T32B-01 INVITED
Segmentation of Axial Depth Along the East Pacific Rise: An Isostatic Response to Magmatic Differentiation that Results From Skew of Mantle Upwelling
Within East Pacific Rise (EPR) ridge segments axial depth typically shoals away from a tectonic offset and toward a local minimum or intra-segment high. Axial depth minima are commonly associated with broad axial summits, intense seafloor hydrothermal activity and the eruption of MgO-rich lavas. A popular hypothesis for explaining these characteristics is that the supply of magma from the mantle to the crust is increased beneath intra- segment highs and reduced near segment offsets. If this conceptual model were correct, we would expect to observe either thickened crust beneath axial depth minima or evidence for the segment-scale, rise-parallel redistribution of magma. Motivated by geophysical studies that are inconsistent with either of these predictions, we propose an alternative model for the origin of rise parallel variations in seafloor depth along fast-spreading ridges. We attribute rise-parallel changes in on- and off-axis seafloor depth to changes in both the density and thickness of the crust, including the Mohorovicic transition zone. We use regional-scale bathymetry, gravity and seismic crustal thickness data from the EPR near 9°N to explore the magnitude of such density variations following two approaches: (1) we determine the variations in crustal density that fit the along-axis gravity, bathymetry and crustal thickness data, and (2) we isostatically estimate the variations in crustal density required to support seafloor depth anomalies. We find that variations in the density and thickness of oceanic crust are consistent with segment-scale, rise-parallel variations in seafloor depth and gravity near the EPR. As there is no evidence for either thicker crust or anomalously low density mantle localized beneath the axial depth minimum in this region, we conclude that crustal density plays a more important role in determining axial depth than heretofore thought possible. By this view, deeper on- and off-axis seafloor depth corresponds to more dense crust, whereas average crustal densities are expected to be less near axial depth minima or where off-axis seafloor depth shoals in the rise-parallel direction. Our modeling results, in combination with seismic observations and simple calculations, can be used to infer the origin of rise-parallel, segment-scale variations in crustal density. We attribute segment-scale variations in crustal density to magmatic differentiation that results from a skew between the axes of mantle upwelling and plate spreading. We use seismic imaging of crustal thickness and upper mantle structure combined with gravity data to provide novel constraints on the temporal evolution of mantle upwelling beneath the EPR and to infer that axial sites of intense volcanic and hydrothermal activity can persist for 105 to 106 years.
T32B-02 INVITED
Thoughts on the Significance of the East Pacific Rise "Undershoot" Seismic Results
Reported EPR seismic results between the Clipperton and Siqueiros transforms indicate that low velocity zones in the uppermost mantle (MLVZ's) correlate well with 3rd order tectonic/volcanic segments (Toomey et al., 2007). This adds to other geologic observations which correlate with segmentation: axial depth, cross-sectional area, crustal magnetization, inferred eruption temperature and fractionation (based on MgO), patterns of hydrothermal activity, inferred eruption effusion rates (based on lava morphology, ie, sheets/lobates vs pillows), and average lava age (based on sediment cover). These data are all consistent with the model that, in a time averaged sense, segment middles mark the surface expression of hotter, more buoyant upper mantle. Stacking several segments together to minimize effects of temporal variations at any one location, the data support the model that magma supply is enhanced near segment centers and is reduced or more episodic near segment ends. One of the 7 MLVZ's occurs up to 10 km off-axis near 9 20-9 35N; 6 occur beneath the EPR axial high. Prior segmentation models did not consider off-axis mantle upwelling effects on axial volcanic processes, due to lack of data, but it is unsurprising if this occasionally occurs. It may explain decreased hydrothermal activity along this one segment where the apparent reduction in axial melt supply could stem from lateral melt transport from off-axis. The new ridge segmentation model proposed by Toomey et al. (2007) is inspired by measurements showing seismic anisotropy canted 5-8 deg. ccw from the EPR topographic axis, which led Toomey et al. to "conclude that the skew of asthenospheric upwelling and transport governs segmentation of the East Pacific Rise". To produce this skewness, the asthenosphere must somehow lead the lithosphere in responding to a change in spreading direction, and this must occur virtually everywhere to explain the ubiquity of ridge segmentation on all spreading ridges. Assuming skewness is adequately resolved in their data, we offer a much simpler local explanation for it. Magnetic anomalies show that the EPR axis has rotated ccw by an amount similar to the observed skewness in response to a recent change in spreading direction. The long 1st order segments N of Clipperton and S of Siqueiros have rotated in response to this change, while the relatively short segment between these transform faults has not. Why? Since Siqueiros steps right and Clipperton steps left, there is transpression across Clipperton and opening across Siquieros. Thus rotation of the segment trapped between these transforms is impeded, while the asthenosphere below is free to rotate in response to the change in spreading direction. Also, this is the shortest 1st order segment bounded by large transforms along the Cocos-Pacific boundary, which may further impede adjustment (interestingly, 3rd and 4th order segments within the axial high ARE starting to show ccw adjustment). This explains skewness in the study area without appealing to a more astonishing universal skewness of mantle upwelling relative to the spreading axis as the cause of ridge segmentation; and, it allows ridge segmentation to exist where spreading directions have not changed. Neither the MLVZ ~10 km off-axis nor a mantle high conductivity zone further off-axis (Constable et al.) produce corresponding surficial evidence (e.g., off-axis volcanism, significant shallow depth anomaly, and/or bright backscatter images). Possibly the MLVZ has only recently arrived in this location. It also is possible that the MLVZ is a geologically ephemeral expression of asthenospheric flow representing instabilities in upwelling.
T32B-03
Detailed investigation of the overlapping spreading center and 9 degrees north on the East Pacific Rise: overview of the Medusa 2007 cruise activities
We undertook a 35-day cruise on the R/V Atlantis (AT15-17; March 24 to April 27, 2007) to investigate seafloor spreading processes along the northern East Pacific Rise. The bulk of our ship time (22 science days) was devoted to studying the 9 deg 03"N overlapping spreading center (OSC). At the OSC site, we surveyed 200 sq. km. with DSL-120A (obtaining 120 kHz backscatter and phase bathymetry, 200 kHz multibeam bathymetry, magnetics, CTD, and MAPR data). This was followed by 16 days of mapping and sampling with ROV Jason II (~7000 digital photographs, recovery of ~300 rock samples, sampling of hydrothermal vent waters and biota) and with the WHOI TowCam (~10,000 digital photographs, CTD and MAPR data). We discovered one high-temperature hydrothermal vent (named "Medusa Vent" to highlight the presence of unusual pink Stauromedusae) and an associated diffuse- flow hydrothermal field (biologic samples included riftia, tevnia, alvinella). We found no evidence of widespread recent eruptions, although loci of recent volcanism were identified on each limb of the OSC. Lavas recovered ranged from basalt to dacite. The data collected will be used to explore the linkages between geochemical, geological, and hydrothermal variations on the seafloor and the subsurface magma supply system at depth. Our preliminary results suggest a complex spatial relationship between magmatism, tectonism and hydrothermal activity observed at the surface and the positions and characteristics of the magma lenses at depth.
T32B-04
Tectono-magmatic Segmentation on the Reykjanes Peninsula
The Mid-Atlantic Ridge (MAR) is highly segmented as it passes through Iceland. Although spreading at 2 cm per year, the magma budget is higher than normal for a slow-spreading ridge.Segmentation of the plate boundary in Iceland occurs at scales similar to those described for the Reykjanes Ridge (RR) and elsewhere on the MAR. Second order segmentation can describe the 4 currently active rift zones, while volcanic systems within the zones are 3rd order segments.Each rift zone is structurally distinct due to differing obliquity with respect to the direction of absolute plate motion.Differences also result from variations in magma budget with distance from the Iceland mantle plume.Additional structural complexity is seen in the pattern of faulting at segment ends.Data from Reykjanes Peninsula (RP), the first on-shore segment in SW Iceland, shows third order segmentation in the distribution of post-glacial eruptive fissures and faults.These have previously been described as either 4 volcanic systems or 5 volcanic fissure swarms striking N40E, in a right-stepping en echelon pattern along a N75E- trending plate boundary zone.Each system is said to have its own center of magmatic production, high temperature geothermal field and set of faults and fissures. Closer examination reveals systematic differences from west to east in the geometry and distribution of post-glacial eruptive fissures within these pre-defined systems. The first on-shore system displays widely spaced en echelon fissures in a zone 12 km wide.Further eastward, overlapping, anastomosing fissures occur in a 4 km wide zone, followed by tightly spaced fissures in a 1 km wide zone.Overall, this progression reflects the evolution from axial volcanic ridges on the RR to more focused magmatic activity in well-developed central volcanoes towards the center of Iceland, and correlates with increases in crustal thickness and magma budget.Individual eruptive fissures are themselves highly segmented.At the 10 to 100 m scale they vary widely in strike and appear to have used pre-existing faults as pathways to the surface.The tip-to-tip strike of a fissure swarm can vary by more than 10 degrees from the average strike of fissure segments within a given swarm.This is attributed to a temporally and spatially inhomogeneous strain field on the RP rift segment.
T32B-05
Co-existing Overlapping-Spreading-Center and Ridge-Transform Geometry
We have studied the plate boundary geometry and recent spreading history in the Quebrada-Discovery-Gofar fracture zone system on the East Pacific Rise at 3-4 °S using multi-beam bathymetry, side-scan sonar, magnetic anomalies, gravity anomalies, and seismicity. Each of the transforms has multiple segments separated by intratransform spreading centers. Collectively, the system currently has 8 active, intratransform centers ranging in length from about 5 to 70 km. Over the last 1 Ma, the geometry has evolved, including the death of two transform segments and the birth of another one. The Discovery transform is the most unusual. Although it has a well-defined bathymetric signature and is seismically active, another transform trace appears to have been recently abandoned and there is no clear, fossil fracture zone outside the active transform region. The current transform is rotated a few degrees counter- clockwise from the direction of plate motion. Unlike most East Pacific Rise ridge segments, the spreading center at the eastern terminus of the Discovery has an axial valley, rather than an axial high. We suggest that part of the Nazca-Pacific separation is taken up by dike intrusion and formation of grabens that extend beyond the nominal end of the spreading center at the western terminus of the Discovery transform and cut across pre-existing seafloor-spreading fabric, creating a form of overlapping spreading center. Another region of recent deformation lies between the eastern and western overlappers near the center of the transform, bounding an older, trapped block or microplate that has been rotated counterclockwise about 45 degrees. The current geometry appears to be a hybrid between a simple ridge-transform boundary and an overlapping spreading center.
T32B-06
Thermal-Mechanical Behavior of Oceanic Transform Faults- Implications for Hydration of the Upper Oceanic Mantle
The presence of water at oceanic transform faults influences the thermal structure, rheology, and petrology of the upper mantle. Serpentinization at ridges and transforms plays an important role for the large-scale water budget of the mantle and eventual flux melting that is responsible for arc volcanism at convergent margins. The extent to which hydrous minerals (e.g., serpentine and talc) are incorporated into the upper mantle at oceanic transform faults is highly dependent on the thermal structure and stress state. Previous numerical modeling studies have suggested that the mantle beneath oceanic transform faults is anomalously cold, with depressed isotherms relative to a half-space cooling model [1,2,3]. However, recent models, that incorporate brittle rheology, show that transform faults may represent a region of enhanced mantle upwelling and elevated temperatures [4]. To investigate the thermal-mechanical behavior of oceanic transform faults, we utilize a 3D finite element model, assuming mantle convection, conduction, and steady-state incompressible mantle flow. Our model incorporates a non-linear viscous rheology with a visco-plastic approximation to simulate lithospheric brittle failure. The introduction of water into the lithosphere causes rheological changes with additional feedbacks on the thermal and rheologic structure such as enhanced conductive cooling and changes in frictional behavior. We incorporate the effects of these feedbacks, and our derived thermal structures are integrated with the estimated zone of permeable fluid flow to approximate the stability fields of hydrous phases in the upper mantle. Through examining a rage of parameters, including spreading rate, fault length, and the efficiency of hydrothermal circulation, we constrain the potential for transform faults to act as a source for mantle hydration, and estimate the amount of water that could be bound in hydrous phases as a result of brittle cracking at oceanic faults. 1. Furlong et al., The Nature and Tectonic Significance of Fault Zone Weakening 2001; 2. Phipps Morgan and Forsyth, JGR 1988; 3. Shen and Forsyth, JGR 1992; 4. Behn et al., Geology 2007
T32B-07
Mapping of Seismic Layer 2A at Intermediate-Spreading Crust Exposed Near Blanco Transform
We present results from a seismic survey over intermediate-spreading crust adjacent to the Blanco Transform using a source array optimized to image the layer 2A boundary. The program acquired ten MCS profiles parallel to and four MCS profiles perpendicular to the transform scarp edge covering a region approximately 10 km (ridge- parallel) x 40 km (ridge-perpendicular). The average two-way travel time to the layer 2A event in our dataset is 0.36 s, which is similar to values measured elsewhere along the Juan de Fuca Ridge and East Pacific Rise. An independent method to estimate two-way travel time to the layer 2A/2B boundary is to model layer 2B refractions recorded on the 6-km streamer; this method gives similar two-way travel time values. The average layer 2A interval velocity calculated from analyses on selected CDP supergathers is 2700 m/s; converting from two-way travel time using this interval velocity results in an average layer 2A thickness of 490 m. Our seismic study was conducted adjacent to a region where the lithologic boundary between the dikes and overlying extrusive section has been mapped by submersible dives over a lateral distance of tens of kilometers. A comparison of the lithologic and seismic boundaries indicates that the mapped depth to the top of the sheeted dikes (average depth 1150 m) is well below the imaged layer 2A/2B boundary. We argue (Christeson et al., Nature, 2007) that the primary control on the depth of the layer 2A/2B boundary is crack closure enhanced by hydrothermal alteration and sealing, and not the structural boundary between lavas and dikes. Mapping reveals regions in the ridge-parallel direction where the layer 2A thickness is consistently thin or thick; these regions do not correlate strongly with seafloor bathymetry. These patterns suggest a temporal nature to the processes controlling the layer 2A/2B boundary.