Advancements in the technique of low fold three dimensional seismic reflection surveying.

Abstract

Three dimensional (3-D) seismic reflection surveying is accepted as the preferred method for imaging complex geology for proving and developing commercial oil and gas fields. However, the cost of 3-D seismic recording and processing is substantial and often can be as expensive as the cost of production drilling. This is particularly the case for land oil field development, where the cost of 3-D surveying is often unacceptably high. Such high costs also restrict its application in coal exploration, where 2-D seismic methods have long been accepted.During the early 1980's, a low fold technique for recording land 3-D data was devised which offered significant cost savings. The technique was adapted by the author for land 3-D surveying over coal fields. Inherent in the technique was a requirement that the data must have a high signal-to-noise ratio, which is not generally the case in land surveying due to the presence of strong source generated surface wave noise. A further major impediment to the technique was its inability to perform an acceptable stacking velocity analysis because of the low number of seismic traces generated. This thesis defines three data collection and processing advancements in low fold 3-D technology which go some way towards resolving these impediments.The first advancement is a method to enhance the signal-to-noise ratio of the stacked seismic data, and consists of a Radon-based transform which stacks shot domain data along a curved trajectory, thereby attenuating surface waves on swath recorded data. This transform is termed the 'Radial Transform' of 3-D data.The second advancement is a statics method to improve the stacked image from a low number of input traces. The method uses the concept that if both the reflected and refracted waves pass through a weathering layer with very similar travel paths, then static corrections to remove the effects of weathering variations on the refraction travel times would be very similar to those required for the reflections. This method, which was patented, is used equally for both 2-D and 3-D field data, and is regularly used in high resolution seismic processing for coal at Curtin University.The third advancement resolves the problem of azimuthal variation of stacking velocities. By predicting the true reflector dip and its azimuth, apparent dip can be included in the normal moveout equation, which is named the Generalized Moveout equation. The requirement for an azimuthally dependent stacking velocity is then no longer an impediment in low fold 3-D processing of coal data.After developing these transforms and applying them to synthetic data, they were tested with success on modelled field data. All field data used within this thesis were either recorded in the field by the author, or were produced with a physical modelling system, which was built by the author at the University of Houston and later at Curtin University.Results indicate that the procedures described in this thesis enable the low fold 3-D technique to be used as a viable method for recording seismic data when survey economics are a major issue. Furthermore, all three advancements are suitable for application in conventional two dimensional (2-D) and swath seismic surveying.