Mantle sources for Central Atlantic Magmatic Province basalts from Hf isotopes

The Central Atlantic Magmatic Province (CAMP) was one of the most voluminous LIP events in Earth history and likely triggered the end-Triassic mass extinction. The tectonic and mantle processes that produced such significant magmatic emplacement are thus of great interest. To further explore the origins of CAMP, we present new 176Hf/177Hf isotope data for a broad geographic sampling of CAMP dikes, sills, and basalt flows. We find that basaltic intrusions from the Carolinas in Eastern North America trend along a shallower slope than the terrestrial array on a diagram of 176Hf/177Hf vs. 143Nd/144Nd. This trend may reflect the presence of variable quantities of sediment-derived material in the mantle source region. This is consistent with previous suggestions that the asthenosphere beneath CAMP has been partially metasomatised by fluids derived from subducted sediments, as well as with isotopic trends observed in other LIP, such as Karoo [Jourdan et al., 2007, Jour. Petrology,doi:10.1093/petrology/egm010]. Distinct from the Carolina trend, we further observe that high-TiO2 basalts from Amazonia exhibit unusually radiogenic 176Hf/177Hf for a given 208Pb/206Pb ratio. The high-TiO2 basalts, which trend towards EM1-type compositions, may be asthenospheric melts that have experienced the addition of melts from local subcontinental lithospheric mantle (SCLM). Similarly high-TiO2 CAMP rocks from Sierra Leone may likewise have incorporated enriched lithospheric melts of lamproite-like composition in the source region [Callegaro et al., JPet, accepted; GSA Abstract #302853, 2017]. Low-TiO2 basalts from the same region in Brazil and of similar age to the high-TiO2 basalts lack the observed radiogenic 176Hf/177Hf ratios. This suggests that the melt source region beneath Brazil was heterogeneous, containing variable material with relatively radiogenic 176Hf/177Hf ratios, perhaps due to the greater age of subcontinental lithosphere and the presence of garnet. It remains unclear, however, whether the hypothesized SCLM source represents lithospheric domains which are still intact, or if this material reentered the convecting mantle by delamination prior to melting