Henry J. B. Dick

Chromian spinel as a petrogenetic indicator in abyssal and alpine-type peridotites and spatially associated lavas

The composition of chromian spinel in alpine-type peridotites has a large reciprocal range of Cr and Al, with increasing Cr# (Cr/(Cr+Al)) reflecting increasing degrees of partial melting in the mantle. Using spinel compositions, alpine-type peridotites can be divided into three groups. Type I peridotites and associated volcanic rocks contain spinels with Cr#<0.60; Type III peridotites and associated volcanics contain spinels with Cr#>0.60, and Type II peridotites and volcanics are a transitional group and contain spinels spanning the full range of spinel compositions in Type I and Type II peridotites. Spinels in abyssal peridotites lie entirely within the Type I spinel field, making ophiolites with Type I alpine-type peridotites the most likely candidates for sections of ocean lithosphere formed at a midocean ridge. The only modern analogs for Type III peridotites and associated volcanic rocks are found in arc-related volcanic and intrusive rocks, continental intrusive assemblages, and oceanic plateau basalts. We infer a sub-volcanic arc petrogenesis for most Type III alpine-type peridotites. Type II alpine-type peridotites apparently reflect composite origins, such as the formation of an island-arc on ocean crust, resulting in large variations in the degree and provenance of melting over relatively short distances. The essential difference between Type I and Type III peridotites appears to be the presence or absence of diopside in the residue at the end of melting.Based on an examination of co-existing rock and spinel compositions in lavas, it appears that spinel is a sensitive indicator of melt composition and pressure of crystallization. The close similarity of spinel composition fields in genetically related basalts, dunites and peridotites at localities in the oceans and in ophiolite complexes indicates that its composition reflects the degree of melting in the mantle source region. Accordingly, we infer from the restricted range of spinel compositions in abyssal basalts that the degree of mantle melting beneath mid-ocean ridges is generally limited to that found in Type I alpine-type peridotites. It is apparent, therefore, that the phase boundary OL-EN-DI-SP +melt⇋OL-EN-SP+melt has limited the degree of melting of the mantle beneath mid-ocean ridges. This was clearly not the case for many alpine-type peridotites, implying very different melting conditions in the mantle, probably involving the presence of water.

An ultraslow-spreading class of ocean ridge

New investigations of the Southwest Indian and Arctic ridges reveal an ultraslow-spreading class of ocean ridge that is characterized by intermittent volcanism and a lack of transform faults. We find that the mantle beneath such ridges is emplaced continuously to the seafloor over large regions. The differences between ultraslow- and slow-spreading ridges are as great as those between slow- and fast-spreading ridges. The ultraslow-spreading ridges usually form at full spreading rates less than about 12 mm yr-1, though their characteristics are commonly found at rates up to approximately 20 mm yr-1. The ultraslow-spreading ridges consist of linked magmatic and amagmatic accretionary ridge segments. The amagmatic segments are a previously unrecognized class of accretionary plate boundary structure and can assume any orientation, with angles relative to the spreading direction ranging from orthogonal to acute. These amagmatic segments sometimes coexist with magmatic ridge segments for millions of years to form stable plate boundaries, or may displace or be displaced by transforms and magmatic ridge segments as spreading rate, mantle thermal structure and ridge geometry change.

Abyssal peridotites, very slow spreading ridges and ocean ridge magmatism

Summary The SW Indian and American-Antarctic Ridges are two of the world’s slowest spreading ocean ridges (less than 1 cm a−1), making them the low end-members for rate of ocean ridge magma supply. Two-thirds of the rocks dredged at the numerous large offset transforms along the ridges are residual mantle peridotites. Gabbroic rocks, however, representing layer 3 and possible palaeo-magma chambers are rare. This suggests a highly segmented crustal structure, with anomalously thin crust near fracture zones that may consist of only a thin veneer of pillow basalt erupted over mantle peridotite. The dredged peridotites underwent high degrees of melting, spanning the range believed to produce abyssal basalt. Their depleted compositions show that the melt was almost entirely removed. At the same time, the spatially associated basalts have a large range of compositions, similar to those from the rift valleys, requiring extensive shallow-level fractional crystallization. Since there is little evidence for magma chambers at these fracture zones, it is concluded that melts formed in the underlying mantle flowed laterally through the mantle beneath the crust towards a magmatic centre at the mid-point of an adjacent ridge segment. Magma was then subsequently intruded down the rift valley fissure system from the magmatic centre to erupt onto the fracture zone floor. Alternatively, the melt was drained from a mantle diapir beneath the midpoint of a ridge segment, prior to lateral flow of the residual peridotite beneath the ridge axis to the fracture zone. These processes suggest behaviour of the partially molten layer beneath ocean ridges analogous to Rayleigh-Taylor fluid instability, where a light less viscous fluid layer floating upwards in a denser medium goes unstable and drains at regularly spaced points into protrusions which rise rapidly to the surface. Evidence for such dynamically driven non-uniform melt flow in the mantle is seen in locally-abundant plagioclase peridotites, where the plagioclase crystallized from impregnated trapped melt. These rocks can contain up to 30% trapped melt, contrasting sharply with the typical abyssal peridotite which contains virtually none. Basalts erupted along these ridges provide a classic case of trace- and major-element decoupling during magma genesis. Despite trace-element and isotopic diversity, basalts from individual ridge segments were derived from primary magmas with similar major-element compositions. These observations can be explained if melt flows locally through the depleted mantle at the end of melting towards the midpoint of a ridge segment. This would cause melts originating at different points in an initially heterogeneous mantle to migrate through and equilibrate with the same section of mantle immediately prior to segregation—which, for the most part, would homogenize the melt’s major-element compositions. However, by virtue of the lever rule, this would have little effect on critical incompatible-trace-element or isotopic ratios of the migrating melts because of the very low incompatible-trace-element content of residual peridotite. Ocean ridges, then, appear to be marked by strings of regularly spaced volcanic centres overlying instability points in the partially molten upwelling asthenosphere much as has been postulated for arc volcanism and early continental rifting. Unlike arcs, the asthenosphere upwells to the base of the crust and the magmatic centres undergo continuous extension. Thus, large volcanoes are not constructed, and instead, ribbons of basaltic crust form parallel to the spreading direction. This is most evident at the SW Indian and American-Antarctic Ridges because of their highly attenuated magma supply. Where the magma supply is more robust and the magma chambers are correspondingly larger, the chambers may merge and eliminate the surficial morphological and chemical expression of punctuated magmatism at ocean ridges.

Mineralogic variability of the uppermost mantle along mid-ocean ridges

Abstract Modal analyses of 273 different peridotites representing 43 dredge stations in the Atlantic, Caribbean, and Indian Oceans define three separate melting trends. Peridotites dredged in the vicinity of “mantle plumes” or hot spots have the most depleted compositions in terms of basaltic components, while peridotites dredged at locations removed from such regions are systematically less depleted. The modal data correlate well with mineral compositions, with the peridotites most depleted in pyroxene also having the most refractory mineral compositions. This demonstrates that they are the probable residues of variable degrees of mantle melting. Further, there is a good correlation between the modal compositions of the peridotites and the major element composition of spatially associated dredged basalts. This demonstrates for the first time that the two must be directly related, as is frequently postulated. The high degree of depletion of the peridotites in basaltic major element components in the vicinity of some documented mantle plumes provides direct evidence for a thermal anomaly in such regions—justifying their frequent designation as “hot spots”. The high incompatible element concentrations in these “plume” basalts, however, are contrary to what is expected for such high degrees of melting, and thus require either selective contributions from locally more abundant enriched veins and/or contamination by a volatile-rich metasomatic front from depth.

Coupled major and trace elements as indicators of the extent of melting in mid-ocean-ridge peridotites

Rocks in the Earths uppermost sub-oceanic mantle, known as abyssal peridotites, have lost variable but generally large amounts of basaltic melt, which subsequently forms the oceanic crust. This process preferentially removes from the peridotite some major constituents such as aluminium, as well as trace elements that are incompatible in mantle minerals (that is, prefer to enter the basaltic melt), such as the rare-earth elements. A quantitative understanding of this important differentiation process has been hampered by the lack of correlation generally observed between major- and trace-element depletions in such peridotites. Here we show that the heavy rare-earth elements in abyssal clinopyroxenes that are moderately incompatible are highly correlated with the Cr/(Cr + Al) ratios of coexisting spinels. This correlation deteriorates only for the most highly incompatible elements—probably owing to late metasomatic processes. Using electron- and ion-microprobe data from residual abyssal peridotites collected on the central Indian ridge, along with previously published data, we develop a quantitative melting indicator for mantle residues. This procedure should prove useful for relating partial melting in peridotites to geodynamic variables such as spreading rate and mantle temperature.

A long in situ section of the lower ocean crust: results of ODP Leg 176 drilling at the Southwest Indian Ridge

Ocean Drilling Program Leg 176 deepened Hole 735B in gabbroic lower ocean crust by 1 km to 1.5 km. The section has the physical properties of seismic layer 3, and a total magnetization sufficient by itself to account for the overlying lineated sea-surface magnetic anomaly. The rocks from Hole 735B are principally olivine gabbro, with evidence for two principal and many secondary intrusive events. There are innumerable late small ferrogabbro intrusions, often associated with shear zones that cross-cut the olivine gabbros. The ferrogabbros dramatically increase upward in the section. Whereas there are many small patches of ferrogabbro representing late iron- and titanium-rich melt trapped intragranularly in olivine gabbro, most late melt was redistributed prior to complete solidification by compaction and deformation. This, rather than in situ upward differentiation of a large magma body, produced the principal igneous stratigraphy. The computed bulk composition of the hole is too evolved to mass balance mid-ocean ridge basalt back to a primary magma, and there must be a significant mass of missing primitive cumulates. These could lie either below the hole or out of the section. Possibly the gabbros were emplaced by along-axis intrusion of moderately differentiated melts into the near-transform environment. Alteration occurred in three stages. High-temperature granulite- to amphibolite-facies alteration is most important, coinciding with brittle^ductile deformation beneath the ridge. Minor greenschist-facies alteration occurred under largely static conditions, likely during block uplift at the ridge transform intersection. Late post-uplift low-temperature alteration produced locally abundant smectite, often in previously unaltered areas. The most important features of the high- and low-temperature alteration are their respective

Magmatic and amagmatic seafloor generation at the ultraslow-spreading Gakkel ridge, Arctic Ocean

A high-resolution mapping and sampling study of the Gakkel ridge was accomplished during an international ice-breaker expedition to the high Arctic and North Pole in summer 2001. For this slowest-spreading endmember of the global mid-ocean-ridge system, predictions were that magmatism should progressively diminish as the spreading rate decreases along the ridge, and that hydrothermal activity should be rare. Instead, it was found that magmatic variations are irregular, and that hydrothermal activity is abundant. A 300-kilometre-long central amagmatic zone, where mantle peridotites are emplaced directly in the ridge axis, lies between abundant, continuous volcanism in the west, and large, widely spaced volcanic centres in the east. These observations demonstrate that the extent of mantle melting is not a simple function of spreading rate: mantle temperatures at depth or mantle chemistry (or both) must vary significantly along-axis. Highly punctuated volcanism in the absence of ridge offsets suggests that first-order ridge segmentation is controlled by mantle processes of melting and melt segregation. The strong focusing of magmatic activity coupled with faulting may account for the unexpectedly high levels of hydrothermal activity observed.

Open system melting and temporal and spatial variation of peridotite and basalt at the Atlantis II fracture zone

In situ ion microprobe analysis of trace and rare earth elements in discrete diopsides in abyssal peridotites from nine transform dredge hauls from the Atlantis II Fracture Zone (All FZ) shows that these samples have a wide range of trace element contents close to the total range found for the entire Southwest Indian Ridge. Though the spread in analyses is large, the average composition of the peridotites is close to that reported for the All FZ by Johnson et al. (1990) and lies at the relatively undepleted end of the spectrum for SW Indian Ridge residual mantle peridotites. A sharp break in peridotite diopside composition and modal mineralogy occurs across the transform, suggesting that it acts as a boundary for different melting regimes and initial mantle compositions. The difference in peridotite compositions is mirrored in spatially associated basalts, which lie on separate parallel liquidus trends in the normative ternary pyroxene-olivine-plagioclase. Basalts from the east side of the transform have higher normative plagioclase contents, indicating that they may be products of lower degrees of mantle melting than basalts from the western side, consistent with greater depletion of peridotites from the western wall or a more depleted initial composition. Basalts from the eastern wall also have consistently lower Fe8.0 and higher Na8.0 than basalts from the western wall and lie parallel to the global along-ridge Fe8.0 − Na8.0 trend (Klein and Langmuir, 1987) and orthogonal to the local melting paths of Klein and Langmuir (1989). Our data provide strong evidence for segmentation of the melting regime, with major mantle discontinuities occurring at transform offsets at slow spreading ridges. Peridotites analyzed along the eastern wall of the fracture zone also show a systematic change in composition with latitude and, with the older peridotites from the median tectonic ridge, define a systematic change in the degree of melting of the mantle occurring beneath the paleoridge axis over the last 11 m.y. Emplacement of mantle showing the lowest degree of melting, or the least depleted parental mantle composition, corresponds roughly to the time of crystallization of Ocean Drilling Program site 735B gabbros. Melting is modeled as a non-steady state, discontinuous process with 0.1–0.5 vol % aggregated melt retained in the porous residue (open system melting). The range in degree of open system melting for the combined suite of All FZ peridotites is 8–20%. Such a large systematic variation would appear to require a dynamically significant change with time, either in the initial temperature and/or a large compositional difference of the mantle beneath the paleoridge axis. This in turn suggests that in the relative reference frame of the ridge axis, mantle flow was non-steady state. This could reflect episodic mantle diapirism beneath the ridge axis or, alternatively, that the ridge axis has moved over a zone of enhanced upflow in the underlying mantle that was fixed in the absolute hotspot mantle reference frame.

Tectonics of ridge-transform intersections at the Kane fracture zone

The Kane Transform offsets spreading-center segments of the Mid-Atlantic Ridge by about 150 km at 24° N latitude. In terms of its first-order morphological, geological, and geophysical characteristics it appears to be typical of long-offset (>100 km), slow-slipping (2 cm yr-1) ridge-ridge transform faults. High-resolution geological observations were made from deep-towed ANGUS photographs and the manned submersible ALVIN at the ridge-transform intersections and indicate similar relationships in these two regions. These data indicate that over a distance of about 20 km as the spreading axes approach the fracture zone, the two flanks of each ridge axis behave in very different ways. Along the flanks that intersect the active transform zone the rift valley floor deepens and the surface expression of volcanism becomes increasingly narrow and eventually absent at the intersection where only a sediment-covered ‘nodal basin’ exists. The adjacent median valley walls have structural trends that are oblique to both the ridge and the transform and have as much as 4 km of relief. These are tectonically active regions that have only a thin (<200 m), highly fractured, and discontinuous carapace of volcanic rocks overlying a variably deformed and metamorphosed assemblage of gabbroic rocks. Overprinting relationships reveal a complex history of crustal extension and rapid vertical uplift. In contrast, the opposing flanks of the ridge axes, that intersect the non-transform zones appear to be similar in many respects to those examined elsewhere along slow-spreading ridges. In general, a near-axial horst and graben terrain floored by relatively young volcanics passes laterally into median valley walls with a simple block-faulted character where only volcanic rocks have been found. Along strike toward the fracture zone, the youngest volcanics form linear constructional volcanic ridges that transect the entire width of the fracture zone valley. These volcanics are continuous with the older-looking, slightly faulted volcanic terrain that floors the non-transform fracture zone valleys. These observations document the asymmetric nature of seafloor spreading near ridge-transform intersections. An important implication is that the crust and lithosphere across different portions of the fracture zone will have different geological characteristics. Across the active transform zone two lithosphere plate edges formed at ridge-transform corners are faulted against one another. In the non-transform zones a relatively younger section of lithosphere that formed at a ridge-non-transform corner is welded to an older, deformed section that initially formed at a ridge-transform corner.

The fingerprint of seawater circulation in a 500-meter section of ocean crust gabbros

A novel strip-sampling technique has been applied to the 500-m gabbroic section drilled at site 735 during Leg 118. Twenty-two continuous strips of 1.1- to 4.5-m length were cut longitudinally from the core, allowing for a more representative sampling of this section of the deep ocean crust. A full suite of trace element and isotopic (Sr, Nd, Pb, Os, δ18O) analyses were conducted on these strip samples; for comparison, analyses were conducted on a small suite of protolith samples, selected for their fresh and unaltered appearance. Amphibole, diopside, and plagioclase from 18 vein samples were also analyzed for Sr and Nd isotopes. Although the evidence for a seawater component in these gabbros is clear (87/86 Sr up to 0.70316; 206/204 Pb up to 19.3; δ18O down to 2.0‰; 187/188 Os up to 0.44), the trace element signatures are dominated by magmatic effects (infiltration and impregnation by late-stage melts derived locally or from deeper levels of the crust). The average upper 500 m 735B gabbro section is somewhat lower than average N-MORB in trace elements such as Ba (30%), Nb (50%), U (40%), and heavy REE (Yb and Lu, 30%), but somewhat enriched in others such as La (23%), Ce (24%), Pb (23%), and Sr (40%). Although the section is largely comprised of cumulate gabbros (Natland et al., 1991), and many of the strip samples show marked Sr and Eu anomalies (plagioclase cumulation), the average composition of the total 500 m section shows no Sr or Eu anomalies (<1%). This implies that there has been local separation of melt and solids, but no large scale removal of melts from this 500-m gabbro section.