Is a Cretaceous Superplume in Pacific Ocean Necessary?

Abstract:

The mid-Cretaceous was a time when large igneous provinces (LIPs) including Manihiki, Hikurangi and Ontong-Java plateaus were produced. These LIPs can be jointed back together by plate reconstruction at $\sim120$ Ma, suggesting that a dramatic volcanic event occurred at that time. The debate over whether the volcanism was caused by a superplume, plate reorganizations, or other tectonic controlled events is still going on. We address the necessity of a superplume in Pacific Ocean with mantle convection simulations. We use time-dependent convection calculations using the finite element code CitcomS with a temperature-dependent Newtonian rheology and a layered viscosity structure. In order to couple the interplay of plate tectonics and mantle convection, we introduce plate velocities from 170 Ma to ``present day'' generated using GPlates as a time-dependent surface boundary condition. In calculations with a thermal superplume rising from core mantle boundary (CMB), we introduce tracers in the initial thermal perturbations just above the CMB to track the location of plume material.
Applying the plate reconstruction model of Seton (2012), the initial large melt region formed at $\sim120$ Ma in our mantle convection simulations is broken up into pieces, which end up close to the ``present day'' locations of Manihiki, Hukurangi, Ontong-Java and Caribbean plateaus. We get similar results using different initial-perturbations at the CMB located within 500 km of the projection of the triple junction at the surface onto the CMB, which implies that the location of initial thermal perturbations is somewhat flexible. We repeat the simulation with no initial thermal perturbation at the CMB to see if the plate evolution alone will allow for a large pulse of magma at the triple junction. While the simulation with plume rising from CMB produce significant melt in the upper mantle, the non-plume simulation controlled by plate tectonics alone does not generate melt in the upper mantle. We also compare geoid and dynamic topographic maps for non-plume and plume simulations.