Introduction

When using passive seismic data to detect and locate sources of seismic energy related to various endogenous processes the typical situation is that these sources have to be determined against a strong background, usually of surface origin. Vanyan and Cole (1990) tried to cope with this problem using a least squares method for subtracting unwanted signals in frequency domain. Here I give some formal basis for applying this method to passive data, describe a more general multichannel approach, and show the results of applying both methods to synthetic datasets.

A detailed description of the least squares method and its application to conventional seismic data can be found, for example, in Jenkins and Watts (1968) and White (1973). The algorithm used to locate sources was computing the semblance coefficient along moveout trajectories for all possible source locations in a 3-D grid (Nikolaev & Troitskiy,1987). Synthetic datasets were created for the seismic array used in the SEP drill-bit source experiment (Cole, 1991), so the synthetics model a realistic case.