The fate of slabs inferred from seismic tomography and 130 million years of subduction

Abstract

The volume and location of subducted plate, since 130 Ma, are reconstructed from magnetic anomalies and updated finite rotation parameters describing relative motions in the ocean basins and between plates and hotspots. The area of subducted plate is calculated from the relative motions between overriding and overridden plates along the length of convergence in the hotspot reference frame in 5 Ma intervals. Ridge locations are reconstructed by rotating magnetic anomalies from their present to their former positions. The distances between trenches and ridges, at a certain point of time, in the spreading direction are thus estimated and are used to determine the ages of trenches, according to the spreading history recorded in present-day oceans. The thickness of the oceanic plate is obtained from the age based on the half-space cooling model. About 3.45 × 10 10 km 3 of oceanic plate has been subducted during the past 130 Ma, with maximum accumulation beneath Southeast Eurasia. At spherical harmonic degree l = 2, where subduction flux peaks, excellent correlation is found between four seismic tomographic models at the top of the lower mantle (about 800–1100 km deep) and predicted slab locations. For some seismic models excellent correlation is also found at other degrees in that depth range. Some models also have good correlations with predicted slab locations in the deep mantle at l = 2. Direct comparison between subduction and seismic tomographic patterns at 800–1100 km shows that subduction history correlates with tomography very well in terms of location and amplitude of anomalies. Cold downwellings, which may be related to slabs, appear to be trapped in the mesosphere, or middle mantle. Correlations between seismic tomography and subduction over different time periods support this conclusion. There are also some correlations in the upper mantle and in the deep lower mantle, although they are generally not as significant as those in the 800–1100 km depth range. There may be a significant boundary in the mantle near the 800–1100 km depth. From a geodynamic and chemical point of view, the lower mantle may start at a depth closer to 1000 km than to 670 km.