Summary: | We report on Nd isotopes and Sm and Nd contents of 46 fresh Mid-Atlantic Ridge basalts over the Azores swell from 30°N to 50°N. The latitudinal ¹⁴³Nd/¹⁴⁴Nd and Sm/Nd profiles anti-correlate with the ⁸⁷Sr/⁸⁶Sr profile previously established [1]. Sharp minima in ¹⁴³Nd/¹⁴⁴Nd, identical in value, are found at the 46°N, 43°N, 39°N, and 35°N short-wavelength geochemical anomalies previously recognized, as well as the 1100 km long gradient south of the Azores platform. A single binary mixing line is observed in ¹⁴³Nd/¹⁴⁴Nd vs. ⁸⁷Sr/⁸⁶Sr space, suggesting that the incompatible element-rich mantle sources underlying these four short-wavelength anomalies have a similar long-term origin. However, at a given ¹⁴³Nd/¹⁴⁴Nd (or ⁸⁷Sr/⁸⁶Sr), the MORB population from 40.5°N to 50°N is systematically lower in incompatible elements and their ratios, such as La/Sm, Nd/Sm, Cl/F and Rb/Sr, than the 40.5°N-30°N population. The required fractionation must be relatively recent. But the bulk chemistry of these MORB rules out that it was generated directly under the ridge by large varying degrees of partial melting of passive heterogeneities embedded in the depleted asthenosphere. An earlier, but not too distant, metasomatic enrichment or depletion event is required. A small 1–2% melt fraction removal by fractional decompression melting of a wet garnet lherzolite, at its wet solidus, can readily account for the depletion of the 46°N and 43°N relative to the 39°N mantle sources. We speculate that the short-wavelength 46°N, 43°N and 39°N anomalies were caused either by: (1) a family of plumes with the same long-term origin and thermal boundary layer source; or (2) by the detachment of blobs from a single bending plume converging with the MAR axis at 39°N and preferentially discharging southwestward along the spreading axis. In either case, the rising plumes, or the detached blobs, at 46°N and 43°N occurred on the east flank of the MAR below the thick lithosphere. The plume-derived material reaching the MAR axis at these latitudes would be ...
|