Theory of Multisource Crosstalk Reduction by Phase‐Encoded Statics

Abstract

SUMMARY Formulas are derived that relate the strength of the crosstalk noise in supergather migration images to the variance of time, amplitude and polarity shifts in encoding functions. A supergather migration image is computed by migrating an encoded supergather, where the supergather is formed by stacking a large number of encoded shot gathers. Analysis reveals that for temporal source static shifts in each shot gather, the crosstalk noise is exponentially reduced with increasing variance of the static shift and the square of source frequency. This is not too surprising because larger time shifts lead to less correlation between traces in different shot gathers, and so should tend to reduce the crosstalk noise. Analysis also reveals that combining both polarity and time statics is a superior encoding strategy compared to using either polarity statics or time statics alone. Signal‐to‐noise (SNR) estimates show that for a standard migration image and for an image computed by migrating a phase‐encoded supergather; here, G is the number of traces in a shot gather, I is the number of stacking iterations in the supergather and S is the number of encoded/blended shot gathers that comprise the supergather. If the supergather can be uniformly divided up into Q unique sub‐supergathers, then the resulting SNR of the final image is , which means that we can enhance image quality but at the expense of Q times more cost. The importance of these formulas is that they provide a precise understanding between different phase encoding strategies and image quality. Finally, we show that iterative migration of phase‐encoded supergathers is a special case of passive seismic interferometry. We suggest that the crosstalk noise formulas can be helpful in designing optimal strategies for passive seismic interferometry and efficient extraction of Green's functions from simulated supergathers.

Efficient full waveform inversion using the excitation representation of the source wavefield

“…On the other hand, the need for more sophisticated algorithms incorporating all kinds of physics (anisotropy, elasticity, viscosity) results in 1582 M. Kalita and T. Alkhalifah extra computational costs in FWI. However, source-encoding schemes (Krebs et al 2009;Choi & Alkhalifah 2011;Schuster et al 2011) have drastically reduced the computational cost, albeit with a trade-off in the accuracy of physics.…”

Plane-wave least-squares reverse-time migration

“…For 30 iterations, the computational cost of the plane-wave LSRTM is about 12 times of the conventional RTM. When high computational efficiency is in demand, the LSRTM can be performed with the dynamic encoding approach of Krebs et al (2009);Schuster et al (2011), where one plane-wave gather is used for each iteration and the ray paramerater p (corresponding to surface shooting angle) is dynamically changed from one iteration to another. 5(d)).…”

Multiscale phase inversion of seismic data

“…The cross-talk can be negligible compared to the summation over the |D(g|s) obs j | 2 δτ (s|g) 2 j terms if the crosstalk phase for each pair of j, k indices is a random variable (Schuster et al, 2011) and N, N g , N s 0 are large. In this case the misfit function reduces to that for wave equation reflection tomography:…”