Back projection using data from extended arrays, such as USarray, has become a powerful tool for examining the high frequency radiation from major earthquakes. Energy localisation is improved when data from more than one array is employed, and so far spatial resolution has been estimated by simulation.
Fortunately, it is possible to build up the spatial resolution for back-projection from an arbitrary distribution of stations by using the configuration of the isochrons from each station passing through the target point in 3-D. It is therefore possible to test out the potential spatial resolution for a station configuration in areas of significance and thereby maximise potential resolution. As in hypocentre estimation the quality of the spatial operator is improved with good azimuthal control. As a consequence back-projection from a limited number of well-distributed stations or small arrays can be competitive with the use of much larger arrays with a limited azimuth span. This simplifies the task of near real-time processing to track the evolution of major events once a hypocentre has been estimated.
We illustrate the potential for source recovery in 3-D for the very deep Mw 8.2 event in the Sea of Okhotsk. We also investigate appropriate station configuration for subduction zone mapping for the Indonesian region where there is considerable potential for megathrust events, but the epicentral distance is too large to allow direct use of North American stations.
