One primary concern in seismic imaging necessarily focuses on determination of the true subsurface medium. Clearly, the accuracy of this information significantly impacts all aspects of the exploration process. Even when we do not have a completely detailed visualization of what is below us a reasonable concept can provide guidelines for surface acquisition that improves subsurface imaging . The underlying Earth model strongly influences what we must do to migrate the data successfully and produce an optimum image.
Another primary concern focuses on which of the myriad available imaging algorithms has the best chance of producing the highest quality image. Making this choice requires an understanding of the most important such technology. Because algorithm development and implementation is a highly mathematical endeavor, acquiring this understanding can be quite difficult.
A third concern arises from the fact that, in general, the Earth does not respond well to high frequency sources. High frequency sound waves are absorbed rather quickly. Depending on rock type they penetrate only to a few thousand meters. On the other hand, low frequency sound waves are known to provide narrow bandwidth images at depths in excess of 30 or 40 kilometers.
Electromagnetic waves are frequently characterized by their penetration depth or their so called skin temperature. Although claims to the contrary abound, the skin temperature of most electromagnetic waves is only on the order of a few hundred meters. This means that to the extent possible, changes in magnetic parameters can be observed only from approximately half of this depth. This is far too shallow to be of much use in exploration. This lack of ability to penetrate deeply into the Earth's interior eliminates most high frequency sources, and strongly implies that we cannot use light, electromagnetic, or radar sources to measure and image the Earth at the depths of interest.
Because of these issues, our best option is to use relatively low frequency sound sources on the order of a few hundred Hertz. When higher frequency sources are routinely available, their responses will be easily incorporated into the general imaging workflow, but until that happens, we must focus on low frequency data sets to achieve our exploration goals. We now know that, from an inversion point of view, accurately determining the subsurface velocity is easier when the low frequency portion of the frequency band is full. High frequencies are certainly important but have much less impact on the velocity estimation problem then lower frequencies. While somewhat contrary to intuition, the importance of very low frequency data cannot be denied.
Perhaps the final concern in seismic imaging is having a clear understanding of how sound propagates. Given a decent understand of the types of rocks we may encounter, this concern can be resolved directly through seismic synthesis or modeling. The ability to generate realistic responses to practical and physical seismic sources should move us a long way down the path toward near optimal application of the entire imaging process
Given these simple concerns, this section attempts to use mathematically based formulations for digitally synthesizing seismic data in the hopes that the issues raised above can be clarified in a relatively simple and intuitive manner.
