A VIEW OF THE MARBLE-CAKE MANTLE FROM THE SOUTHEAST INDIAN RIDGE

Abstract:

Along the Southeast Indian Ridge, variations in axial depth, crustal thickness, hydrothermal venting [1], basaltic major elements and U-series disequilibria [2] all indicate a west-to-east decrease in magma supply and mantle temperature from the Amsterdam-St. Paul hotspot to the Australian-Antarctic Discordance. Paired Hf-Pb isotopes in closely spaced glasses (5-10 km) from 81-100°E define two populations revealing compositional streaks in the upper mantle [3]. The number density of the streaks follows a Poisson distribution with a characteristic thickness of ~20 km. K/Ti and Na₈ do not correlate with Pb or Hf isotopes, and both isotopic domains encompass N- and E-MORB types indicating the variations represent mantle source heterogeneities. ³He/⁴He varies from 7.5 - 10.2 R_A, more than half the range in global MORB away from hotspot influence [4]. No systematic relationship exists between ³He/⁴He and Pb or Hf isotopes. A general negative correlation between K/Ti and Fe₈ for the SEIR resembles that for MORBs globally, with higher K/Ti associated with lower ³He/⁴He. Collectively the observations suggest the presence of lithologically heterogeneous mantle. Lower ³He/⁴He derives from a source containing a few percent pyroxenite or ecologite, while ³He/⁴He > 9 R_A arises from peridotite. Mantle convection has folded together distinct composite reservoirs of heterogeneous mantle, and stretched them into streaks that remain discernible units. The mantle ‘unit’ giving rise to each MORB sample represents a ‘mixture of mixtures’ with a multi-stage mixing history. Spectral analysis of the length scales of Hf, Pb and He isotopic variability allows a visual representation of this upper mantle ‘texture’. The dominant length scales reflect large (1000, 500 km) and regional scale (100-150 km) structures in mantle flow, and sampling of heterogeneities during partial melting (20-30 km).

1-Baker et al., doi:10.1002/2014GC005344; 2-Russo et al., doi:10.1016/j.epsl.2008.11.016; 3-Hanan et al., doi:10.1016/j.epsl.2013.05.028; 4-Graham et al., doi:10.1002/2014GC005264