The implications of picritic lavas for the mantle sources of terrestrial volcanism

Abstract

The success of recent dynamic models of adiabatic partial melting of the Earth's mantle has lead to an increasing acceptance that much of the compositional spectrum of terrestrial basalts can be explained in terms of variations in the physical parameters of melting of a common anhydrous-lherzolite mantle source. A comparison of normalized compositions of terrestrial picritic lavas with the results of recent lherzolite melting experiments indicates, however, that picritic magmas from intraplate hot spots, such as Hawaii, can not be derived from the mantle source which gives rise to the picritic magmas associated with mid-ocean ridges or subduction zones. MORB picritic magmas could have equilibrated with a lherzolite residue wose composition resembles that of lherzolite xenoliths. Picritic magmas associated with subduction zones have compositions which range from those similar to MORB picrites to the ankaramitic picrites of the southwest Pacific, which appear to have equilibrated with a more-depleted harzburgite residue. The mantle sources for intraplate picritic magmas, on both the continents and in the oceans, are richer in Fe and Ti than the lherzolite mantle source which gives rise to the picritic magmas associated with plate boundaries. Moreover, the major and trace element compositions of hot spot picrites, such as those of Hawaii, present a paradox. The behaviour of the heavy rare earth elements has convinced geochemists (Hofmann et al., 1984; Frey and Roden, 1987) that residual garnet remains in their mantle source, yet melting experiments on both Hawaiian picrites (Eggins, 1992a) and model lherzolite mantles (Hirose and Kushiro, 1993) indicate that magmas of Hawaiian picrite composition have equilibrated with a Ca and Al-depleted harzburgite residue. The Hawaiian paradox is unresolved at present, but may indicate that garnet-equilibrated magmas have been forced to re-equilibrate with depleted harzburgite mantle or that the breakdown of amphibole in their mantle source resulted in melting under hydrous conditions. Alternatively, the existence of the paradox could be interpreted as evidence that hot spot picritic magmas are derived from piclogitic rather than lherzolitic mantle source regions (Anderson, 1984).