Fernando Martinez

Back-arc basin basalt systematics

The Mariana, east Scotia, Lau, and Manus back-arc basins (BABs) have spreading rates that vary from slow ( 100 mm/yr) and extension axes located from 10 to 400 km behind their island arcs. Axial lava compositions from these BABs indicate melting of mid-ocean ridge basalt (MORB)-like sources in proportion to the amount added of previously depleted, water-rich, arc-like components. The arc-like end-members are characterized by low Na, Ti and Fe, and by high H2O and Ba/La; the MORB-like end-members have the opposite traits. Comparisons between basins show that the least hydrous compositions follow global MORB systematics and an inverse correlation between Na8 and Fe8. This is interpreted as a positive correlation between the average degree and pressure of mantle melting that reflects regional variations in mantle potential temperatures (Lau/Manus hotter than Mariana/Scotia). This interpretation accords with numerical model predictions that faster subduction-induced advection will maintain a hotter mantle wedge. The primary compositional trends within each BAB (a positive correlation between Fe8, Na8 and Ti8, and their inverse correlation with H2O(8) and Ba/La) are controlled by variations in water content, melt extraction, and enrichments imposed by slab and mantle wedge processes. Systematic axial depth (as a proxy for crustal production) variations with distance from the island arc indicate that compositional controls on melting dominate over spreading rate. Hydrous fluxing enhances decompression melting, allowing depleted mantle sources just behind the island arc to melt extensively, producing shallow spreading axes. Flow of enriched mantle components around the ends of slabs may augment this process in transform-bounded back-arcs such as the east Scotia Basin. The re-circulation (by mantle wedge corner flow) to the spreading axes of mantle previously depleted by both arc and spreading melt extraction can explain the greater depths and thinner crust of the East Lau Spreading Center, Manus Southern Rifts, and Mariana Trough and the very depleted lavas of east Scotia segments E8/E9. The crust becomes mid-ocean ridge (MOR)-like where the spreading axes, further away from the island arc and subducted slab, entrain dominantly fertile mantle.

Mantle wedge control on back-arc crustal accretion.

At mid-ocean ridges, plate separation leads to upward advection and pressure-release partial melting of fertile mantle material; the melt is then extracted to the spreading centre and the residual depleted mantle flows horizontally away. In back-arc basins, the subducting slab is an important control on the pattern of mantle advection and melt extraction, as well as on compositional and fluid gradients. Modelling studies predict significant mantle wedge effects on back-arc spreading processes. Here we show that various spreading centres in the Lau back-arc basin exhibit enhanced, diminished or normal magma supply, which correlates with distance from the arc volcanic front but not with spreading rate. To explain this correlation we propose that depleted upper-mantle material, generated by melt extraction in the mantle wedge, is overturned and re-introduced beneath the back-arc basin by subduction-induced corner flow. The spreading centres experience enhanced melt delivery near the volcanic front, diminished melting within the overturned depleted mantle farther from the corner and normal melting conditions in undepleted mantle farther away. Our model explains fundamental differences in crustal accretion variables between back-arc and mid-ocean settings.

Backarc spreading, rifting, and microplate rotation, between transform faults in the Manus Basin

The Manus Basin in the eastern Bismarck Sea is a fastopening backarc basin behind the New Britain arc-trench system. Within the basin, motion between the Pacific and Bismarck plates about a pole located at 11° S, 145° E, occurs along three major leftlateral transform faults and a variety of extensional segments. We interpret SeaMARC II sidescan and other geophysical data to show that a Brunhes age plate reorganization created new extensional boundaries and a microplate between the NW-trending Willaumez, Djaul, and Weitin transforms. Two linked spreading segments formed in backarc basin crust between the Willaumez and Djaul transforms: the ESE-trending extensional transform zone (ETZ) in the west and the Manus spreading center (MSC) in the east. Positively magnetized crust on the MSC forms a wedge varying in width from 72 km at its southwest end to zero at its northeast tip, with corresponding Brunhes spreading rates varying from 92 mm/yr to zero. The MSC forms the northwestern boundary of the 100 km-scale Manus microplate and opens at 51°/m.y. about a pole near its apex at 3°02′S, 150°32′E. Opposite the MSC, bordering the arc margin of New Britain, the microplate is bound by a zone of broadly distributed strike slip motion, extension, and volcanism. Within this area, the Southern Rifts contain a series of grabens partially floored by lava flows. Left-lateral motion between the Pacific and Bismarck plates appears to drive the counterclockwise pivoting motion of the Manus microplate and the complementary wedge-like opening of the MSC and the Southern Rifts. The pivoting motion of the microplate has resulted in compressional areas along its NE and SW boundaries with the Pacific and Bismarck plates respectively. East of the microplate, between the Djaul and Weitin transforms and within the arc margin of New Ireland, another zone of broad extension referred to as the Southeast Rifts takes up opening in a pull-apart basin. There, en echelon volcanic ridges may be the precursors of spreading segments, but erupted lavas include calcalkaline volcanics. Kinematic modeling and marine geophysical observations indicate that the responses to similar amounts of extension in the eastern Manus Basin have varied as a function of the different types of pre-existing crust: arc crust tectonically stretched over a broad area whereas backarc crust underwent relatively little stretching before accommodating extension by seafloor spreading.

Why is the Challenger Deep so deep

Abstract Recent sidescan surveys of the deepest segment of the southern Mariana Trench in the western Pacific Ocean provide the first detailed images of this plate boundary, which includes the world’s greatest ocean depth, the Challenger Deep. The surveys reveal details of the southern Mariana plate margin, identify another deep rivaling the Challenger, and document widespread deformation of the overriding plate. Our data show a subduction-generated deep ocean trench, not the transform fault boundary suggested by other work [D.E. Karig, Geol. Soc. Am. Bull. 82 (1971) 323–344; D.E. Karig et al., J. Geophys. Res. 83 (1978) 1213–1226; D.E. Karig, B. Ranken, in: The Tectonic and Geologic Evolution of Southeast Asian Seas and Islands, Part 2, American Geophysical Union, Washington, DC, 1983, pp. 266–280; K. Fujioka et al., Geophys. Res. Lett. 19 (2002) 1–4]. We present the geological characteristics of the region, including seismic evidence for a tear in the subducting plate that has influenced the deformation of the overriding plate. The rollback caused by this tear creates greater depths along the southern part of the trench than elsewhere along its length.

A serpentinite-hosted ecosystem in the Southern Mariana Forearc

Several varieties of seafloor hydrothermal vents with widely varying fluid compositions and temperatures and vent communities occur in different tectonic settings. The discovery of the Lost City hydrothermal field in the Mid-Atlantic Ridge has stimulated interest in the role of serpentinization of peridotite in generating H2- and CH4-rich fluids and associated carbonate chimneys, as well as in the biological communities supported in highly reduced, alkaline environments. Abundant vesicomyid clam communities associated with a serpentinite-hosted hydrothermal vent system in the southern Mariana forearc were discovered during a DSV Shinkai 6500 dive in September 2010. We named this system the “Shinkai Seep Field (SSF).” The SSF appears to be a serpentinite-hosted ecosystem within a forearc (convergent margin) setting that is supported by fault-controlled fluid pathways connected to the decollement of the subducting slab. The discovery of the SSF supports the prediction that serpentinite-hosted vents may be widespread on the ocean floor. The discovery further indicates that these serpentinite-hosted low-temperature fluid vents can sustain high-biomass communities and has implications for the chemical budget of the oceans and the distribution of abyssal chemosynthetic life.

The East ridge system 28.5–32°S East Pacific rise: Implications for overlapping spreading center development

Abstract We report here on geophysical data from the East ridge and surrounding areas of the large-offset overlapping spreading centers (OSCs) that accommodate Pacific-Nazca opening between 28.5° and 32°S. The East ridge overlaps and is offset from the West ridge system by ∼ 120 km, forming the largest known pair of OSCs. In this area spreading rates reach the fastest currently active on Earth of ∼ 149 mm/yr. Although the East ridge is composed of 4 morphologically defined segments separated by 3 small OSCs, other geophysical characteristics imply 1 upwelling segment. All the active ridge segments in this area (including the propagating tips of the East and West ridges) form relative topographic highs with respect to the flanking sea floor; however, identified abandoned ridge tips form deeps. We interpret these data in terms of a model in which the propagating segment represents an overshoot of a surficial rupture of the brittle lithospheric layer, only partially coupled to the diverging flow of a more broadly distributed ductile deformation zone (DDZ), surrounding the steady-state ridges and crossing the offset between the OSCs. The topographic high of the propagating segment may be maintained primarily by along-axis melt migration from the stable spreading segments rather than by direct upwelling from beneath the ridge. The large overlapping ridges are inherently unstable and continued extension causes the overlapping axes to become offset from the stably spreading segments, cut off from the supply of melt, and replaced by a new set. The failed rift tips, for a period of time, overlie the broad DDZ and preferentially undergo continued extension and subsidence. The DDZ surrounding the ridge axes may be very broad in this area because of the very fast spreading rate, creating a very thin lithosphere susceptible to perturbation by relatively small mantle heterogeneities advected near the ridge axis, leading to the formation of the smaller OSCs observed.

Tectonic/volcanic segmentation and controls on hydrothermal venting along Earth's fastest seafloor spreading system, EPR 27°–32°S

[1]xa0We have collected 12 kHz SeaBeam bathymetry and 120 kHz DSL-120 side-scan sonar and bathymetry data to determine the tectonic and volcanic segmentation along the fastest spreading (∼150 km/Myr) part of the global mid-ocean ridge system, the southern East Pacific Rise between the Easter and Juan Fernandez microplates. This area is presently reorganizing by large-scale dueling rift propagation and possible protomicroplate tectonics. Fracture patterns observed in the side-scan data define structural segmentation scales along these ridge segments. These sometimes, but not always, correlate with linear volcanic systems defining segmentation in the SeaBeam data. Some of the subsegments behave cohesively, with in-phase tectonic activity, while fundamental discontinuities occur between other subsegments. We also collected hydrothermal plume data using sensors mounted on the DSL-120 instrument package, as well as CTDO tow-yos, to determine detailed structural and volcanic controls on the hydrothermal vent pattern observed along 600 km of the Pacific-Nazca axis. Here we report the first rigorous correlation between coregistered hydrothermal plume and high-resolution marine geophysical data on similar scales and over multisegment distances. Major plume concentrations were usually found where axial inflation was relatively high and fracture density was relatively low. These correlations suggest that hydrothermal venting is most active where the apparent magmatic budget is greatest, resulting in recent eruptions that have paved over the neovolcanic zone. Areas of voluminous acoustically dark young lava flows produced from recent fissure eruptions correlate with many of the major hydrothermal vent areas. Increased crustal permeability, as gauged by increased fracture density, does not enhance hydrothermal venting in this area. Axial summit troughs and graben are rare, probably because of frequent volcanic resurfacing in this superfast spreading environment, and are not good predictors of hydrothermal activity here. Many of the hydrothermal areas are found in inflated areas near the ends of segments, suggesting that abundant magma is being supplied to these areas.

Hydrothermal cooling along the Eastern Lau Spreading Center: No evidence for discharge beyond the neovolcanic zone

Heat transported from the mantle beneath spreading centers creates an astonishingly narrow ribbon of convective heat discharge at plate boundaries, as apparently demonstrated by exhaustive exploration for hydrothermal discharge sites over the last three decades. Recent observations and models are now challenging this assumption of exclusively axis-centric high-temperature venting. One example is the proposal that intense cooling along the vertical boundaries of a broad low-velocity volume (LVV) of hot crust could generate high-temperature fluids several kilometers off axis. To test the hypothesis that substantial hydrothermal discharge might occur beyond the LVV, we conducted a dense survey grid of the ridge and surrounding seafloor (up to ±5 km) along 175 km of the Eastern Lau Spreading Center and Valu Fa Ridge (∼1800 km of track line). Our sampling array extended from ∼50 to 400 m above bottom and included light-scattering, oxidation-reduction potential, and hydrographic sensors attached to the tow line and beneath the IMI120 sonar mapping system. The surveys successfully mapped plumes from several vent fields in the neovolcanic zone (∼±1.5 km about the axis) but did not detect evidence of significant discharge anywhere farther off-axis. At a few locations on the Valu Fa Ridge, however, we did record oxidation-reduction potential anomalies with hydrographic density signatures that imply low-temperature hydrothermal sources on the axial flank. Although these sites are hundreds of meters deeper than the adjacent crest, they are above, not beyond, the previously mapped LVV. Our results thus do not support a simple picture of high-temperature fluids ascending undiluted through the crust to the seafloor several kilometers off-axis. However, we cannot exclude the possibilities that the largely unmapped LVV is narrower here than seen on other ridges, that hydrothermal fluids formed beyond the LVV are channeled to the axis, or that discharge beyond the neovolcanic zone occurs only as dispersed, very low-temperature fluids. Our observations do demonstrate that high-temperature discharge predominantly exits the seafloor within a narrow (∼±1.5 km) axial ribbon, regardless of the presence or absence of an axial magma chamber.

Pito Rift: How a large-offset rift propagates

The Pito Rift area is the site of actively deforming oceanic lithosphere that has been primarily under extension for at least the past million years, based on kinematic reconstructions. The major morphologic features, Pito Deep and Pito Seamount, are aligned toward the Euler pole for relative motion between the Easter and Nazca plates. SeaMARC II side-scan and bathymetry data indicate that there are two general modes of faulting currently active in the Pito Rift area. One is associated with incipient rifting of old (∼3 Ma) Nazca lithosphere by large NW-SE normal faults, and the other is associated with a broad area of right-lateral transform shear between the Nazca and Easter plates. This transform shear is distributed over a broad region because of the northward growth of the East Rift and parallel tectonic rifting within the Pito Rift area. The majority of the Pito Rift area is composed of preexisting blocks of Nazca plate that are back-tilted away from Pito Deep and strike perpendicular to present and previous relative plate motions. This observation suggests that block-faulting and back-tilting are the primary mechanisms responsible for the distributed lithospheric extension, in agreement with gravity and magnetic analyses (Martinez et al., this issue).The only recent volcanic flows observed in side-scan data are from the Pito Seamount area and to the outside of the outer pseudofault of the East Rift. The significance of the young flows near the outer pseudofault is not understood. We interpret the flows extending northwest from the Pito Seamount as representing a newly formed seafloor spreading axis within the Pito Rift area. Gravity and magnetic analyses (Martinez et al., this issue) together with SeaMARC II bathymetry and side-scan data support this interpretation.Based on the tectonic evolution of the Easter microplate, we propose an evolutionary model for the formation of the Pito Rift area, where new ‘tectonic’ grabens form immediately west of the previous graben and with slightly more counterclockwise orientation. The duration and history of tectonic activity for each graben are not well constrained.

Magmatic and tectonic extension at the Chile Ridge: Evidence for mantle controls on ridge segmentation

We use data from an extensive multibeam bathymetry survey of the Chile Ridge to study tectonomagmatic processes at the ridge axis. Specifically, we investigate how abyssal hills evolve from axial faults, how variations in magmatic extension influence morphology and faulting along the spreading axis, and how these variations correlate with ridge segmentation. The bathymetry data are used to estimate the fraction of plate separation accommodated by normal faulting, and the remaining fraction of extension, M, is attributed primarily to magmatic accretion. Results show that M ranges from 0.85 to 0.96, systematically increasing from first-order and second-order ridge segment offsets toward segment centers as the depth of ridge axis shoals relative to the flanking highs of the axial valley. Fault spacing, however, does not correlate with ridge geometry, morphology, or M along the Chile Ridge, which suggests the observed increase in tectonic strain toward segment ends is achieved through increased slip on approximately equally spaced faults. Variations in M along the segments follow variations in petrologic indicators of mantle melt fraction, both showing a preferred length scale of 50u2009±u200920 km that persists even along much longer ridge segments. In comparison, mean M and axial relief fail to show significant correlations with distance offsetting the segments. These two findings suggest a form of magmatic segmentation that is partially decoupled from the geometry of the plate boundary. We hypothesize this magmatic segmentation arises from cells of buoyantly upwelling mantle that influence tectonic segmentation from the mantle, up.