Resolution of prominent crustal scatterers near the Noress small-aperture array

Abstract

Our goal is to identify prominent scatterers near seismic arrays so that the secondary phases they generate can be anticipated and recognized as being of local, not teleseismic, origin. Toward this end we have developed a technique to scan seismic coda recorded by arrays for phases generated locally by scattering of teleseismic energy from large topographical and crustal heterogeneities. In this paper we test the technique by using the NORESS small-aperture array in southern Norway to image free surface secondary Rg sources excited by incoming teleseismic P waves. In essence, our technique suppresses the teleseismic energy and migrates the residual array records to image-coherent local sources. To assess the azimuthal and radial resolution that the NORESS array provides we have analysed point-source synthetics. A preliminary analysis of recorded data provides an initial view of the scatterers in the vicinity of NORESS and identifies limitations of imaging using the full coda of individual events. Single events can be used to image scatterers but more robust images, with higher azimuthal resolution, can be obtained by using a broadly distributed suite of events. Radial resolution of scatterers can be enhanced by deconvolution. A composite image generated by simultaneously considering a suite of 31 deconvolved teleseisms indicates that two secondary Rg sources are present at, or near, the free surface near NORESS. One, 27 km to the south-west of the array, appears to be related to dramatic topographic relief at the south-west edge of Lake Mjosa. Another, roughly 10 km to the east of the array is judged by a bootstrap-resampling analysis to be only marginally significant. This apparent source does not coincide with obvious topography and is likely buried. Although the time delay that separates the scattered and ‘parent’ phases depends somewhat on the angle of approach of the incoming waves, in general these secondary sources are contributing wave trains to the seismic coda that begin ˜ 4 and 11s, respectively, after the primary onset.