Introduction

In recent years, several studies have sought to obtain additional information from the parts of seismic records which are treated as obstacles in traditional studies. Examples of such data are records of microseismic noise and teleseismic coda waves. In the case of coda waves it is possible to detect irregularities in the structure of the Earth's crust, which act as sources and scatter teleseismic waves. Study of microseisms may also yield information about processes in the Earth's crust. Microseismic noise can be due to different sources, such as zones of anomalous temperature or pressure in oil and natural gas reservoirs (particularly during industrial exploitation) or active seismic faults. Seismic noise research can be applied to the monitoring of such processes.

In the department of experimental seismology at the Institute of Physics of the Earth (Moscow, USSR), methods for processing noise and noise-like data are being developed (Nikolaev and Troitskiy, 1987). One algorithm constructs a 3-D grid in the subsurface, and then computes the semblance coefficient for energy arriving from a source at each point in the grid. Using this method, records of coda waves and microseismic noise, both recorded by the NORSAR array in Norway, were processed. Also, in 1988, the same group conducted an experiment in Iceland to measure microseisms using a sparse 2-D array of geophones. The same semblance method was used to locate subsurface sources. The results are reviewed in this paper.

One of the problems inherent in such work is the presence of conflicting signals arriving from many noise sources. In particular, surface noise sources are relatively strong compared to endogenous sources or scattered energy. We discuss a frequency-domain least-squares method that can suppress the signal generated by strong surface source and improve the detection of weak subsurface sources.

At SEP, an experiment was conducted to record ambient noise (Nichols et al, 1989; Cole, 1989) using a compact, more densely-sampled array, with the goal of observing scattering of ambient noise energy and of energy from quarry blasts. Stacking-based methods were used to first examine the plane wavefield incident on the array, and then to look for scattering in a manner similar to that used in the Russian experiments. In this paper we use some of the techniques developed at SEP to supplement our review of the Russian work and the processing of the microseismic dataset from Iceland.