Today's world is still quite far from overcoming its dependence on hydrocarbons. Unfortunately, finding hydrocarbons is getting more and more difficult. They are no longer found by simply looking for surface anticlines. We now look for hydrocarbons deep within the earth in stratigraphic or structural traps (). In this thesis, I am concerned with improving our ability to create images of complex structural traps using seismic data.
A good location for a structural trap is at the edges of salt bodies in the subsurface. These salt bodies are often found under deep water, making it economically unfeasible to just drop a well (wildcatting). Therefore, we want to image the subsurface surrounding the salt as well as possible. Unfortunately, salt has a very high seismic velocity (often twice as high as the surrounding subsurface) that significantly reduces the amount of seismic energy that gets through to the surrounding rock (Figure , seismic energy is represented by rays). This energy may be directed outside of the bounds of the seismic survey (Figure ), or become evanescent (die out) at the salt boundaries. This poor illumination makes it very difficult to image the potential hydrocarbon traps.
Imaging the subsurface is commonly done with some type of migration. Migration tries to move the seismic energy recorded at the earth's surface back to where it belongs in the subsurface, thereby creating an image of the subsurface. Poor illumination will cause ``shadow zones'' in the resulting image, in the areas where we hope to find the potential hydrocarbon traps. These shadow zones are clear in Figure , which shows the result of trying to migrate the seismic data collected over the subsurface shown in Figure . Shadow zones often contain some seismic signal that may be recovered with better imaging techniques. However, some of the information we would need to fill in these areas does not exist in the data at all.