Comments on stacking noisy data
John C. Bancroft
Stacking assumes that we have seismic traces that are contaminated with noise, and that the sum of similar traces divided by the number of traces will reduce the noise by the square-root of the number of traces. This stacking process assumes the standard deviation (SD) of the noise is similar in each trace. In applications where the SD of the noise varies from trace to trace, some form of weighing each trace by the inverse of its SD may improve the overall quality of the stack. In prestack migration gathers, where the noise is much greater than the data in a trace, balancing the amplitudes by dividing by the square-root of the fold may improve the stack.