Cross‐correlation weighted migration: towards high‐resolution mapping of mantle heterogeneities

SUMMARY We present a new type of Born migration scheme to obtain a high‐resolution image of small‐scale scatterers in the mantle. Seismic migration is a technique that projects scattered energy back to its origin by shifting and stacking seismograms using theoretical traveltimes. Migration methods in global seismology, because of the limited distribution of sources and receivers, tend to suffer from considerable smearing owing to isochronal artefacts. This type of artefacts is caused by migration grids that have the set of theoretical traveltimes (from a given source to all receivers) similar to that for a true scattering point. Previously, the information of slowness and backazimuth was explicitly explored to reduce such migration artefacts. Here we propose a simpler and more computationally efficient method but with similar effects, in which the coherency of shifted seismograms measured by cross‐correlation is exploited to suppress isochronal artefacts. The new migration scheme makes it possible to compute a large number of migrated images efficiently and thus evaluate the robustness of observed scattering objects by statistical approaches such as bootstrap resampling.