Nonstationary deconvolution as applied to the Blackfoot broadband survey

Alana R. Schoepp and Gary F. Margrave


Nonstationary deconvolution (NSD) is designed to approximately correct seismic data for source signature and anelastic attenuation. The amplitude spectrum of the forward operator is designed by calculating the time-variant amplitude spectrum of the input data and then smoothing it to remove reflectivity. The amplitude spectrum of the operator can be combined with a minimum-phase spectrum, estimated with the Hilbert transform of the natural logarithm of the amplitude spectrum. The forward operator is then inverted and applied to the input data with nonstationary filter techniques. NSD is a data-driven method and can be thought of as a combination of stationary deconvolution and inverse-Q filtering.

NSD has been developed in two different modes. The NSD operator can be designed and applied for each individual trace (trace-by-trace NSD), or a single operator can be designed from the ensemble and applied to each trace (profile-mode NSD). The advantage to profile-mode NSD is that it has greater statistical leverage because it uses many traces to design one operator. This can lead to improved lateral stability of reflectivity estimates in noisy data.

Nonstationary deconvolution was applied to the Blackfoot broadband survey and the results were compared to the results from combinations of gain and Wiener deconvolution, and Wiener deconvolution and inverse-Q filtering. The NSD result displayed relative enhanced vertical resolution and improved reflectivity character. In addition, it delineated the target, an incised channel, more clearly than the other methods.

View full article as PDF (1.68 Mb)