MANTLE AND CRUSTAL EFFECTS ON THE GEOCHEMISTRY OF PROTEROZOIC DIKES AND SILLS IN SWEDEN

Chemical and Sr-Nd isotopic data are presented for 14 Proterozoic dole rite intrusions from Sweden. The dolerites are unmetamorphosed and bel ong to two age groups at 1.26 and 0.95 Ga. The analysed rocks were sel ected from over 80 samples characterized for initial Sr-87/Sr-86 ratio , and represent those from each intrusion that have little to no crust al contamination effect on Sr isotopes. In these tholeiitic to mildly alkaline basalts, MgO varies from 4 to 14%, and TiO2 from 1.0 to 4.7%. Incompatible-element abundance patterns show similar shapes at widely varying overall enrichment levels; La varies from eight to 64 times p rimitive mantle. Many samples have apparent Sr anomalies; positive for the less incompatible-element enriched rocks, negative for those with the highest abundances. Hence Sr varies by a factor of only two, wher eas neighboring elements vary by a factor of seven. Small Eu anomalies correlate with the larger Sr variations. All samples have negative Nb anomalies, correlated with epsilon(Nd) i values that vary from -0.5 t o + 3.5 for both dolerite age groups. Sr-87/Sr-86i values are systemat ically different in the two groups: 1.26-Ga dolerites vary from 0.7026 to 0.7035, and 0.95-Ga dolerites from 0.7041 to 0.7051. Rather than u rging any particular model of subcontinental mantle evolution, we use the dolerite results to deconvolute the effects of crustal contaminati on, differentiation including cryptic differentiation effects, and man tle-lithosphere interactions on possibly asthenosphere-derived magmas. Problems related to addition of Rb and Sr-87 giving false isochrons, as well as raised Sr-87/Sr-86 from syn- or post-crystallization fluid effects, were eliminated by the sample selection criteria. In the sele cted dolerites, crustal contamination is not a major control on geoche mistry, because all incompatible elements, including those of both hig h and low crustal abundance, are positively and similarly correlated. Overall abundances have been affected by migrations of residual magma after much olivine and plagioclase crystallization. Only Sr is an appr oximate indication of the parent magma's enrichment, whereas all other incompatible elements may be depleted or enriched by factors greater than two. On entering the crust, the magmas varied by a factor of only two in overall incompatible-element enrichment. They had Nb anomalies correlated with epsilon(Nd), and the 0.95-Ga dolerites possessed high er Sr-87/Sr-86 than had characterized the 1.26-Ga dolerites. All mantl e effects could be interpreted as resulting from a lithospheric magma source that experienced addition of subduction-related components 1.9- 1.7 Ga ago, and maybe locally 1.6-1.0 Ga ago. We prefer a model where asthenosphere-derived melts interact with lithospheric wall rocks, acq uiring variable epsilon(Nd) down to -0.5 and Nb anomalies. Up to two-t hirds of the budget of incompatible elements, but not much Nb, were ad ded to the magmas. Ultimately, the contribution of any mantle source i s unconstrained, except that significant lithosphere has to have been involved. The mantle-related geochemical effects, which result in only weak constraints on process and source, are very difficult to isolate in an objective approach to intracontinental basalts, because of supe rimposed differentiation and contamination effects.