Impact of source and receiver distributions on imaging of dipping reflectors with passive reflection seismic interferometry

Abstract

SUMMARY Passive reflection seismic interferometry (PRSI) facilitates imaging the subsurface structure using passive sources but according to the general theory, the target needs to be illuminated equally from all directions to obtain a kinematically correct result. In practice, this requirement is almost impossible to meet which can introduce artefacts into the PRSI results. Our study was motivated by an unsuccessful attempt to image a couple of known dipping reflectors by applying PRSI to a data set of local microearthquakes clustering around the glacially triggered Burträsk fault. Dipping reflectors are a special case since they introduce a directionality into the seismic-interferometry problem that makes the results especially sensitive to the source azimuths. To investigate which source distributions are favourable in such a case and to study the range of artefacts occurring, we analyse a number of acoustic and elastic synthetic data sets calculated using a simple model of a dipping fault. Our results show that the main contribution for imaging such a fault with PRSI comes from sources in the hangingwall whereas contributions from the footwall are often weak and kinematically incorrect. The type and position of the occurring artefacts depend upon both the source azimuth and the type of modelling. In the acoustic case, the main artefact is a gently dipping reflection caused by insufficient cancellation of the direct reflection at the fault. In the elastic case, the artefacts are dominated by a set of both gently and steeply dipping reflections related to P–S converted waves. These artefacts are present even for ideal illumination due to the use of source records containing both P- and S-wave contributions. During interpretation, it is essential to be able to distinguish between physically meaningful reflections and artefacts. We found that both acoustic and elastic artefacts stack best at lower than expected normal moveout velocities. If data quality is insufficient for velocity analysis, our results can serve as a reference point for the interpretation of dipping features in PRSI images.