Time-lapse full-waveform inversion for ocean-bottom node seismic data with seawater velocity changes

Xin Fu, Kristopher A. Innanen

As a powerful tool for 4D seismic data inversion to monitor subsurface reservoir changes and/or CO2 storage, full waveform inversion (FWI) has the ability of high-resolution imaging of physical properties for subsurface media, and it can solve the problem of non-repeatable receiver/source positions in time-lapse seismic surveys. In this report, we develop a three-stage time-lapse (TTS) FWI strategy for ocean-bottom node seismic data, in which the first stage is to use FWI to estimate the seawater velocities in the baseline model and the monitor model, respectively; the second stage is to obtain a relatively good common starting model that is close to the final inversion result, to guide the next baseline FWI and monitor FWI to converge to local minima that are closing to each other; the third stage is to employ the better starting models acquired in the second stage to carry out the final convergence and reflect the time-lapse differences. The tests using synthetic data obtained from acoustic models with different levels of seawater velocity changes have demonstrated the feasibility and stability of our new method.