Hararo basalts. Furthermore, the development of secondary phases is unable to increase the Eu/Eu* and to generate positive Eu anomalies, ruling out this explanation.
Positive Ba, Sr and Eu anomalies in basaltic rocks, are usually explained by plagioclase accumulation or digestion. However this process is unable to produce Sr anomalies as high as those displayed by the Hayyabley or Manda Hararo basalts without increasing drastically the Al2O3 contents of the resulting rocks. The Al2O3 abundances of the LREE-depleted basalts are not anomalously high, which immediately rules out this explanation.
The Hayyabley basalts show elevated Sr isotopic compositions and low εNd
values relative to Aden Gulf or Red Sea MORB (Schilling et al., 1992; Volker et al., 1993; Hase et al., 2000). The assimilation of a continental component could explain this shift from usual N-MORB values, but incompatible trace element ratios give no support to this interpretation. Contamination of MORB-like melts with continental crust would produce significant changes in incompatible trace element ratios. The Hayyabley basalts, like the Manda Hararo LREE-depleted basalts, lack the classical negative Nb or Ta anomalies observed in crust-contaminated basalts, and show a limited range of Ce/Pb ratios from 24 to 28, similar to values measured in oceanic basalts (e.g., Sun and McDonough, 1989). Thus, there is no indication for assimilation of significant amounts of material derived from the crust in the LREE-depleted basalts. In the case of the Manda Hararo basalts, this conclusion is strengthened by their δ18O values close to 5.5
‰, which are typical of mantle composition (Barrat et al., 2003).
As previously pointed out for the Manda Hararo basalts (Barrat et al., 2003), the positive Sr, Ba and Eu anomalies and the particular Sr-Nd-Pb displayed by the Hayyabley basalts are more likely a genuine feature inherited from their mantle sources.
The same conclusions have been reached for depleted basalts with similar positive anomalies from Iceland. As an alternative mechanism, Chauvel and Hémond (2000), Skovgaard et al. (2001), and Kokfelt et al. (2006) have suggested that the sources of Icelandic lavas contained an old recycled oceanic lithosphere component and that melting of the gabbroic portion of this lithosphere led to the formation of basalts that exhibit large positive Ba, Sr and Eu anomalies. At first glance, such an explanation is attractive because if this recycled gabbroic component has been hydrothermally altered, one may expect much more radiogenic 87Sr/86Sr ratios than those of typical MORB.
Hence, the involvement of such component could account for the relatively high
87Sr/86Sr ratios of the Manda Hararo and Hayyabley depleted basalts. However, an old light REE depleted gabbroic component would be certainly characterized by high εNd
values. To the converse, the Manda Hararo and Hayyabley lavas display εNd values unexpectedly low (εNd= 5-7) for depleted basalts. We conclude that this model fits at best only partially the observations.
5.2. The depleted components in the sources of the basalts from Ethiopia and Djibouti
Previous geochemical studies have emphasized the participation of a depleted component during the genesis of the NE Africa basalts. In the case of basalts emitted by the young oceanic ridges from the Red Sea or the Aden Gulf, major involvement of MORB-related sources has been proposed (e.g., Barrat et al., 1990, 1993; Schilling et al., 1992; Volker et al., 1993). On land, in Afar and Ethiopia, where enriched basalts have been massively erupted, the geochemical and isotopical features of the depleted reservoirs which have been involved during the genesis of the lavas, are very difficult to constrain. Sometime, rare lavas display the signatures of such components or some mixing trends permit to deduce some of their features. Two distinct light REE depleted components have been unambiguously detected: (1) a MORB-like component is involved in the genesis of Quaternary basalts from Northern Afar; the Sr-Nd-Pb isotopic relationships displayed by the Erta’Ale basalts point to the participation of two mantle end-members, namely a HIMU component and a depleted component undistinguishable of the source of the Red Sea MORB (Barrat et al., 1998);
furthermore, a similar depleted component has been detected in the sources of the Oligocene lavas from the Northwestern Ethiopian volcanic province (Meshesha and Shinjo, 2007); the entrainment of depleted mantle during plume ascent is a plausible explanation for the contribution of this component in the sources of some of the basalts erupted in Afar and Ethiopia. (2) the light REE depleted basalts from Manda Hararo and Hayyabley point to a depleted endmember chemically and isotopically distinct from the previous one; in addition, a single light REE depleted basalt that displays the same chemical features as the depleted basalts from Afar has been collected
in Ethiopia (Pik et al., 1999); although, its isotopic compositions are not exactly the same as the Hayyabley basalts (see figures 4 to 6, sample E88), this sample indicates that a depleted component distinct from the MORB source was involved in this area at this time. We suggest that a depleted component, intrinsic to the plume, has contributed to these young and old lavas related to the Afar hot spot.