SLICING THE EARTH: A LAYER-STRIPPING METHOD EMPLOYING A CAUSALITY-BASED IMAGING CONDITION

摘要

We formulate the theory for a direct nonlinear seismic inversion method in the acoustic approximation. It is a completely data-driven method, aiming at the determination of subsurface properties directly from the data. The theory is presented for the full three-dimensional, laterally varying case. For this situation we derive a layer replacement method based on the reciprocity theorem and we derive an imaging condition based on causality. Next we simplify the theory for the one-dimensional case and give some synthetic results for this case. We explain how the limited bandwidth of seismic data influences our method. We propose a solution method to deal with the lack of low frequencies in the data. This method uses the absolute value of the data. Finally we present a synthetic inversion example for a laterally varying earth model using common-midpoint techniques.