Median Valley Research Articles

An evolving ridge system around the Indian Ocean triple junction

A 2°×2° map of spreading centres and fracture zones surrounding the Indian Ocean RRR triple junction, at 25.5°S, 70°E, is described from a data set of GLORIA side-scan sonar images, bathymetry, magnetic and gravity anomalies. The GLORIA images show a pervasive fabric due to linear abyssal hills oriented parallel to the two medium-spreading ridges (the Central Indian Ridge (CIR) and Southeast Indian Ridge (SEIR)). A cuvature of the fabric occurs along fracture zones, which are also located by lows in the bathymetry and gravity data and by offsets between magnetic anomalies. The magnetic anomalies also record periods of asymmetric spreading marking the development of the fracture zones, including the birth, at anomaly 2A, of a short fracture zone 50 km north of the triple junction on the CIR, and its death near the time of the Jaramillo anomaly. In some localities, a fine-scale fabric corresponds to a coarser fabric on the opposite flank of the CIR, possibly indicating a persistent asymmetry in the faulting at the median valley walls if the fabric has a tectonic and not a volcanic origin. A plate velocity analysis of the triple junction shows that both the CIR and Southwest Indian Ridge (SWIR) are propagating obliquely; the CIR appears to form an oblique trend by segmenting into a series of almost normally-oriented segments separated by short-offset fracture zones. For the last 4 m.y., the abyssal hill lineations indicate that the CIR segment immediately north of the triple junction has been spreading with an average 10° obliquity. The present small 5 km offset of the centres of the CIR and SEIR median valleys (Munschy and Schlich, 1989) is shown to be the result of this obliquity and a 30% spreading asymmetry between anomaly 2 and the Jaramillo on the CIR segment immediately north of the triple junction.

Isostatic compensation of tectonic features of the Mid‐Atlantic Ridge: 25–27°30′S

We investigate the isostatic compensation of tectonic features along the Mid‐Atlantic Ridge 25°–27°30′S through a detailed examination of the relationship between bathymetry and gravity anomalies The study area includes three ridge segments, spreading at an average full rate of 35 mm/yr, their flanks out to about 6 m.y. old crust, and the intervening Rio Grande and Moore fracture zones. In a three‐dimensional analysis of gravity and Sea Beam bathymetry data, we focus on crustal thickness variations and mantle density anomalies by removing from the observed fields the predicted contribution of simple crustal and mantle models. Positive residual gravity anomalies over the northern wall of the Rio Grande fracture zone indicate that in spots the crust is 2–3 km thinner than average there, in contrast to the Moore fracture zone where little thinning is observed. The greater than average depths of the fracture zones as a whole are not locally compensated by thin crust, but the deepest basins within the active transform parts of the fracture zones are partially compensated by thin crust or cooler mantle. A narrow linear ridge that crosses one of the inactive branches of the Moore fracture zone had been suggested to be the product of an episode of excess volcanism but is found to be underlain by thinner, not thicker crust. There is no indication of thicker crust beneath topographic highs at the inside corners of ridge‐transform intersections. Presumably, these highs are dynamically maintained as must be the median valleys characteristic of most the length of the ridge segments judging from the absence of axial residual gravity anomalies. There is a residual gravity low associated with an unusually shallow 15‐ to 20‐km section of the ridge segment between the Rio Grande and Moore offsets where the median valley virtually disappears. Our analysis suggests that while part of the anomalous elevation of this section of the ridge axis is attributable to excess volcanism and a thicker crust, much of the elevation contrast is simply caused by the diminution of the dynamic mechanism responsible for median valley formation. Two‐dimensional Fourier transforms of the gravity and bathymetry fields show that seafloor topography is strongly aligned parallel to either ridges or transforms while density anomalies are more randomly oriented. This implies that the tectonic processes that control the gross ridge‐transform topography are not the dominant control on the magmatic processes that determine the upper mantle and crustal density structure. Statistical analysis of the coherence between the gravity and bathymetry fields indicates that the average, effective, elastic plate thickness in the study area is about 6 km.

Emplacement of deep crustal and mantle rocks on the west median valley wall of the MARK area (MAR, 23°N)

Exposures of deep crustal and mantle rocks within the axial rift valley characterize accretion processes in the slow-spreading Mid-Atlantic Ridge. Nine dives of the submersible Nautile explored the western axial valley wall in the northern cell of the MARK area (Mid-Atlantic Ridge/Kane fracture zone), where peridotite and gabbro outcrops had been previously reported, in order to constrain the structure and determine emplacement mechanisms. The ridge/transform intersection massif on the western wall shows a section of gabbros from 6000 to 2500 mbsl, locally overlain by metabasalts and metadolerites, capped by slightly weathered basalts. The morphology is controlled by ridge-parallel faults, dipping moderately (40–65°) to the east. Transform-parallel scarps, present in the northernmost dives, become rare toward the south. Brittle deformation, along moderately dipping fault scarps, produced a dense microcrack network filled with greenschist facies minerals (chlorite, actinolite, epidote, quartz), locally overprinting high-temperature ductile deformation fabrics. The small hill located at 23°20′ in the western wall shows good exposures of serpentinized peridotite between 3700 and 3100 mbsl. Above 3100 mbsl, the summit of the hill is composed of pillow-basalts and sediments. The peridotite outerops display a strong schistosity dipping 20–40° to the east, parallel to striated normal fault planes. Some steeper east-facing fault scarps truncate the lower-dipping fault surfaces. The serpentinites, deriving from a harzburgitic protolith, are cut by rare dikelets of highly differentiated gabbro. These new data, combined with previous results of Alvin dives, are used to draw a generalized geological map of the western axial valley wall. This map suggests a variation in the thickness of crustal units and composition along strike in the northern cell of the MARK area: the gabbro body disappears toward the south where basalts appear to directly overlie the mantle peridotites which are cut by isolated gabbroic dikelets. Low-angle stretching affecting this heterogeneous lithosphere is probably responsible for the exposure of gabbros in the intersection massif and of peridotites further south.

The Mid-atlantic Ridge (31°S–34°30′S): Temporal and spatial variations of accretionary processes

The ridge located between 31° S and 34°30′S is spreading at a rate of 35 mm yr−1, a transitional velocity between the very slow (≤20 mm yr−1) opening rates of the North Atlantic and Southwest Indian Oceans, and the intermediate rates (60 mm yr−1) of the northern limb of the East Pacific Rise, and the Galapagos and Juan de Fuca Ridges. A synthesis of multi-narrow beam, magnetics and gravity data document that in this area the ridge represents a dynamically evolving system. Here the ridge is partitioned into an ensemble of six distinct segments of variable lengths (12 to 100 km) by two transform faults (first-order discontinuities) and three small offset (< 30 km) discontinuities (second-order discontinuities) that behave non-rigidly creating complex and heterogeneous morphotectonic patterns that are not parallel to flow lines. The offset magnitudes of both the first and second-order discontinuities change in response to differential asymmetric spreading. In addition, along the fossil trace of second-order discontinuities, the lengths of abyssal hills located to either side of a discordant zone are observed to lengthen and shorten creating a saw-toothed pattern. Although the spreading rate remains the same along the length of the ridge studied, the morphology of the spreading segments varies from a deep median valley with characteristics analogous to the rift segments of the North Atlantic to a gently rifted axial bulge that is indistinguishable from the shape and relief of the intermediate rate spreading centers of the East Pacific Rise (i.e., 21°N). Like other carefully surveyed ridge segments at slow and fast rates of accretion, the along-axis profiles of each ridge segment are distinctly convex upwards, and exhibit along-strike changes in relief of 500m to 1500 between the shallowest portion of the segment (approximate center) and the segment ends. Such spatial variations create marked along-axis changes in the morphology and relief of each segment. A relatively low mantle Bouguer anomaly is known to be associated with the ridge segment characterized by a gently rifted axial bulge and is interpreted to indicate the presence of focused mantle upwelling (Kuo and Forsyth, 1988). Moreover, the terrain at the ends of each segment are known to be highly magnetized compared to the centers of each segment (Carbotte et al, 1990). Taken together, these data clearly establish that these profound spatial variations in ridge segment properties between adjoining segments, and along and across each segment, indicate that the upper mantle processes responsible for the formation of this contrasting architecture are not solely related to passive upwelling of the asthenosphere beneath the ridge axis. Rather, there must be differences in the thermal and mechanical structure of the crust and upper mantle between and along the ridge segments to explain these spatial variations in axial topography, crustal structure and magnetization. These results are consistent with the results of investigations from other parts of the ridge and suggest that the emplacement of magma is highly focused along segments and positioned beneath the depth minimum of a given segment. The profound differences between segments indicate that the processes governing the behavior of upwelling mantle are decoupled and the variations in the patterns of axis flanking morphology and rate of accretion indicate that processes controlling upwelling and melt production vary markedly in time as well. At this spreading rate and in this area, the accretionary processes are clearly three-dimensional. In addition, the morphology of a ridge segment is not governed so much by opening rate as by the thermal structure of the mantle which underlies the segment.

Block-tilting, transfer faults, and structural control of magmatic and hydrothermal processesin the TAG area, Mid-Atlantic Ridge 26°N

Seven Alvin dives (14 km total) and numerous deep-towed camera traverses using ANGUS and NOAA camera systems provide dense coverage of a 12-km2 portion of the eastern wall of the Mid-Atlantic Ridge in the TAG area (26°N lat.). These data, in conjunction with recent Soviet Mir submersible data, provide important constraints on the tectonic, magmatic, and hydrothermal history of this spreading center segment. Active hydrothermal venting occurs near the junction of the median valley floor and eastern median valley wall and appears to be tectonically controlled by the intersection of major fault zones. An east-west fault-line scarp interpreted as an accommodation zone intersects escarpments associated with 020°-trending (ridge-parallel) normal faults that bound the median valley floor. The accommodation zone permits differential extension and rotation between major crustal blocks to the north and south. On the basis of the distribution of tilted chalk beds and geochemical anomalies in sediments, this fault zone has been intermittently active for at least 5x104 yr. The accommodation zone has apparently provided a conduit of high permeability oriented at a high angle to the ridge axis. Observations and samples from areas surrounding active and inactive vent sites provide evidence for three distinct episodes for hydrothermal outflow driven by separate magmatic events. The geometry of this active system may have implications for the location of hydrothermal systems in active spreading regimes and for massive sulfide exploration in ophiolite terranes.

Hundreds of small volcanoes on the median valley floor of the Mid-Atlantic Ridge at 24–30° N

WE have identified 481 seamounts from ∼6,000 km2 of Sea Beam swaths of the median valley floor of the Mid-Atlantic Ridge (MAR) between 24 and 30° N. We find that seamount abundances are an order of magnitude larger (on average, 80 per 1,000 km2 with relief above 50 m) than estimated in previous studies along mid-ocean ridges in the Pacific1 and South Atlantic2, suggesting that at this section of the MAR, seamount volcanism plays an important part in the formation of new oceanic lithosphere. The neovolcanic ridge, often observed on the inner valley floor3, appears to be constructed totally of individual coalesced seamounts, so that our counts represent a minimum estimate. The frequency distribution of summit heights of the MAR seamounts is consistent with the exponential model describing the heights of off-axis seamounts in the eastern Pacific4–6, but with different parameters. Such a distribution of heights may be a fundamental aspect of seamount volcanism, related to lithospheric properties and/or magmatic plumbing. Seamount-dominated axial volcanism must create a crustal structure very different from classic fissure-fed models.

Magnetic properties and opaque mineralogy of rocks from selected seafloor hydrothermal sites at oceanic ridges

Magnetic properties (natural remanent magnetization (NRM), susceptibility (χ), Curie point temperature (Tc), saturation isothermal remanent magnetization (IRMs), and Köenigsberger ratio (Q)) and opaque mineralogy were determined for basalts, diabases, gabbros, peridotites, and serpentinites collected by dredging and submersible from the rift valley at five hydrothermal sites (15°N, 17°N, 23°N (the Snake Pit hydrothermal field), 26°N (the TAG hydrothermal field) on the Mid‐Atlantic Ridge, and 42°N (the Sea Cliff hydrothermal field) on the northern Gorda Ridge). The magnetomineralogy is interpreted in terms of deuteric, hydrothermal, metasomatic, and ambient seawater alteration. Evidence for unequivocal magnetic mineral modification by hydrothermal action is present only in a small percentage of extrusive basalts but is pervasive in diabases, gabbros, and ultramafic rocks. This lithologiec distribution suggests a progressive increase in alteration intensity with depth in the oceanic crust and upper mantle from minor low‐temperature alteration in young surface basalts to more pervasive high–temperature alteration in diabases, gabbros, and peridotites. The studies reveal distinct correlations between magnetization intensity, thermomagnetic behavior, and magnetic mineralogy, grain size, style, and intensity of alteration and rock type. Basalts have the highest NRM intensities and the lowest Curie points (105°–294°C); basalts with NRM/IRMs ratios less than 10−2 may have been remagnetized and have NRM intensities and Q values much lower (one–third and one–half, respectively) than basalts with NRM/IRMs ratios of 10−2 or greater. In the oceanic lithospheric suite examined, remanent (Q &gt; 1) magnetization is present in basalts and induced magnetization (Q &lt; 1) occurs in metabasalts and the other rock types. Layer 2A basalts are the source of median valley magnetic anomalies. The magnetic source may shift from the surface to deeper horizons with progressive seafloor aging and enhanced deuteric oxidation in layer 2B dikes and layer 3 gabbros and the formation of magnetite in serpentinized peridotites. This shift is influenced by fluctuations in the Curie isotherm related to the duration of active magma chambers and convective hydrothermal cells. The magnetic and mineralogie properties of oceanic rocks determined show the effects of a wide range of alteration processes that are variously related to depth in oceanic lithosphere, magmatic cooling history, and hydrothermal circulation.

Uplift and a possible moho offset across the Dead Sea transform

The East Africa rifts and the Red Sea spreading center are characterized by uplifted shoulders and sunken median valleys. The Dead Sea transform, on the northern extension of the Arabian-African plate boundary, has a similar morphological character despite its well-documented strike-slip motion. To understand the “rift” morphology and the crustal structure beneath the transform we compiled a 320 km long gravity and topography profile perpendicular to the Dead Sea transform. The gravity field and the topography in the region surrounding the profile are generally parallel to the trend of the Dead Sea transform, justifying the two-dimensional approximation of the analysis. The gravity profile was modeled using constraints from seismic refraction, borehole and surface geology data. The observed gravity anomaly can be explained by the juxtaposition of two different sedimentary and crustal sections which have been offset by a 105 km left-lateral displacement across the transform boundary, suggesting a step in the depth to Moho. The existence of a significant density anomaly under the median valley is not required by the model. The current elevation of the transform shoulders appears not to be compensated locally. Assuming a state of local isostasy prior to the development of the transform, the magnitude of the uplift in the vicinity of the profile is estimated at 700–900 m, with a half width of 100–125 km. This topography (uplift) can be fitted equally well by models which assume either dynamic support or regional compensation. It is suggested that if the source of the uplift is thermal, it is located within the upper mantle, since there is no requirement for a shallow density anomaly under the rift itself or in the crust nearby.

Earthquake swarms on the Mid‐Atlantic Ridge: Products of magmatism or extensional tectonics?

By analogy with the microearthquake swarms which accompany terrestrial volcanic eruptions, teleseismically observed earthquake swarms along mid‐ocean ridges are commonly thought to be indicators of volcanism, particularly dike emplacement. We test this assumption by investigating the spatial and temporal patterns and other characteristics of earthquakes in 34 swarms on the Mid‐Atlantic Ridge. Resolution of the relative locations of swarm events is improved by the use of a multiple‐event relocation technique. Most swarms extend 5 to 15 km across the ridge axis and 5 to 20 km along it. None of the swarms displays the large length‐to‐width ratio and temporal migration of epicenters along axis which are characteristic of earthquake swarms in the well‐studied rift zones of Hawaii and Iceland. The relative locations, magnitudes, and source mechanisms of swarm events are most easily explained if teleseismically observable swarms are expressions of extension across a substantial portion of the median valley and perhaps into the rift mountains. The possibility that some of these swarms include events directly associated with magmatic intrusion cannot be excluded, however. Like their terrestrial counterparts, purely volcanic earthquakes on the mid‐ocean ridges probably occur dominantly at magnitudes below the teleseismic detection threshold of global seismic networks (mb &lt; 4.5). The largest normal‐faulting earthquakes on mid‐ocean ridges and many of the teleseismically observed swarms likely occur on the faults marking the inner rift wall, where the crust first experiences significant faulting and is broken into numerous closely‐spaced normal faults. We speculate that the high degree of order in the dimensions, spacing, throws, and mechanical properties of these faults, combined with stress concentrations arising naturally from the mechanical behavior of the lithosphere at rifted, slow‐spreading ridges, may account for the common occurrence of earthquake swarms.

Magnetic anomalies across the southern Central Indian Ridge: evidence for a new transform fault

Bathymetric and magnetic studies along seven northeast-southwest profiles across the Central Indian Ridge segment between latitudes 18 and 22°S and longitudes 66 and 69°E reveal a northwest-southeast trending median valley, associated with a broad and high-amplitude magnetic anomaly (the central anomaly) along the ridge crest. A series of linear magnetic anomalies parallel to the ridge are identified as sea-floor spreading lineations 2,2A,3,3A and 4. A half spreading rate of 2.2 cm y −1 is estimated for the last 10 Ma. The ridge jump between the anomalies 2-2A (approx. 2.5 Ma) and a new left lateral transform fault offsetting the ridge axis about 50 km are indicated to explain the magnetic pattern.

Axial topographic relief associated with ridge-transform intersections

The axial depths of mid-ocean ridges are consistently observed to increase along-strike from the central part of a ridge segment to its intersection with a transform fault. Variations in the magnitude of this axial relief probably reflect variations in the magmatic and tectonic processes that characterize plate accretion and seafloor spreading at each individual segment. Constrained by observations from 41 ridge-transform intersections, we investigate the relative effect on axial depths of transform offset, segment length, regional depth, and spreading rate. We first solve for a model of ridge segment relief as a function of the age offset of the transform and the segment length, both scaled by spreading rate. Through a series of regression analyses covering a range of spreading rate scalings, we determine a model which predicts segment relief at both fast and slow spreading centers to within 240 m standard deviation. We find that half-segment scale deepening is more than a direct thermal effect of the juxtaposition of a cold lithospheric edge and the end of a ridge at a transform fault. A second inversion problem is solved to determine the parameters that contribute to the absolute depth of the ridge axis. We can show that the absolute depth of the shallowest section of a ridge segment is strongly dependent on regional depth and that this dependence is more complex at slow spreading ridges reflecting the common, but not ubiquitous, existence of a median valley there. These results, in combination with the segment relief analysis, clearly indicate that the axial depth at a ridge-transform intersection is a function of both long wavelength processes, such as may control hot spot distribution, and local tectonic factors including, but not limited to, transform offset.

Mechanisms of lithospheric extension at mid-ocean ridges

We examine the extensional deformation of oceanic plates at mid-ocean ridges, especially within an axial yield zone where pervasive faulting occurs. Thermal models of ridges are developed which include the effects of lithospheric thickening on the mantle flow, the heat of magmatic crustal accretion at the ridge axis, and the hydrothermal cooling due to seawater circulation in the crust. When hydrothermal circulation occurs, the brittle lithospheric plate at slow-spreading ridges could be as thick as 8–9 km, thicker than the crust; while the plate at fast-spreading ridges is only 1–2 km. For a typical slow-spreading ridge, several kilometres of plate thickening are expected within a distance of 15 km from the ridge axis. When subjected to the extensional force due to horizontal stretching, shear failure by normal faulting will occur pervasively in an axial zone, where the lithospheric plate is the thinnest. Adopting perfectly plastic rheology as a continuum description of deformation on the distributed faults, we obtain approximate solutions for the stress distribution in the yield zone. Within this yield zone, sea-floor topography increases significantly away from the ridge axis so that the resulting gravity sliding force balances the differential horizontal extensional force due to the thickening of the lithospheric plate. Basal stresses induced by the viscously deforming asthenosphere could significantly influence the stresses inside the lithospheric plate only if the mantle viscosity beneath the ridge is on the order of 1020 Pa s, significantly higher than generally accepted values of 1018 Pa s. Model calculations reveal that although the sea floor topography at the Mid-Atlantic Ridge at 13–15 °N is regionally compensated, it is locally supported by stresses in the lithospheric plate. The deviation of the lithospheric plate from its thermal isostatic equilibrium position can be explained by necking due to plastic stretching in the axial yield zone and elastic deflection of strong plates outside the yield zone. The best fit models require the yield zones to have a half width of 10–15 km. We find systematic variations in the gravity and topography of the East Pacific Rise, which indicate strong influence of plate spreading rate on the ridge thermal and mechanical structure. At the 16–17 °N area, where the half-spreading rate is 4.3 cm yr−1, a prominent axial topographic high and an axial mantle Bouguer gravity low exist, implying a crustal or sub-crustal low density body. The gravity low disappears at the 20–21 °N area, where the half-spreading rate is 3.6 cm yr1. As the plate spreading rate decreases from 4.3 cm yr1 at the 16–17 °N area to 2.7 cm yr1 at the 22–23 °N area, axial ridge topography changes from higher than the thermal isostatic equilibrium position to lower than the isostatic position. The low axial topography at the 22–23 °N area can be explained by the existence of a low amplitude median valley.

Centroid Depths of Mid‐Ocean Ridge Earthquakes: Dependence on Spreading Rate

From inversions of teleseismic P and SH waveforms we have determined the source parameters of 50 large earthquakes that occurred during 1962–1983 on slowly spreading mid‐ocean ridges. All events are characterized by predominantly normal faulting on planes that dip at approximately 45° and strike parallel to the local trend of the ridge axis. Centroid depths range from 1 to 6 km beneath the seafloor. The P waves from these earthquakes show strong water column reverberations, suggesting that fault rupture extended to the seafloor. Under the assumption that the centroid depth marks the mean depth of fault slip, earthquake faulting extended to a depth of 2–10 km for these earthquakes. The water depths inferred from the predominant periods of the water column reverberations constrain the epicenters of these earthquakes to lie within the relatively deep inner floor of the median valley. The maximum centroid depths of ridge crest earthquakes decrease with increasing spreading rate, and the maximum seismic moment may also decrease with increasing spreading rate. These results indicate a general decrease with spreading rate in maximum thickness of the mechanically strong layer beneath the ridge axis region. The concentration of large earthquakes within the deepest parts of the median valley and the depth extent of seismic faulting support lithospheric necking models for the origin of the median valley.

Microearthquakes beneath Median Valley of Mid‐Atlantic Ridge near 23°N: Tomography and tectonics

Data from a microearthquake experiment in the median valley of the Mid‐Atlantic Ridge near 23°N in 1982 are used to measure earthquake source parameters, to determine the laterally heterogeneous seismic velocity structure across the inner floor, and to develop a kinematic tectonic model for this portion of the median valley. Fifty‐three microearthquakes occurred over a 10‐day period beneath the median valley inner floor and eastern rift mountains. Twenty of 23 well‐located inner floor epicenters define a line of activity, about 17 km long, having a strike of N25°E and located near an along‐axis depression some 300–400 m deeper than surrounding regions. Earthquakes with well‐resolved hypocenters generally have focal depths of 4–8 km beneath the seafloor of both the inner floor and the rift mountains; the hypocentral locations are robust with respect to plausible lateral variations in seismic velocity structure. Composite fault plane solutions for inner floor events indicate normal faulting on planes dipping at angles near 45°. Normal faulting mechanisms, although poorly constrained, are also indicated for the rift mountain microearthquakes. Seismic moments, approximate fault dimensions, and average stress drops for the largest events recorded are 1019–1020 dyn cm, 200–400 m, and 1–70 bars, respectively. A twodimensional tomographic inversion of P wave travel time residuals from microearthquakes and local shots indicates a well‐resolved lateral heterogeneity in crustal velocity structure across the median valley inner floor. P wave velocities at 1–5 km depth within a zone less than 10 km wide beneath the central inner floor are lower by several percent than in surrounding regions. The most likely explanation for the low velocities is that the region is the site of the most recent local magmatic injection and remains pervasively fractured as a result of rapid hydrothermal quenching of the newly emplaced crustal column. By this view, the seismic velocity structure at the ridge axis evolves, probably by the sealing of cracks and pores, within the first few hundred thousand years of crustal accretion. Consideration of the detailed Sea Beam bathymetry in this region of the inner floor, the characteristics of large earthquakes that the region has experienced during the past 25 years, and the results of the microearthquake and tomography analysis suggests that this section of the median valley has been undergoing continued horizontal extension and block faulting without significant crustal injection of magma for at least the past 104 years.

Horizontal extent of anomalously thin crust near the Vema Fracture Zone from the three‐dimensional analysis of gravity anomalies

Recent seismic refraction and reflection experiments have demonstrated that parts of oceanic fracture zones are characterized by very thin crust, with total crustal thickness less than half that of normal oceanic crust. The horizontal extent of this anomalous crust is not well known. To map the lateral distribution of thin crust in a major fracture zone, we have constructed free‐air anomaly, bathymetry, and sediment isopach maps of a 220 by 280 km region surrounding the western intersection of the Vema fracture zone with the Mid‐Allantic Ridge. We isolate those components of the gravity field that are associated with variations in crustal thickness by first subtracting the attraction of the water‐crust interface, the sediment layer, and the crust‐mantle interface, assuming constant crustal thickness and density. These calculations are performed using a fully three‐dimensional model. The remaining mantle Bouguer anomaly is dominated by the effects of the cooling lithosphere. Variations in density of the crust and upper mantle associated with the cooling of the plate with increasing age of the seafloor are predicted for a thermal model that incorporates advection and the three‐dimensional conduction of heat. The attraction of this three‐dimensional density structure is subtracted from the mantle Bouguer anomaly to generate a map of the residual gravity anomaly which should directly reflect variations in crustal thickness or density. The residual gravity anomaly map indicates that there is thin crust along much of the length of the Vema fracture zone. Thin crust is concentrated primarily beneath the walls of the fracture zone, rather than beneath the center of the fracture zone valley, an interpretation confirmed by seismic refraction data. The degree of thinning along the length of the fracture zone is variable, ranging from one section with no appreciable thinning to others in which the crustal thickness is less than 3 km. The excess depth of the fracture zone compared to surrounding seafloor is not isostatically compensated by local thinning of the crust. Other major topographic features, such as the median valley and the elevated transverse ridge along the south wall of the fracture zone, are also not locally compensated by changes in crustal thickness.

The seismic moment budget of slowly spreading ridges

Mid-ocean ridges with well developed median valleys are commonly the sites of large normal-faulting earthquakes1,2. A recent global study3–6 of the centroid depths and source mechanisms of such earthquakes has yielded two important conclusions: first, that all the largest mid-ocean ridge earthquakes occur beneath the inner floor of the median valley, and second, that the maximum thickness of the seismogenic layer increases with decreasing spreading rate. These results lend support to the hypothesis that the median valley arises from the necking of a strong, brittle lithosphere under horizontal extension7–9. Here we compare the global rate of seismic moment release along slowly spreading mid-ocean ridges with the rate predicted if widening of the median valley occurs entirely by seismogenic extension of the brittle lithosphere beneath the inner floor. We find that earthquakes account for no more than about 10–20% of the rate of plate separation across the central median valley.

Structure and development of the Midocean Ridge Plate Boundary in the Gulf of Aden: Evidence from GLORIA Side Scan Sonar

A GLORIA side scan sonar survey along the crest of the midocean ridge in the Gulf of Aden has provided information on the small‐scale morphology of its spreading centers and fracture zones. Spreading center offsets increase in length to the east. This is thought to result from changes in the behavior of the mantle with increasing distance from the Afar hot spot. A 300‐km‐long spreading center in the western Gulf of Aden is oriented obliquely to the spreading direction. Nested within the obliquely oriented median valley is an en echelon arrangement of a number of smaller, elongated basins, each oriented approximately normal to the spreading direction. Close to the axis of spreading, volcanic and closely spaced tectonic structures run approximately normal to the spreading direction. Higher on the walls of the median valley, more‐widely‐spaced normal faults have formed subparallel to the valley. In the western Gulf of Aden there are no indications of sonar targets that might represent transform faults parallel to the NNE spreading direction. In the east, spreading center offsets of 15–25 km are made at fracture zones which are best described as narrow zones of highly oblique spreading. Offsets of greater than 25 km are achieved on single transform faults, and the associated tectonic fabric has a spreading direction trend. The boundaries between crusts of different ages along inactive transform fault extensions lie along the bases of fracture zone scarps. These scarps are situated in the older crust and face the younger. This is the opposite of the situation predicted on the basis of simple plate cooling models. At some places in the Gulf of Aden the fracture zone topography is very subdued, and this is attributed to the failure of the two slabs of lithosphere at the end of the transform fault to couple completely before the younger lithosphere on the nontransform side of the spreading center completes its ascent of the median valley wall. There is evidence of a change of spreading direction 3.5 m.y. ago, from about 040° to 022°. It is argued that the small‐scale morphology of the transform fault/spreading center intersections can be understood in terms of the transfer of shear fracture on a transform fault to tensile fracture on a spreading center propagating away from the intersection. It is concluded that the transform fault must be weaker than the spreading center.

Two- and three-dimensional inversions of magnetic anomalies in the MARK area (Mid-Atlantic Ridge 23° N)

Magnetic data collected in conjunction with a Sea Beam bathymetric survey of the Mid-Atlantic Ridge south of the Kane Fracture Zone are used to constrain the spreading history of this area over the past 3 Ma. Two-dimensional forward modeling and inversion techniques are carried out, as well as a full three-dimensional inversion of the anomaly field along a 90-km-long section of the rift valley. Our results indicate that this portion of the Mid-Atlantic Ridge, known as the MARK area, consists of two distinct spreading cells separated by a small, zero-offset transform or discordant zone near 23°10′ N, The youngest crust in the median valley is characterized by a series of distinct magnetization highs which coalesce to form two NNE-trending bands of high magnetization, one on the northern ridge segment which coincides with a large constructional volcanic ridge, and one along the southern ridge segment that is associated with a string of small axial volcanos. These two magnetization highs overlap between 23° N and 23°10° N forming a non-transform offset that may be a slow spreading ridge analogue of the small ridge axis discontinuities found on the East Pacific Rise. The crustal magnetizations in this overlap zone are generally low, although an anomalous, ESE-trending magnetization high of unknown origin is also present in this area. The present-day segmentation of spreading in the MARK area was inherited from an earlier ridge-transform-ridge geometry through a series of small (∼ 10 km) eastward ridge jumps. These small ridge jumps were caused by a relocation of the neovolcanic zone within the median valley and have resulted in an overall pattern of asymmetric spreading with faster rates to the west (14 mm yr−1) than to the east (11 mm yr−1). Although the detailed magnetic survey described in this paper extends out to only 3 Ma old crust, a regional compilation of magnetic data from this area by Schoutenet al. (1985) indicates that the relative positions and dimensions of the spreading cells, and the pattern of asymmetric spreading seen in the MARK area during the past 3 Ma, have characterized this part of the Mid-Atlantic Ridge for at least the past 36 Ma.

The morphology and tectonics of the Mark area from Sea Beam and Sea MARC I observations (Mid-Atlantic Ridge 23° N)

High-resolution Sea Beam bathymetry and Sea MARC I side scan sonar data have been obtained in the MARK area, a 100-km-long portion of the Mid-Atlantic Ridge rift valley south of the Kane Fracture Zone. These data reveal a surprisingly complex rift valley structure that is composed of two distinct spreading cells which overlap to create a small, zero-offset transform or discordant zone. The northern spreading cell consists of a magmatically robust, active ridge segment 40–50 km in length that extends from the eastern Kane ridge-transform intersection south to about 23°12′ N. The rift valley in this area is dominated by a large constructional volcanic ridge that creates 200–500 m of relief and is associated with high-temperature hydrothermal activity. The southern spreading cell is characterized by a NNE-trending band of small (50–200 m high), conical volcanos that are built upon relatively old, fissured and sediment-covered lavas, and which in some cases are themselves fissured and faulted. This cell appears to be in a predominantly extensional phase with only small, isolated eruptions. These two spreading cells overlap in an anomalous zone between 23°05′ N and 23°17′ N that lacks a well-developed rift valley or neovolcanic zone, and may represent a slow-spreading ridge analogue to the overlapping spreading centers found at the East Pacific Rise. Despite the complexity of the MARK area, volcanic and tectonic activity appears to be confined to the 10–17 km wide rift valley floor. Block faulting along near-vertical, small-offset normal faults, accompanied by minor amounts of back-tilting (generally less than 5°), begins within a few km of the ridge axis and is largely completed by the time the crust is transported up into the rift valley walls. Features that appear to be constructional volcanic ridges formed in the median valley are preserved largely intact in the rift mountains. Mass-wasting and gullying of scarp faces, and sedimentation which buries low-relief seafloor features, are the major geological processes occurring outside of the rift valley. The morphological and structural heterogeneity within the MARK rift valley and in the flanking rift mountains documented in this study are largely the product of two spreading cells that evolve independently to the interplay between extensional tectonism and episodic variations in magma production rates.

Variations in axial processes on the Mid-Atlantic Ridge: The median valley of the MARK area

Submersible observations and photogeology document dramatic variations in the distribution of young volcanic rocks, faulting, fissuring, and hydrothermal activity along an 80 km-long segment of the Mid-Atlantic Ridge south of the Kane Transform (MARK Area). These variations define two spreading cells separated by a cell boundary zone or a small-offset transform zone. The northern spreading cell is characterized by a median ‘neovolcanic’ ridge which runs down the axis of the median valley floor for 40 km. This edifice is as much as 4 km wide and 600 m high and is composed of very lightly sedimented basalts inferred to be < 5000 years old. It is the largest single volcanic constructional feature discovered to date on the Mid-Atlantic Ridge. The active Snake Pit hydrothermal vent field is on the crest of this ridge and implies the presence of a magma chamber in the northern spreading cell. In contrast, the southern cell is characterized by small, individual volcanos similar in size to the central volcanos in the FAMOUS area. Two of the volcanos that were sampled appear to be composed of dominantly glassy basaltic rocks with very light sediment cover; whereas, other volcanos in this region appear to be older features. The boundary zone between the two spreading cells is intensely faulted and lacks young volcanic rocks. This area may also contain a small-offset ( < 8 km) transform zone. Magmatism in the northern cell has been episodic and tens of thousands of years have lapsed since the last major magmatic event there. In the southern cell, a more continuous style of volcanic accretion appears to be operative. The style of spreading in the southern cell may be much more typical for the Mid-Atlantic Ridge than that of the northern cell because the latter is adjacent to the 150 km-offset Kane Transform that may act as a thermal sink along the MAR. Such large transforms are not common on the MAR, therefore, lithosphere produced in a spreading cell influenced by a large transform may also be somewhat atypical.