Inferring Mantle From Basalt Composition

Abstract:

Isotope ratios in oceanic basalts, first reported by Gast and co-workers 50 years ago, are unique tracers of mantle composition, because they are expected to mirror the composition of their mantle sources. While the latter is certainly true for homogeneous sources, the plethora of studies over the last 50 years have shown that mantle sources are isotopically heterogeneous on different length scales. Isotopic differences exist between basalts from different ocean basins, volcanoes of individual ocean islands, lava flows of a single volcano, and even in µm sized melt inclusions in a single mineral grain. Diffusion, which acts to homogenize isotopic heterogeneity over Gyr timescales, limits the length scale of isotopic heterogeneity in the mantle to anywhere between several mm to 10s of meters. Melting regions, however, are typically several 100 km wide and up to 100 km deep. The scale of melting is thus generally orders of magnitude larger than the scale of isotopic heterogeneity. How partial melts mix during melting, melt transport, and melt storage then inevitably influences how isotopic heterogeneity is conveyed from source to melt. The isotopic composition of oceanic basalts hence provides an integrated signal of isotopically diverse melts. Recent mixing models and observed isotopic differences between source (abyssal peridotites) and melts (MORB) show that the range of isotopic heterogeneity of erupted melts need NOT directly reflect that of their source(s), nor need observed isotopic endmembers in source and melts be congruent. Many geochemical models, however, implicitly assume equivalence of source and melt composition. Especially when attempting to infer spatial patterns of isotopic heterogeneity in the mantle from those observed in erupted melts, or for linking isotopic diversity to geophysical structures in the mantle requires a more profound understanding to what extent erupted melts represent the isotopic composition of their mantle sources.