine green though have sometimes a narrow rim of reddish weathered olivine. The sometimes huge size of the samples is an appealing fact. The biggest sample of dredge haul PS66-238 is a boulder 70x50x40cm weighing over 100 kg, directly followed by a 50x40x30cm boulder of 52 kg. Onboard hand sample and thin section investigations confirm the near complete lack of alteration. Most samples show 100% preservation of all primary mineral phases. After hand specimen and thin section investigations, plagioclase-bearing and plagioclase-free mantle peridotites occur in almost equal proportions (38% and 44%, respectively). Among those are most of the mantle lithologies found on mid-ocean ridges, ranging from enstatite-dunites up to lherzolites with 12% cpx. Two pyroxenite samples are found as well and some samples show pyroxene-rich bands. Another 17% consists of matrix-supported breccias with clasts of the aformentioned lithologies. The degree of deformation ranges from undeformed to one mylonitic sample. Overall, dredge haul PS66-238 will provide enough absolutely fresh sample material to allow detailed studies of mantle processes and might be therefore of great interest for the mantle petrology community.
T13B-1359 1340h
Hydrosweep Measurements During the Expedition ARK XX-2 to Lena Trough and Western Gakkel Ridge
The region of Lena Trough and Western Gakkel Ridge in the Arctic Ocean was the object of an expedition in the summer of 2004. This region is of particular geoscientific interest because of its extremely slow spreading rates and the variety of morphologic forms that are produced in this tectonic environment. Therefore, the multibeam measurement system was of particular importance to the scientific goals of the cruise. The main characteristic of the Hydrosweep DS-2 deep-water sounding system aboard RV Polarstern is the $90\deg$ or $120\deg$ coverage angle in which the seafloor is depicted with 59 specific values for water depths perpendicular to the ship's long axis. The accuracy of the measurement is approx. 1% of water depth, the frequency of the acoustic signal is 15.5 kHz. The refraction of the sonar beams was corrected by automatic crossfan calibration. By regular transmission and measurement of a sweep profile in the ship's longitudinal direction and comparison of the slant beams with the vertical beam, the mean sound velocity over the vertical water column is determined and is used for the depth computation. The data collected include depth, sidescan (2048 values per scan), and backscatter information on each of the 59 beams. During this cruise, the Lena Trough was surveyed systematically for the first time by a multibeam sonar system. The recorded area has an expanse of approx. 100000 km$^{2}$ and connects previously mapped areas of the Eurasian - North-American plate boundary between Fram Strait and Gakkel Ridge. The region of Western Gakkel Ridge, mapped in 2001 (AMOR-Expedition) by RV Polarstern and USCGC Healy (USA), was extended by two more profiles (each 220 km long) along the ridge. In order to produce working maps for the expedition, the multibeam sonar data were gridded with a spacing of 50 m, producing plots with various contour line intervals. For further morphological interpretation of Lena Trough and Gakkel Ridge slope magnitude maps, slope direction maps and two-dimensional slope histograms were produced. They show clearly the predominant slope direction and are helpful in distinction between sediment basins, volcanos and ridge flanks.
T13B-1360 1340h
Spatial and Geochemical Variations of Lavas Exposed Along a Crustal Section in the Blanco Transform: Insights into Accretion of the Upper Oceanic Crust at the Southern Juan de Fuca Ridge
Opportunities to examine the architecture of the upper oceanic crust are limited to rare tectonic windows, ocean drilling and ophiolites. Studies at the Hess Deep Rift, which exposes fast-spread crust, suggest a 4-D model for accretion in which dikes transport magma along-axis and construction of the extrusive unit continues off-axis. Additionally, density filtering restricts lava compositions to those of lowest density while higher density magmas represented in the dike unit never reach the surface. Within this context, we examined a suite of lavas exposed along the Blanco Transform (BT). The western portion of the BT consists of a deep trough and a steep northern scarp, providing a window into upper crust generated at the southern end of the intermediate-spreading Juan de Fuca Ridge. During the 1995 Blancovin dive program, 53 samples of lavas were collected along an $\sim$11 km transect of the north wall. Whole rocks were analyzed for major and high abundance trace elements by DCP, and for low-abundance trace elements by ICP-MS. All of the samples are incompatible-element depleted N-MORB, with variations in trace element ratios (e.g., La/Sm, Zr/Y) suggesting minor mixing between a depleted and enriched source. As a whole, the lavas exhibit a large range in extent of crystal fractionation, extending from relatively primitive MORB (Mg\# 64) to evolved FeTi basalts (Mg\# 40, 15wt% Fe$_{2}$O&_{3}$, 3wt% TiO$_{2}$). The diversity of compositions among the BT lavas contrasts with the more limited range in lava compositions found at Hess Deep and Hole 504B. In terms of spatial variations, five vertical dives located $\sim$1 km apart in the center of the study area show systematic variations in extent of crystallization with depth. Primitive upper lavas (avg. Mg\# 60) grade downward into lower evolved lavas (avg. Mg\# 48). This finding contrasts with a simple view of lava eruption during which an eruptive cycle begins with the emplacement of primitive (recently replenished) magmas, forming the base of the lava sequence, followed by continued cooling and crystallization, leading to more evolved lavas toward the top of the unit. Calculated density relations for these lavas reveal, however, that the highest density lavas (low MgO, high FeO) form the base of the lava unit and are overlain by progressively lower density magmas. Density variations both between dikes and lavas and within the lava unit may be key to understanding upper crustal architecture and accretion in space and time.
T13B-1361 1340h
The Influence of Ridge Migration on Magmatic Segmentation of the Mid-Atlantic Ridge, 22-36 deg N and 25-35 deg S
The mid-ocean ridge (MOR) system is not uniform. Systematic changes in depth and shape subdivide the ridge into discrete spreading segments. It is not well understood how such morphological changes develop, but this spatial diversity is probably a result of variations in magma supply. Recent observations of ridge bathymetry (Carbotte et al. Nature, 2004) indicate magmatic segmentation at fast and intermediate spreading centers is linked to the migration velocity (rate and direction) of the spreading axis over the mantle. In this investigation, inter-segment elevation changes are examined within the context of absolute plate motions, with particular emphasis on slow-spreading MORs. Multibeam bathymetry data from the slow-spreading Mid-Atlantic Ridge (MAR 22-36 deg N and 25-35 deg S latitude) is used to inspect the relationship between the direction of ridge migration and various morphological ridge characteristics. Moving averages, with swaths of varying widths, smooth the axis depth in order to eliminate local topographic anomalies. For each first- and second- order segment, the axial depths of the segment's ends and midpoint, as well as average depth, are measured and compared across transform faults (TFs) and non-transform offsets (NTOs). These values are compared as a function of distance from each TF or NTO. At fast and intermediate spreading centers Carbotte et al. observed that leading segments, whose offsets are in the direction of ridge migration, are typically shallower and morphologically robust, relative to trailing segments. Our preliminary results along the MAR suggest a weak correlation in the slow spreading environment.
T13B-1362 1340h
An Investigation of Shallow off-axis Crustal Structure and Ridge Segment Morphology Along the Southeast Indian Ridge
Along the Southeast Indian ridge (SEIR) there is a systematic variation in axial morphology and axial depth with distance away from the Australian Antarctic Discordance, an area of cold mantle downwelling. Since spreading rate (72-76mm/yr) and mantle geochemistry appear constant along this portion of the SEIR, variations are attributed to a gradient in mantle temperature. In this study we report on a multichannel seismic survey investigation of axial and ridge flank structure along the SEIR. There are three distinct forms of ridge crest morphology found within our study area: axial highs, rifted axial highs, and shallow axial valleys. Axial highs have shallow magma lens (1330-1820 m), and thin layer 2a (240-440 m). Rifted axial highs have a deeper magma lens (2000-2400 m) and thicker layer 2a on-axis (320-540 m). Beneath shallow axial valleys, no magma lens is imaged, and layer 2a is thick on-axis (440-990 m). There is a step-like transition in magma lens depth, and layer 2a thickness with changes in morphology along the SEIR. We have also found that the presence of a melt lens on-axis corresponds with areas of lower relief abyssal hills on the ridge flank. This relationship suggests that the portions of ridge segments with a melt lens at present have persisted as regions of greater melt delivery in the past. This study focuses on the shallow off-axis crustal structure. We will investigate how the geometry and thickness of layer 2a differs within and between segments as a function of axial and ridge flank morphology. There is evidence of significant thickening of layer 2a off-axis (within 1-3 km) at axial high segments, with no thickening seen at rifted axial high, and shallow valley segments. The relationship between layer 2a structure (geometry and thickness) and abyssal high relief will be examined to assess modes of abyssal hill formation at this ridge (e.g. via split axial volcanoes or horst and graben).
T13B-1363 1340h
Geochemistry of basalt from the Ayu Trough, equatorial western Pacific
We analyzed dredge rock samples collected from the Ayu Trough, a divergent margin located between the Philippine Sea and Caroline plates in the equatorial western Pacific. The region may be divided into three sections from north to south on the basis of geochemistry and axial morphology. The recovered Ayu Trough basalt shows a large geochemical variation from N-MORB to relatively enriched-MORB. The N-MORBs were mainly derived from the southern and northern sections, while the E-MORBs were sampled from the middle section ($2\deg$20'N - $4\deg$N). The variations in major and trace elements between the N-MORB and E-MORB are too large to be explained by melting of uniformly depleted source mantle alone. Instead, mantle heterogeneity where both depleted mantle and enriched end-member exist may explain the overall geochemistry. The composition of depleted component in the basalt appears to be close to that of the depleted mantle source, whereas the enriched end-member may be close to that of EM. Sr-Nd-Pb isotope data also supported the mantle heterogeneity argument. We infer that the enriched component has different mineralogy with respect to the depleted component and suggest garnet-pyroxenite vein in the mantle as the source of the enrichment. According to this argument, the samples from the middle section then may have originated from pyroxenite-rich part of the mantle. It appears that such enriched parts are unevenly distributed under the Ayu Trough. In addition, variations in the degree of melting may have also contributed to the variability of incompatible elements especially in the middle section. From one of our dredges that targeted the axial volcanic ridge in the middle section, we found two distinct populations of samples where one has much lower degree of melting than the other. Although the exact cause of the contrasting degree of melting is unclear, it is possible that the samples with low degree of melting were produced after the spreading has halted or when the trough was slowing down, which then led to the emplacement of melt from the lower part of melting regime.
T13B-1364 1340h
Thermal Models for Southern Mexico and Guatemala and the Position of the Volcanic Belt
A finite element method is applied to model the thermal structure of the subducted Cocos plate and overlying mantle wedge beneath the southern part of Mexico and Guatemala. A numerical scheme solves a system of 2D Navier-Stokes equations and a 2D steady state heat transfer equation with strong temperature dependent viscosity. The geometry of the subducting Cocos slab is inferred from seismicity. In Southern Mexico (Chiapas) the modern volcanic front has a remote and tilted position, corresponding to $\sim$200 km depth from the slab surface. In Guatemala the position of the volcanic chain is in good agreement with the high mantle wedge temperature zone ($>$1300$^{o}$ C). On the other hand the remote position of the modern volcanic front in Chiapas does not correspond with maximum temperature zone in mantle wedge predicted by the steady state thermal models. The mantle wedge high temperature zone is situated just below the coast, corresponding to the position of the old Miocene volcanic arc in this region. Recent studies suggest that the Cocos plate just south of Tehuantepec Ridge (TR) has had a non-rigid behaviour, probably acting as a microplate. A tectonic model constrained by the structural and morphological features observed in the bathymetry of the TR and surrounding ocean floor in the Guatemala Basin, shows that the velocity of this microplate decreased down to $\sim$2 cm/yr between 14 to 9 Myr ago. This slowing down would have induced a strong disturbance in the mantle wedge flow beneath Chiapas, and therefore a migration of the volcanic front to the more inland present day position. Since the steady state thermal models are not able to explain the position of the modern volcanic arc in Chiapas, we suggest that a non steady-state thermal model needs to be applied in this region, taking into account a possible variation of the slab geometry and age through the time.
T13B-1365 1340h
Inferences about the geometry and distribution of magma supply to oceanic crust at slow and intermediate spreading rates from AMS studies in the Troodos Ophiolite, Cyprus.
We investigate magma flow directions in the sheeted dike complex of the Troodos Ophiolite, Cyprus, using anisotropy of magnetic susceptibility (AMS) data. Suites of oriented cores collected from the dike chilled margins were analyzed to determine the AMS orientations, and infer the magma flow directions. With newly collected data from the Kionia Peak area, our dataset includes 115 dikes sampled over an area of about 8 km along the ridge axis, and 2 km across the ridge axis. The inferred magma flow vectors are corrected for tectonic rotation and tilting, and present day elevation variations serve as estimators of individual site depths in the pseudostratigraphy along the paleo ridge axis. Analysis of flow directions as a function of depth indicates more vertical flow directions in the shallow section of the sheeted dike complex. Also, the horizontal component of magma velocity is found to correlate with the along-axis distance from the presumed segment center. The flow vectors rotate from more vertical directions by about 30 degrees toward horizontal over a distance of 8 km from the presumed magma source toward the bounding transform fault. Chemical analysis shows horizontal flow directions in the shallow section also correlate with low Ti magma chemistry. We use the observed magma flow orientations in the sheeted dike complex to constrain the along-axis distribution of magma supply from the mantle source to the crust. The two end-member models are: a predominantly vertical magma supply to the crust from the underlying mantle, and a highly focused supply to the center of the segment with subsequent lateral redistribution via blade-like dike injections in the shallow magma plumbing system. We show that the flow directions inferred from AMS are not a direct proxy for the mode of magma supply to the crust, as some vertical flow directions may be imprinted at the leading edge of a horizontally propagating dike, while horizontal flow directions may result from magma flowing along the level of neutral buoyancy in dikes that travel vertically from an underlying mantle source. A fraction of up to 1/3 of magma flow directions toward the presumed segment center indicate that not all dikes are emplaced laterally, and at least some melt is generated in the mantle underlying the segment ends.
T13B-1366 1340h
Axis-Parallel Depth and Gravity Profiles: Implications for Intra-Segment Crustal Thickness Variations and Near-Axis Tectonic Disruption
Axial depths within a ridge segment generally increase regularly along the axis away from an intrasegment minimum. These along-axis depth variations are widely assumed to be an isostatic response to along-axis intrasegment variations in crustal thickness, particularly in the Atlantic where MBA gravity lows are consistently found at the shallowest portion of segments, leading to the widespread use of gravity anomalies as a proxy for along-axis crustal thickness variations. However, Neumann and Forsyth [1993] have argued that the ridge axis cannot be locally compensated because the axial valley is a dynamic feature. To the extent that along-axis depth variations reflect crustal thickness variations, they should persist onto the ridge flanks as the isostatic expression of the changes in crustal thickness. Similarly MBA gravity anomalies resulting from crustal thickness variations should continue onto the ridge flanks. We examined ridge-parallel depth and gravity variations, both on-axis and at various distance from the axis, at the EPR, segments of the SEIR with axial highs, and at the MAR. The areas on the EPR and SEIR were chosen because they have large along-axis depth variations.. We picked the ridge axis location at 1' intervals and sampled the bathymetry and gravity grids at 1 km intervals along the flow lines through the ridge axis locations. The bathymetryprofiles were filtered using several techniques to remove the abyssal hill relief and were used to construct ridge parallel profiles at various distances from the axis. Along-axis depth variations at intermediate- and fast-spreading ridges do not extend out onto the ridge flank. Within segments where there is significant relief along the axis, axis-parallel profiles on the ridge flanks are either nearly flat or have a steady long-wavelength regional gradient across the segment. Profiles across the axis demonstrate that changes in axial depth result from variations in the form and relief of the axial morphology. This is particularly clear at intermediate spreading ridges where there is greater along-axis depth variation, but can also be seen at 13-15\deg N portion of the EPR where there is significant along-axis depth variation. At a few segments on the SEIR, a small (100 m) along-axis off-axis bathymetric high is found along the same flow line as the much larger axial bathymetric high. Along-axis depth and MBA gravity profiles at the slow-spreading MAR show a close correlation. However, he characteristic humped axial morphology does not extend off-axis. The interior of segments on the ridge flanks is shallower than the ridge discontinuities, but the shallowest off-axis depths are not in the center of the segment but rather near the ends. The center of the segment, the location of the shallowest axial depth, is several hundred meters deeper than near the segment ends. This appears to be the result of tectonic processes near the segment ends, which form segment-bounding ridges parallel to the axis. These ridges are most prominent on the "inside corners", but are usually present at both on both sides of the axis. At the 33\deg S "Plume" segment on the southern MAR, the off-axis MBA profiles have the same shape as the axial profile, although with slightly smaller amplitude. At shorter segments, the off-axis gravity anomalies do not show a coherent pattern. It thus appears that, at short segments, tectonic activity as the crust is transported up and out of the rift valley may disrupt the crustal thickness distribution established at the ridge axis. At longer segments, disruption by faulting is limited to the segment ends leaving the gravity pattern in the center of the segment intact.
T13B-1367 1340h
Poisson's Ratios in The Crust and Mantle North of The Kane Fracture Zone From 87-147 Ma
We present preliminary results of modeling mode-converted S-wave arrivals recorded by 14 Ocean Bottom Seismometers (OBS) deployed during the 2001 Far-Offset Active Source Imaging of the Mantle (FAIM) experiment in the Western Atlantic. Wide-angle seismic data were acquired along two transects: Line 1, an 800-km transect extending along a plate-kinematic flow line on 87-147 Ma lithosphere, and Line 2, orthogonal to Line 1. Previously, we demonstrated that crustal thickness along Line 1 decreases by ~1.5 km as paleo-spreading rate decreases from slow (13 mm/yr) to ultraslow (8 mm/yr). Results of FAIM mantle refraction and gravity modeling studies suggest that the equivalent of ~1.5 km thickness of gabrroic inclusions are distributed within the upper 30 km of the slow-spreading mantle, thus balancing the crustal thickness deficit and implying that spreading rate controls melt extraction rather than melt production. Here, we investigate whether some part of the lower crust consists of serpentinized peridotites. We identified PSP type mode-converted S-wave arrivals Sg, SmS, and Sn, and modeled those travel-times to constrain Poisson's ratio. The range of measured Poisson's ratio within the lower crust is 0.27-0.29 (-0.01) at all instrument locations, including crust created during ultraslow spreading, and the range of lower-crustal Vp is 6.8-7.0 km/s. Peridotite has a Poisson's ratio of 0.29-0.02 and peridotites containing ~30% serpentinite have Poisson's ratios of 0.29-0.31. A mixture of 30% serpentinite and peridotite with Vp of 8.2 km (the measured uppermost-mantle velocity beneath Line 1), would have Vp of 7.30 km/s at 2 kb. Thus, measured lower-crustal Vp and Vs are marginally consistent with a partially serpentinized peridotite (30-40% serpentinite) at the base of the crust, but are more consistent with a gabbroic (Vp ~6.8, Poisson's ratio 0.27-0.30) lower crust. This consistency, coupled with the observation that the sharpness of the crust/mantle boundary and the lower-crustal Poisson's ratio are both laterally uniform along Line 1 and Line 2, suggest that the crust along the FAIM transects is primarily melt-derived igneous crust.
T13B-1368 1340h
Seismic Structure of the Daito Ridges on the Northwestern End of the Philippine Sea Plate
The northwestern end of the Philippine Sea plate is characterized by several bathymetric highs, the Daito Ridges. We conducted two wide angle seismic and multi-channel seismic (MCS) profilings across the Oki-Daito Ridge, a member of the Daito Ridges in May_|July 2004, which provided the first images of the deep structure in this region and important clue to origin of these bathymetric highs. In this seismic experiment, the controlled sources were a tuned array of 36 airguns with a total volume of 8,040 inch$^{3}$ and two explosives. We shot the airgun array at an interval of 90 s, 100 m along EW 240 km and NS 660 km long lines. We used 251 ocean bottom seismographs (OBS) at an interval of 3-5 km, and a 6 km long, 480-channel hydrophone streamer as receivers. Both of the OBS and MCS data were of good quality. Clear PmP signals and reflections from the bottom of the lower crust were detected in many OBS records. These data were modeled by a tomographic inversion and two-dimensional ray tracing. Preliminary results shows the upper crust of the Oki-Daito Ridge has P wave velocity of 3.5-6.5 km/s with a thickness of about 7 km and the lower crust is 6.5-7.0 km/s with 20 km thickness. The total crustal thickness exceeds 25 km, which is thicker than the island arc curst obtained at the northern part of the Izu-Bonin arc. Several previous studies suggest that the origin of the Oki-Daito Ridge is a remnant island arc. The obtained crustal model, however, is similar to that of the southern Kerguelen Plateau (Indian Ocean), one of the largest oceanic plateaus in the world, that is a representative large igneous province (LIP), rather than that of the remnant island arc.