Efficient source-independent Q-compensated least-squares reverse time migration with LNCC imaging condition

Source-independent Q-compensated least-squares reverse time migration based on the convolutional misfit function constitutes an amplitude-preserving methodology that effectively compensates for seismic attenuation and alleviates the constraints imposed by the source wavelet. Nevertheless, its conventional implementation, which constructs the gradient through the cross-correlation between background wavefield and adjoint wavefield propagating in opposite directions, incurs exorbitant storage and computational expenses. To ease the computational and storage pressures, we developed an efficient source-independent Q-compensated scheme by introducing a local Nyquist cross-correlation imaging condition to formulate gradient. Instead of employing entire wavefields for migration, the local Nyquist cross-correlation imaging condition, in combination with the Nyquist rate, adopts only the local wavefields around the excitation amplitude time. Consequently, the proposed scheme considerably diminishes the storage requirement as well as the additional time resulting from frequent input-output operations, thereby enhancing computational efficiency. Numerical examples conducted on the 2D layered model, Marmousi model, field data, and the 3D Overthrust model reveal that the proposed scheme is capable of attaining imaging accuracy comparable to that of the conventional scheme, while exhibiting superior storage and computing performance, and possesses higher feasibility in 3D applications.