A statistical array processing approach for elastic wavefield data

Three-component seismic array data provide a multidimensional, spatially and temporally unaliased sampling of the seismic wavefield and the potential for a nearly comprehensive decomposition and characterization of the observed wavefield based on its frequency, wavenumber, and polarization characteristics. In this paper I outline a statistical array processing approach for three-component seismic array data that allows us to exploit and fully utilize the abundance of information provided by these unique data. My approach entails a succession of principal component analysis, factor analysis, and principal component analysis. The initial principal component analysis is conducted in an effort to separate the coherent and random components of the wavefield into orthogonal subspaces. The ensuing factor analysis allows us to isolate superimposed signals and estimate their frequency-wavenumber characteristics. The final principal component analysis provides estimates of the signals´ polarization characteristics. This adaptive formulation provides high-resolution estimates of the signals´ frequency, wavenumber, and polarization characteristics