STATISTICAL OPTIMIZATION OF SEISMIC HOLOGRAPHY FOR ARRAY DATA PROCESSINGE. A. Gulko, A. F. Kushnir, and P. A. Troitskiy
Mapping sources of seismic radiation due to scattering of body waves at sharp inhomogeneities within the Earth is treated in this paper as a statistical estimation problem. The case is studied where the experimental data are records of a small-aperture seismic array providing a high coherence of body wave signals. By scanning all points of a given volume in the medium the problem is reduced to estimating the power of seismic radiation from a given point. Several estimating algorithms are proposed, most of them depending on matrix power spectral density (MPSD) of body wave signals at array sensors. Only those algorithms have practical value whose performance is not impaired when the true MPSD is replaced by its sample estimate. These algorithms evaluate the power of scattered radiation and suppress coherent body waves which interfere with scattered waves and hinder estimation of its power. Synthetic data based on real records of NORSAR array are generated and analyzed by the algorithms. These mapping algorithms are shown to be preferable to methods using conventional beamforming.