In recent years, many important discoveries have been made in the marine deepwater hydrocarbon exploration in the South China Sea, which indicates the huge exploration potential of this area. However, the seismic prediction of deepwater reservoirs is very challenging because of the complex sedimentation, the ghost problem, and the low exploration level with sparse wells in deepwater areas. Conventional impedance inversion methods interpolate the low frequencies from well-log data with the constraints of interpreted horizons to fill in the frequency gap between the seismic velocity and seismic data and thereby recover the absolute impedance values that may be inaccurate and cause biased inversion results if wells are sparse and geology is complex. The variable-depth streamer seismic data contain the missing low frequencies and provide a new opportunity to remove the need to estimate the low-frequency components from well-log data. Therefore, we first developed a broadband seismic-driven impedance inversion approach using the seismic velocity as initial low-frequency model based on the Bayesian framework. The synthetic data example demonstrates that our broadband impedance inversion approach is of high resolution and it can automatically balance between the inversion resolution and stability. Then, we perform seismic sedimentology stratal slices on the broadband seismic data to analyze the depositional evolution history of the deepwater reservoirs. Finally, we combine the broadband amplitude stratal slices with the impedance inversion results to comprehensively predict the distribution of deepwater reservoirs. Real data application results in the South China Sea verify the feasibility and effectiveness of our method, which can provide a guidance for the future deepwater hydrocarbon exploration in this area.
Estimation of reservoir fluid saturation from seismic data: amplitude analysis and impedance inversion as a function of noise