Region Of Partial Melt Research Articles

Transform faults and longitudinal flow below the Midoceanic Ridge

Transform faults offset the pipe-shaped region of partial melting and magma generation below the Midoceanic ridge. Hence if there is flow along the pipe, it will be blocked or at least impeded at major transform faults. There is evidence on the Reykjanes ridge that minor transform faults, with offsets of a few tens of kilometers, may be converted to oblique spreading axes by asthenosphere flow. Quantitative estimates of the extent of blocking are derived from a Parker-Oldenburg law of plate thickness increasing as the square root of crustal age. There are several kinds of evidence that the subcrustal partial melts generally moving away from long-wavelength topographic and gravity highs on the Midoceanic ridge (hot spots) are partly blocked at transform faults: (1) elevations of particular isochrons, including the present spreading axis, frequently jump discontinuously across major transform faults (the best examples of this are in the northeast Atlantic); (2) morphology and seismicity change abruptly across the Blanco fracture zone, which separates a ridge influenced by a hot spot (the Juan de Fuca ridge) from one not so influenced (the Gorda ridge); (3) a zone of relatively high magnetic amplitudes is associated with the hot spot influenced Juan de Fuca and central Galapagos ridges (this zone may delineate how much crust was produced by the Fe/Ti-rich melts of hot spot origin (whether due to distinctive source composition or subsequent fractionation); on both ridges the high-amplitude magnetic zones are terminated on both ends at transform faults, again suggesting blockage of the hot spot melts); (4) prominent ridges, such as the Mendocino and Charlie Gibbs ridges, may form on the ‘upstream’ side of some fracture zones at certain times during their evolution (such fracture ridges seem to be constructional volcanic piles formed continuously at or near the fracture-ridge crest junction; they are here interpreted as being due to excess basalt melts, produced from the partial melts ponded on the upstream side of the transform fault; fracture ridges are proposed to record the past intensity of pipe flow and hence hot spot activity); and (5) recent studies by Solomon (1973) show a region of high S wave attenuation and low Q at the southern end of the Reykjanes ridge, below the prominent fracture ridge just north of the Charlie Gibbs fracture zone. Taken together, the various observations support the existence of pipelike flow at shallow depths (up to a few tens of kilometers) below the Midoceanic ridge. The interaction between mantle hot spots and nearby plate boundaries discussed in this paper may produce features, such as the Ninetyeast and Broken ridges, that exhibit a strong plate tectonic fabric (structures parallel or perpendicular to transform faults) but nevertheless attest to special convective processes in the mantle.

Shear wave attenuation and melting beneath the Mid-Atlantic Ridge

Because the attenuation of seismic waves is sensitive to variations in temperature and to partial melting, the mapping of seismic Q beneath the mid-ocean ridge systems is a useful tool to outline boundaries between lithosphere and asthenosphere and to constrain the mechanics of the intrusion process. Our approach is to measure the differential attenuation of long-period shear waves, using a spectral ratio technique, from earthquakes on the ridge and to look for variations in attenuation with propagation direction. We correct for propagation distance and, where known, the upper mantle attenuation beneath the receiving stations. The azimuthal dependence of attenuation of S waves from an earthquake offset from the ridge axis on a transform fault indicates the existence of a low-Q zone, no wider than 100 km and confined to depths shallower than about 50 to 150 km, beneath the crest of the mid-Atlantic ridge. The absence of appreciable azimuthal variation in shear wave attenuation for an earthquake on the ridge crest suggests that the low-Q zone is at least 50 km wide. Q within such a zone must be 10 or less for long-period S waves. The most likely explanation of such a low-Q zone of limited spatial extent and with sharply defined boundaries is that the zone is a region of extensive partial melting, probably at temperatures in excess of the anhydrous solidus of mantle material. Such a region of large melt concentration is consistent with the chemistry of rocks from the mid-ocean ridges and with models of the temperature field derived from numerical calculations of flow beneath spreading centers.