Crust-Poor Lithosphere at Cold Spots in the Mid Atlantic and SW Indian Ridges

The Equatorial portion of the Mid Atlantic Ridge is thought to reflect a thermal minimum in the subridge structure, with deeper than normal axial topography underlain by high upper mantle seismic velocities revealed by tomography. This stretch of Ridge is intersected by a number of long offset transforms, the longest being the Romanche (offset ~950 km corresponding to ~50 Myr). As the Mid Atlantic Ridge axis approaches the Romanche transform from the south, it gradually deepens; its rift valley disappears, and, starting roughly 50 km from the transform, the basaltic crust becomes patchy and then disappears, leaving mantle ultramafics outcropping on the sea floor. Modelling the "cold edge" effect of the transform on the axial Ridge segment shows that partial melting of the subridge mantle decreases as the transform is approached. Crust-free lithosphere outcrops continuously for several hundred kilometers in a ~30 km wide belt south of the Romanche, indicating that the present-day lack of crustal production has been prevailing for at least 30 million years. The mantle derived serpentinized peridotites are of two types. The majority of the samples show evidence of strong impregnation by basaltic melts. The mineral chemistry of the samples that are free of impregnation suggests that they have undergone a very low degree of melting. These results suggest a quasi-crust-free lithosphere, produce by a mantle that has undergone little or no partial melting, unable to expel the small quantities of melt it generates. The small quantities of basalt produced in this area tend to have alkali affinity and are strongly enriched in H2O. Their REE content show a strong garnet signature, suggesting that they were generated mostly in the garnet peridotite mantle zone (> 20 kbar). This quasi-crust-free impregnated lithosphere probably prevails at cold spots along mid ocean ridges. Peridotites obtained recently from the SW end of the Andrew Bain transform, that offsets the SW Indian Ridge by ~750 km (~50 Myr), are strongly impregnated by basaltic melt, in a situation very similar to that observed near the Romanche. In contrast, peridotites from the NE end of the Andrew Bain transform are not impregnated, and are residual of a significant degree of melting, probably due to the influence of the Marion plume located a few hundred km away.