Report 4 Acquisition Footprint Removal by Least Squares Migration

Least squares migration (LSM) is applied to synthetic data with the purpose of removing the ”acquisition footprint” noise from migrated sections. Results for simple point scatterer and meandering stream channel models suggest that LSM can significantly attenuate the acquisition footprint noise. For comparison, the corresponding Kirchhoff migrated sections show noticeable artifacts due to a coarse sampling of the wavefield. For the Overthrust synthetic data provided by the Society of Exploration Geophysicists, the LSM method encountered unexpected problems. In these data there are strong reflections that emanate from a deep layer and reverberate off the free surface. This type of multiple is not accounted for in the Kirchhoff forward modeling operator, and so confuses the iterative model updates. This confusion tends to spoil parts of the migrated image around the strong reflector. Thus, it appears that an effective use of the LSM method must use modeling operators that take into account strong multiple generation. However, muting the multiples allowed the undamped LSM to produce images with a higher interface resolution than the standard migration result, but with a much greater amount of high frequency noise. For the Gulf of Mexico field data, the undamped poststack LSM images were of noticeably higher resolution than the standard migration image. Faults were better resolved and the detail in layering was more pronounced with the LSM images. However, the LSM images contained more high frequency noise than the standard migration images. The preliminary conclusion is that poststack undamped LSM appears to be a somewhat robust imaging method, except for the introduction of high frequency noise at later iterations. It is expected that a regularized LSM procedure will eliminate