Appraising the Reliability of Scattered Wavefield Imaging: Application to the 410 km Discontinuity

Passive seismic array wavefield imaging methods currently suffer from a lack of methods to appraise the reliability of the results. This is a more important issue than it is in the seismic reflection method because of variations in illumination created by the uncontrolled source geometry. We introduce three approaches that build on error methods commonly used in travel time tomography: (1) isolated point scatters used to appraise point resolution; (2) a checkboard test to appraise resolution on a single surface; (3) hypothesis testing using forward simulation of a model to be tested. To accomplish this we developed an elastic, single scattering approximation forward modeling program that duplicates the exact geometry of the data being simulated. We implement the synthetics by discretizing the integral equation for forward P-to-S scattering based on the Born approximation. This allows simulation of P to S conversions from a surface of any shape by a surface integral of point sources. We apply the new algorithm to appraise results from imaging the 410 km discontinuity with USArray data estimated with a recently developed 3D plane wave migration method. Our results suggest that the 410 discontinuity has topography and lateral variations in scattering strength that are consistent with mineral physics models of the 410 discontinuity. The implication is that the 410 km discontinuity is not really a discontinuity but is a rough surface at the scale of short-period, teleseismic P waves.