ISOTOPIC EVIDENCE FOR CRUST MANTLE EVOLUTION WITH EMPHASIS ON THE CANADIAN SHIELD

Nd, Sr and Pb isotope data from carbonatites and related silicate rocks spanning an age range from 0.1 to 2.7 Ga in the Grenville and Superior provinces, Canadian Shield, trace the secular geochemical evolution of large ion lithophile element-depleted mantle sources beneath those provinces over the past 2.2 Ga. Sr data from 2.67-Ga-old syenites from the Abitibi, Quetico and Wabigoon belts in the Superior Province do not indicate depleted mantle sources, although ε{lunate}Nd for the same rocks averages +1. Initial Pb ratios in the syenites yield single-stage model ages that agree closely with their radiometric ages, while the model ages for the younger complexes are substantially younger than their radiometric ages. This suggests either that Pb in the 2.67-Ga complexes was not derived from aged depleted mantle sources, or that any recycled crustal component with high U/Pb had previously experienced only a short crustal residence time. The combined Pb and Sr data suggest that the depleted mantle beneath the Superior Province sampled by the complexes was initiated ∼3 Ga ago, in agreement with previously published results. Available data on carbonatites from the Fennoscandian Shield parallel the Canadian Shield isotope evolution pattern, although data from the southern hemisphere show considerable divergence, indicating regional variations. However, isotope evolution data from Precambrian marine precipitates suggest that major contributions of Sr to the oceans from sialic crust commenced ∼3 Ga ago. Assuming a complementary relationship between sialic crust and depleted mantle, the Canadian and European Sr data agree with average world-wide evolution of depleted mantle, and therefore appear to have more than regional significance. Models to reconcile the differences between the Nd and Sr results are considered. With present information several models are viable, some of which involve the first appearance of major volumes of depleted mantle around 3.0 Ga due to formation of sialic continental crust. The contrast between the Sr and Nd data may be due to differences in the way in which the bulk silicate Earth isotope evolution models for Nd compared to those for Sr and Pb are derived. Until the cause of the differences between Nd and the Sr and Pb data are better understood the presence of substantial volumes of sialic continental crust prior to ∼3.0 Ga ago is open to question. © 1990.