Application of seismic interferometry by Multidimensional Deconvolution to USArray data

Seismic interferometry by crosscorrelation has found wide application over the last decade. With which, the Green’s function is retrieved between two receiver positions. Because the bulk of the ambient seismic field consists of surface wave energy, most applications in seismology utilize the surface wave part of the Green’s function. The correct reconstruction of interferometric (surface wave) responses, however, relies on a number of assumptions to be fulfilled. Violation of these assumptions reduces the accuracy of the retrieved responses. Seismic interferometry by multidimensional deconvolution (MDD) has been shown to improve the retrieved response on synthetic data. A point-spread function (PSF) computed from the same ambient noise is deconvolved from the response acquired by crosscorrelation and the result is a better focused response. In this thesis, it is demonstrated with numerical modeling that seismic interferometry by MDD yields better results than seismic interferometry by crosscorrelation. Moreover, a different formulation of the underlying theory is considered. The new formulation obviates the need to separate in-and outgoing wavefields, which is a drawback of the original formulation. The MDD method resulting from the new formulation is also numerically tested and subsequently applied to the data recorded by a passive deployment of broadband sensors (Transportable USArray) along the east of the continental USA. Limited response improvement is achieved mainly due to two reasons: receivers aparature (large nominal receiver spacing/limited width and height of receivers array) and illumination gaps.