Regional Correlation Analysis between Seismic Heterogeneity in the Lower Mantle and Subduction in the Last 180 Myr: Implications for Mantle Dynamics and Rheology

Abstract

We have carried out a regional correlation analysis between the seismic structure of the lower mantle and the reconstructions of subduction sites in the past 180 Myr with the aim of estimating individual styles of slab motion over different parts of the earth. The correlation patterns obtained for three subduction branches (West Pacific, East Pacific and Alpine-Himalayan) are remarkably different. In the West Pacific, the subducting slabs tend to be stagnant beneath the 660-km discontinuity, while basically no subducted lithosphere has been detected below the depth of 1000 km. In contrast, the lithosphere subducted beneath the Americas seems to penetrate through the lower mantle continuously, showing correlation peaks at depth intervals of 800 -1100 km and 1900 -2500 km. In the AlpineHimalayan region, significant correlation has been found below the 660-km discontinuity for recent subduction and in the mid-mantle for subduction younger than 120 Myr. An increase in the correlation close to the core-mantle boundary nevertheless indicates that, under certain circumstances, the slabs can reach the bottom of the mantle in the West Pacific and in the Alpine-Himalayan regions as well. The correlation peak at a depth of around 1000 km is common to all the subduction branches. However, its depth rather varies for different subduction zones and, thus, it is not clear whether this correlation maximum may be associated with a global mid-mantle discontinuity.

A comparison of tomographic and geodynamic mantle models

“…We have, however, found that fast anomalies are imaged consistently in the midmantle where we would expect slabs in the absence of a long‐term barrier to flow at 660 km (Figures 14 and 17). This substantiates previous studies [ van der Hilst et al , 1997; Čížková et al , 1998; Bunge et al , 1998], and slab penetration is found to be a common phenomenon (Figure 18). As discussed, it can be expected that future flow models will do a better job in predicting slab locations and current discrepancies should lead to a better understanding of the nature of the subduction process.…”