Secondary Hawaiian volcanism formed by flexural arch decompression

[ 1] This paper explains two forms of secondary Hawaiian volcanism ( rejuvenated onshore and Hawaiian Arch offshore) as a direct consequence of lithospheric flexural uplift that surrounds growing shield volcanoes. This uplift decompresses the underlying asthenosphere, which is assumed to be chemically and isotopically heterogeneous, near its solidus, and derived from the Hawaiian mantle plume. Lithospheric uplift is modeled as the axisymmetric response of an elastic plate to a ( volcanic) point load that grows linearly in time. To model flow in the asthenosphere, the rate of flexure of the lithosphere is taken as the upper boundary condition on an isoviscous, incompressible, fluid half-space. This model successfully explains the majority of spatial gaps between secondary and active shield volcanism due to the flexing of a lithospheric plate with an effective elastic thickness of 25 - 35 km. Second, this work demonstrates that the flexural model can produce realistic magmatic fluxes if magma is focused toward individual eruption sites from the mantle over an area two to ten times the eruption area. Next, this model addresses the isotopic distinction between secondary and shield lavas. In this model, the same heterogeneous mantle plume feeds the secondary and shield lavas, but the compositional components are sampled by melting at rates that differ between the two settings. Flexural decompression mostly melts the component that begins melting shallowest, which we assume to be depleted in incompatible elements with relatively low Sr-87/Sr-86 and high Nd-143/Nd-144. Melting in the center of a mantle plume is assumed to generate shield volcanism and is predicted to mostly melt components that begin melting deepest, which we assume to be enriched in incompatible elements with higher Sr-87/Sr-86 and lower Nd-143/Nd-144. Thus the models successfully predict the observed mean Sr-87/Sr-86 and Nd-143/Nd-144 isotopic compositions of secondary and shield lavas to arise out of the melting process alone. The fourth feature addressed is that secondary lavas are alkalic, having formed from relatively low extents of partial melting, and shield lavas are dominantly tholeiitic, consistent with more extensive partial melting. Indeed, models predict lower mean extent of melting for secondary lavas compared to shield lavas if the source material, which is mostly peridotite, contains at least some pyroxenite. Results show that model predictions are consistent with the geochemical constraints for a range of reasonable starting mantle compositions, lithospheric thicknesses, and plume temperatures.