Abstract

This work aims to present a comparison between the f-k and bandpass filtering methods in reducing ground roll energy in a land seismic dataset. The f-k filter used is based on the 2-D Fourier transform of the input data in the t-x domain. A triangular-shaped f-k filter was incorporated as the reject-band which is then applied to areas where the ground rolls are located without affecting the signal in the same area. The f-k filter attenuates the ground roll using the differences between noise and signal in two aspects: frequency and amplitude, with the data outside the window still intact. However in the bandpass filtering, the band pass filter of truncation frequencies of 5-15-55-65 Hz (10-60 Hz) was used to attenuate the ground roll, and this eliminates low‐frequency reflection signals which contribute to the character of reflection events which are very important in inversion for acoustic impedance and advanced seismic processing. This analysis is applied to the D-10 dataset, which demonstrates a successful performance using this comparison.