Accurate and automatic refraction statics in large 3D seismic dataset

Abstract

Inversion for refraction statics is a key part of three-dimensional (3D) reflection seismic processing. The present thesis has two primary goals directed toward improvement of refraction statics inversion. First, I attempt to improve the quality of the travel-time data right at the beginning of the processing sequence and before any inversion. Any error in the travel times or geometry caused during acquisition or processing would propagate into the resulting model and may harm the resulting image. To implement rigorous, model-independent data quality control, I view the first-arrival travel times as surfaces in 3D, which allows utilization of the travel-time reciprocity condition to check for errors in geometry and in first-arrival picking.

The second goal of this study is in development of a new inversion approach for refraction statics specifically for 3D seismic datasets. The first-break travel-times are decomposed by using a ô-p parameterization, which allows an automatic derivation of a high-quality initial subsurface model. This model is further improved by using accurate, multi-layer ray-tracing and inversion techniques to obtain accurate refraction statics. An iterative inversion scheme based on the Simultaneous Iterative Reconstruction Technique is utilized, and its performance is measured and discussed. To assess the quality of the inverse and establish the optimal grid sizes, I use several types of resolution tests. Finally, the surface consistent statics is calculated and applied to a real dataset from southern Saskatchewan. A comparison of the resulting statics model with statics calculated by using standard industry software is made, and the statics correction is incorporated in seismic processing.

An overall result of this study is in demonstration that the fully 3D, ô-p based travel-time inversion method works, is applicable to large seismic datasets, and results in detailed shallow subsurface models and reliable statics solutions. Several recommendations for extending and improving the proposed approaches are also made.