Tackling the Stationary Zero-phase Assumption

The assumption that the phase of seismic data is known, and is stationary with respect to time and space, is one of the few assumptions which underpin almost the entire spectrum of seismic interpretation methods. We demonstrate that the phase of 3D seismic data can be robustly estimated using an iterative kurtosis-based method. Our methodology is first applied to a noisy synthetic seismic volume and is shown to correct the non-stationary phase that is present. Results of the application of the method to a real seismic dataset indicate that non-stationary phase is present. We attempt to correct for this phase variation in 3D to leave only local apparent phase anomalies that may be attributable to geological variation. We believe that all seismic data should be subject to phase interrogation before applying techniques that require an assumption about phase to be made, and offer this method as a way to do so.