Interpolation of Near Offsets with Multiples and Prediction-Error Filters

Interpolation of near offsets is an important issue in marine surveys. Pseudo-primary data is created from the Sigsbee 2B model using surface-consistent cross-correlation. This creates pseudo-data in near offsets that were missing from the original data. A non-stationary prediction-error filter is estimated on the pseudo-primaries, and is used to interpolate the missing near offsets in the original data, with promising results. The prediction-error filter can also be used to gauge the benefit of the pseudo-primaries by convolving it with the recorded data.