Robert L. Fisher

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.

Evolution of the East: Central Indian Ocean, with Emphasis on the Tectonic Setting of the Ninetyeast Ridge

The meridional Ninetyeast Ridge in the eastern Indian Ocean separates the deep Central Indian Basin from the deeper Wharton Basin (or Cocos Basin–West Australian Basin) to the east. The flattish-summited ridge extends slightly east of north from near 32° S. directly to 7° S. where it appears segmented as a series of en echelon northeast-southwest–trending highs, then in a northerly direction disappears beneath the sediments of the Bengal Fan system near 9° N. Linear parallel to subparallel troughs border this linear ridge on the east side; on the west, from results of magnetic observations and preliminary deep drilling information, Ninetyeast Ridge apparently is bonded to the Indian plate. A second extensive north-south topographic rise and magnetic boundary zone, herein named the Investigator Fracture Zone, lies near 98° E. in the Wharton Basin. Easterly trending magnetic-anomaly lineations identified as numbers 5 through 16 and numbers 21 through 33b, increasing in age northward and with spreading rates variable through time, have been recognized in the Central Indian Basin. East of Ninetyeast Ridge in the Wharton Basin, anomalies 19 through 27, with spreading rates varying in concert with those of comparable age west of the ridge, have been found to increase in age toward the south. Older anomalies 28 through 33 have been identified in both basins; their divergent trends provide evidence that spreading rates decrease markedly westward during the time span they cover in the Late Cretaceous. From deep-sea drilling information supplementing and supporting magnetic, topographic, and gravity data obtained principally by research ships and PROJECT MAGNET since 1962, we interpret Ninetyeast Ridge to be an extrusive pile with a low-density shallow root, rather than a horst or an uplift resulting from the convergence of plates. The trough system that is partially buried with sediment east of the ridge and the north-south Investigator Fracture Zone several hundred kilometers farther to the east are remnants of formerly active transform faults that marked huge relative offsets between the spreading centers separating the Indian and Antarctic-Australian plates from anomaly 33b (Late Cretaceous) to anomaly 19 (Eocene) time. During the Late Cretaceous, Ninetyeast Ridge and Chagos-Laccadive Ridge had similar settings, marking paired offsets of an active spreading center around the southern tip of India. Both features terminated as active transform faults with the cessation of north-south spreading and the commencement of northeast-southwest spreading close to the time of anomaly 11 (Oligocene). The here-interpreted oceanic data is strong but not conclusive support for fitting India to Enderby Land in Antarctica in the Early Cretaceous. With presently available information, we have been unable to establish the precise time at which the spreading center in the Wharton Basin ceased to function.

Granitic to ultramafic rock complexes of the Indian Ocean ridge system, western Indian Ocean

During five expeditions of the Scripps Institution of Oceanography to the western Indian Ocean, more than 4,500,000 sq km of the Central Indian Ridge and its branching Southeast Indian Ridge and Southwest Indian Ridge were explored by bathymetric, magnetic, and seismic-reflection profiling. In some 2,800,000 sq km of this region, igneous rocks of the crust, lower crust, and possible upper mantle are exposed by faulting or volcanism. Fifty-six dredge hauls of these igneous rocks were obtained, largely from the major cross-fractures (transform faults) or clefts trending athwart the volcanically active ridges. From north to south, the cross-fractures most intensively sampled were the Vema Fracture Zone, which crosses the crestal area near 9°S, Argo Fracture Zone near 13°30′S, Marie Celeste Fracture Zone near 17°30′S, and the newly delineated “Melville Fracture Zone” trending north-south for more than 600 km near 60°30′E on the Southwest Indian Ridge. Our field and laboratory studies indicate that under a capping of young flow basalt, there is a regional complex of igneous rocks produced by magma generated under the ridges, trapped and differentiated into sill-like, podiform, and larger, crudely stratified to well-stratified sheets. Rocks from the stratiform masses include abundant Iherzolite and minor harzburgite, orthopyroxenite, olivine- and two-pyroxene gabbros, Ti-ferrogabbros, norite, and anorthosite. Some associated diabase intrusions are granophyric and are cut by late-stage dikelets of quartz monzonite and Na-rich trondhjemite. Both calc-alkaline and alkalic lines of differentiation are indicated. The granitic dikelets contain clear, doubly-terminated crystals of zircon, unusual in a terrane of large-cation–depleted rocks. The overlying basalt flows are pillowed with chemical and mineralogical characteristics typical of olivine-bearing tholeiite from the ridge-rise systems of the world oceans. The ubiquitous nature of the crustal complex found throughout the western Indian Ocean, together with data from the Atlantic and Pacific Oceans, suggest that similar rock complexes, dominated in their lower parts by stratiform bodies, are characteristic of most of the igneous crust throughout the world oceans.

Evolution of the Central Indian Ridge, Western Indian Ocean

Topographic, magnetic, and earthquake epicenter data from the wholly submerged Central Indian Ridge were interpreted, using the Theory of Plate Tectonics. The pole of relative motion between the Indian and Somalian plates, lying at 16.0° N., 48.3° E. and with opening at 6.2 × 10−7 deg/yr, was obtained from the strike of fracture zones taken as transform faults and the spreading rates based on magnetic anomaly patterns. Since this pole appears to have moved little since the Miocene, the plate positions at that past time can be obtained by finite rotation about the present rotation pole. Such a reconstruction shows that the complicated nature of the present plate margins results from Miocene to Recent opening along a north-south fracture zone that existed in this area during an interval of rapid spreading in the late Cretaceous and early Tertiary.

Mineralogic Studies of the Residues of Mantle Melting: Abyssal and Alpine-Type Peridotites

Abstract Abyssal peridotites dredged from the ocean ridges range from diopside-poor harzburgite to lherzolite, but all contain enstatite saturated with diopside, indicating that melting of the abyssal mantle was constrained by the pseudo-invariant point Ol+En+Di+Sp+Melt. We find systematic regional differences, which suggest that a range of primary melt compositions and large variations in the apparent degree of melting exist in the mantle beneath ocean ridges. We note in particular that North Atlantic peridotites are highly depleted relative to the average abyssal peridotite. Alpine-type peridotites overlap the range for abyssal peridotites, but extend to far more depleted and enriched compositions. Many contain enstatite undersaturated with respect to diopside. Frequently alpine-type peridotites contain highly magnesian Al-poor and Cr-rich minerals lying outside the abyssal range. Melting of many alpine peridotites, therefore, has occurred well into the three phase field Ol+En+Sp+Melt under different conditions than for abyssal peridotites. Alpine-type peridotites evidently represent parageneses extending from relatively undepleted sub-continental upper mantle to periodtites melted in a volcanic-arc or near-arc oceanic environment.

Topography and structure of the Peru-Chile trench☆☆☆

During the Scripps Institutions IGY Expedition Downwind of 1957–1958 the research vessels Horizon and Spencer F. Baird spent 41 ship-days conducting geological-geophysical studies in two portions of the Peru-Chile Trench, concentrating their efforts off northern Chile and central Peru. Data from 5400 miles of echo-sounding traverses and eight seismic-refraction stations (three off Peru, five off Chile) are here reported. The peru-Chile Trench, lying off southern Ecuador to central Chile, is interrupted off southern Peru by the northeast-trending Nasca Ridge. North of Nasca Ridge the trench reaches a maximum depth of nearly 6500 m. Characteristically, the trench floor is flat; this flatness is attributed principally to land-derived sediment, introduced by the intermittenly-flowing rivers of the coastal zone, passing down the trench flank. South of Nasca Ridge, off the Atacama Desert, relatively little sediment reaches the narrow trench bottom. Here the maximum depth is slightly more than 8000 m. South of 27° 30′S the trench shoals in a series of nearly flat-floored basins. River-transported solid phases here form a major portion of the trench-floor sediment. Seismic refraction studies of two sections across the trench off Chile and Peru are very similar in that both show that crustal thickening begins westward of the trench and reaches a thickness of about 11 km beneath the trench axis. The Mohorovicic Discontinuity, which is found at a depth of about 17 km below sea level at the trench axis, plunges steeply to the east with a slope not inconsistent with the depths of about 65 km beneath the Andes, estimated from the observations of Tuve and Tatel.

Middle America Trench: Topography and Structure

From 1952 to 1959, during nine expeditions of the Scripps Institution of Oceanography and one of the U. S. Navy Electronics Laboratory, research vessels recorded 31,950 miles of echo-sounding traverses in and adjacent to the Middle America Trench, which extends from the Islas Tres Marias off western Mexico to the Cocos Ridge southwest of Costa Rica. The Middle America Trench is continuous at depths greater than 2400 fathoms (4400 m) for 1260 miles, except off Manzanillo and Zihuatanejo, Mexico, where submarine mountains lie in the trench. It is deeper than 3000 fathoms (5500 m) for 380 miles as the Guatemala Deep. Northwest of Acapulco it is generally U-shaped in cross section, with a steeper shoreward flank and a flat bottom suggesting sedimentary fill. From Acapulco southeast to the west side of the Gulf of Tehuantepec, the trench shoals, in a series of basins, to 2700 fathoms (5000 m). To the southeast it widens and deepens abruptly to a maximum 3500 fathoms (6400 m) off western Guatemala, then shoals gradually to merge into the sea floor off Costa Rica. The southeast segment is also asymmetrical in cross section but is V-shaped with irregular bottom. A northeast-trending band of ridge-and-trough topography, 60 miles wide, separates the 1800- to 1900-fathom sea floor outside the trench off southern Mexico from the 2100- to 2200-fathom Guatemala Basin. This zone has been traced from several hundred miles offshore to an intersection with the trench near the west side of the Gulf of Tehuantepec. Seismic-refraction studies reported in an accompanying paper (Shor and Fisher, 1961) were employed in determining the trench structure. Three refraction stations were taken along the axis of the trench west of Acapulco and two along its axis off Guatemala and El Salvador. Another station was shot on the shelf and one 60 miles seaward of the trench off Guatemala. Thick sediments were found in the Tres Marias Basin off Manzanillo and at the shelf station off Guatemala. Arrivals from rock with compressional wave velocity of 4–6 km/sec were observed at the Tres Marias Basin and Guatemala shelf stations. Off Guatemala, on a section normal to the trench, the depth below sea level to the M discontinuity is interpreted from these seismic data as about 9 km (Pacific Basin), 10 km (outer ridge), 16 km (trench), and 17 km (shelf). Below the sea floor the crust thickens from 5–7 to 10–17 km along this section. The M discontinuity is deeper and the crust below the sediments thicker under the two southern stations than under the two central trench stations. The mantle is deeper under the Tres Marias Basin, where thick (1½ km) sediments are found, than under the central stations. The Gulf of Tehuantepec marks a major change in trench configuration and possibly in age. Northwest of Tehuantepec the flat trench bottom developed in most places suggests a greater age. Southeast of the gulf the deep V-shaped trench, with thicker crustal layers but very little fill, borders a volcanically active coast. The zone of ridge-and-trough topography trending southwest from Tehuantepec may be another evidence of this boundary.

Petrology and Geochemistry of Igneous Rocks from the Tonga Trench: A Non-Accreting Plate Boundary

Petrologic and geochemical examination of a varied suite of intermediate, mafic, and ultramafic rocks dredged from the deep flanks of the Tonga Trench between 20°S and 21°S show that the landward slope has not developed by accretion of material from the subducted Pacific plate. The lowermost trench slopes (> 9000 m) are part of a graben in the Mesozoic Pacific plate; the west and east walls of this graben expose normal- and enriched-type ocean-ridge basalts. The distribution of recovered rock types suggests that the shoaler (< 9000 m) nearshore flank is crudely layered from peridotite (up to 8500 m), to gabbro (to 7000 m), to volcanic rocks (7000-5000 m). Peridotites are fresh to moderately serpentinized harzburgites, with some dunite and minor lherzolite. They are distinctly more depleted in clinopyroxene and have more magnesian mineral compositions than are characteristic of tectonite periedotites from ocean-ridges, fracture zones, and many ophiolites. Volcanics from the upper nearshore slope are basalts, andesites, and dacites. Unlike the N- and E-type basalts being subducted, they are most similar to low-Ti, low-Ba arc volcanics from the Lau Ridge and Mariana forearc. Overall, the coherent 4 + km crustal section of the nearshore flank exposed above the structural plate boundary at 9000+ m is geochemically unlike crust generated in mid-ocean. It may be primitive island-arc crust, older oceanic crust modified by island-arc volcanism, or atypical ocean-ridge crust upon which the Tonga arc was constructed. There is no evidence of accretion of material like that being subducted-non-accretion or tectonic erosion are inferred to be the principal processes which have shaped this margin.

Ultramafic and Basaltic Rocks Dredged from the Nearshore Flank of the Tonga Trench

Deep dredging in the Tonga Trench (Southwest Pacific Ocean) at a depth of 9150 to 9400 m yielded fresh to granulated and serpentinized peridotite and dunite. Other rocks recovered there and at three stations deeper than 7000 m include basalts, tuffs, and tuffaceous agglomerates. Chemical analyses of the fresh peridotite, with combined H2O < 0.10 weight percent, indicate that the rock consists of Si, Mg, Fe (6 percent), and Cr + Ni about 0.7 percent. Mineralogically, the peridotite contains forsteritic olivine and enstatite with minor spinels. The ultramafic mass exposed at 9400 m probably is an accumulate exposed by faulting.

Lherzolite, anorthosite, gabbro, and basalt dredged from the mid-Indian ocean ridge.

The Central Indian Ridge is mantled with flows of low-potassium basalt of uniform composition. Gabbro, anorthosite, and garnet-bearing lherzolite are exposed in cross fractures, and lherzolite is the bedrock at the center of the ridge. The Iherzolites are upper-mantle rock exposed by faulting.