Inversion of post-stack seismic data into acoustic impedance is a classic trace-based process (Russell and Hampson, 1991). The usual assumption behind post-stack inversion methods is that a seismic trace in a stacked section satisfies the convolutional equation, which can be written as
where is the seismic trace, is the earth's normal incidence reflectivity, and is the seismic wavelet. According to equation 1, by deconvolving the seismic wavelet, one can acquire the earth's normal incidence reflectivity, which, in turn, is related to acoustic impedance through the recursive equation (Lindseth, 1979)
In practice, when the subsurface shows dipping layers, the convolutional model no longer holds true. Neither will the equation that relates earth's normal-incidence reflectivity to acoustic impedance (equation 2), because seismic traces in this case cannot be considered as simple 1D normal-incidence seismograms.
In order to improve the accuracy of seismic-derived acoustic impedance, I propose to employ the stratigraphic coordinate system (Karimi and Fomel, 2014,2011), in which the convolutional model assumption is more accurate.