Geochemical constraints on magma formation and transport processes
Abstract:Primitive basalts provide an invaluable probe of the mantle's thermo-chemical structure. What these samples show is that the Earth's interior is widely variable in its trace element, isotopic and even major element composition, on the km to the hemispherical scale. This heterogeneity has profound implications for not only the history of the solid Earth, but the oceans and atmosphere as well, as it represents ~4 billions of years of elemental transport back into the mantle via subduction recycling of oceanic crustal to mantle sections. Reconstructing planetary evolution through the volcanic record of mantle composition is therefore a primary aim of igneous geochemistry.
However, between the solid mantle and our chemical analyses lie a series of melt generation, aggregation and transport processes, themselves poorly understood, that are potentially critical in controlling the amplitude and style of chemical heterogeneity preserved in an erupted basalt. If these processes are also sensitive to mantle potential temperature, the degree of melting and the presence of lithological heterogeneity, then the geochemical record may not only be biased as a whole, but biased in a relative sense between different geodynamic settings: such a dichotomy may be represented by ocean islands and mid-ocean ridges.
Here we combine observational and modelling approaches to understand how varying conditions of melt generation and transport affect basalt chemical variability. Focusing first on Iceland, we combine new and existing melt inclusion data to investigate how chemical variability may be controlled by tectonic parameters (on versus off rift) and source enrichment. We find that on Iceland the key parameter controlling variability is enrichment, with the most enriched basalts preserving diminished variability compared with more depleted eruptions. However, on a larger scale enriched sources preserve the greatest variability: we see this both in terms of the greater variability of Iceland as a whole compared to that preserved in primitive Reykjanes Ridge melt inclusions, and from a statistical analysis of global mid-ocean ridge basalts. On the local scale enrichment-variability systematics can be understood in the context of simple statistical models of melt aggregation and mixing. On larger scales source dominates.