Estimates of the signal band of the Blackfoot broad-band data

Gary F. Margrave

A unique seismic dataset was acquired over the Blackfoot field (owned by PanCanadian Petroleum Ltd) in the summer of 1995. The acquisition was designed to collect a dataset which would allow a thorough analysis of the factors controlling the bandwidth of seismic data and, hopefully, would lead to new strategies for extending that bandwidth. The dynamite source was recorded simultaneously into four distinct receiver configurations: conventional 6 phone (10 Hz) vertical component strings, single 10 Hz 3 component phones, single 4.5 Hz 3 component phones, and 2 Hz 2 component phones. This work presents a spectral analysis of the data, both raw and processed, to compare the performance of the different receivers, to identify the derived signal band of the processed data and to estimate the potential signal band on the raw data. It is found that the different receiver types performed in a broadly similar fashion though the low frequency models recorded markedly more power at the low end. No high end signal distortion was observed on any model. The signal band of the processed vertical components, as defined by f-x phase coherence, tops out near 100 Hz in the shallow section and decays to 75 Hz of so near basement. Corner frequencies estimated from the raw data give similar results indicating that the data processing has fully extracted the available signal. On the radial components, the processed data tops out in the 30-40 Hz range while the raw data show corner frequencies near 65 Hz. Either the data processing has failed to image the upper P-SV frequencies or P-P energy has leaked onto the radial components. On the low end, all processed data show f-x phase coherence to well less than 5 Hz. It is anticipated that this represents extended signal band and that impedance inversions should benefit from it.