Publications

Image distortion due to gas-filled shallow channels poses challenges to velocity model building. The slow velocity and strong absorption from the gas accumulation cause complex velocity and absorption variations in the overburden. Both the velocity and quality factor (Q) models need to be accurately estimated for proper imaging. Full-waveform inversion (FWI) has established its capability for high-resolution and high-fidelity velocity model building, while Q estimation for shallow channels is still a challenge. Ray-based Q tomography, using frequency-dependent amplitude decay along reflection ray paths to estimate Q, works well for less complex areas but struggles in complex shallow areas due to limited offset information and resolution. Moreover, Q tomography often enters at the later stages of the velocity model building flow with the purpose of enhancing final imaging rather than velocity estimation, meaning systematic velocity errors might already exist. In this case study in the South Mahakam area of offshore Indonesia, multiple layers of gas-filled channels with sharp velocity contrasts and strong absorption exist from the water bottom down to 600 meters. The target reservoir, at a depth of 3.5 kilometres, experiences depth uncertainties related to the complex overlying geology. We employed a visco-acoustic full-waveform inversion (Q-FWI) approach to jointly estimate the velocity and Q models. Q-FWI can invert for high-resolution velocity and Q models of the shallow channel system, providing superior imaging when using Q-compensated pre-stack depth migration (Q-PSDM). The cumulative improvements help to increase interpretation confidence and mitigate future drilling risks.