Passive seismic imaging applied to synthetic data

Introduction

Conventional seismic reflection methodology relies on having a source of seismic waves at the surface and studying the reflections from impedance contrasts in the earth. Ambient noise present in the subsurface is also reflected by impedance contrasts in the same way. Therefore, do we need the source or can we determine earth structure simply by listening to the background noise for a long enough time ?

With 4-D seismic surveys aimed at monitoring fluid movements becoming more and more popular, a number of oil fields are having geophone arrays laid out permanently to reduce acquisition costs. These geophones will typically only record while a survey is being shot, and for most of their life they will be turned off. However, if the geophones are left recording while surveys are not being shot, we hypothesize the information contained in the background seismic energy could be used for imaging the subsurface between main surveys.

For this to be realized a technique has to be developed that: firstly, is able to extract the useful information from the background noise; and secondly, is able to do this quickly enough, ideally in real-time, so the huge amounts of raw data that would be recorded would not have to be stored.

With this in mind, a technique for creating `pseudo-cmp gathers' out of background noise traces using cross-correlation is explored and tested on a variety of synthetic models.

Passive seismic imaging applied to synthetic data