Regional–residual separation of bathymetry and revised estimates of Hawaii plume flux

Observations of the temporal variations in the volume flux of a plume can provide useful constraints on geodynamic models of plumes and plume-plate interactions. Furthermore, they can be compared with observations at other plumes and may be analysed further to understand the nature and cause of the variations. The volume plume flux is typically derived from a sum of edifice and compensation root volumes. The former can be obtained via the application of regional–residual separation procedures that split the observed relief into regional (swell) and residual (edifice) components, while the latter is generally inferred from the former using local (Airy) or regional (flexural) compensation models. Most regional–residual techniques used in past studies give non-unique results and provide no estimates of the uncertainty in the separation, which impacts the significance of the results. Here, the optimal robust separator (ORS) method achieves a unique separation for the swell and edifice components of the Hawaiian Ridge and furthermore obtain confidence bounds on the total volume flux. A fast spectral method for plate flexure with different edifice and infill densities is used to determine compensation volumes. Although my flux estimates have assigned confidence bounds, these are much smaller than the flux estimates themselves. A comparison of my new results to published volume flux curves shows that my revised flux estimates are lower by a factor of 2–3. Reproducing the prior higher results demonstrates that these discrepancies appear to be related to shortcomings in the implementation of the methodology used in the separation. The variability in the Hawaiian plume flux occurs at two different time scales: A short (1–2 Myr) periodicity related to the spacing of islands and seamounts, which ultimately is related to plume-plate flexural interactions, and a much longer (10–15 Myr) periodicity that may be related to plate kinematic changes. Superimposed on these trends may be an exponential increase towards more recent times, but this trend may also be explained by a higher flux level during the period when the plume was positioned beneath the relatively younger lithospheric segment bracketed by the Murray and Molokai fracture zones. Landslides and erosion of the edifice may imply an underestimation of total volumes by 5–10 per cent. The main uncertainty facing studies of plume flux is related to the unknown quantity of magmatic underplating.