Quantitative saturation prediction driven by Fluid Mobility Variation with Angle

This is the authors' abstract. We don't add key points for this paper.

Accurate saturation prediction is fundamental to reservoir characterization and hydrocarbon exploration, yet conventional seismic attributes often fail to provide the sensitivity required for reliable estimation in complex reservoirs. To address this, we extend fluid mobility attribute from post stack to pre stack data and derive a new fluid-sensitive attribute, Fluid Mobility Variation with Angle. This attribute is constructed using Sparse Inverse Spectral Decomposition combined with Amplitude Variation with Offset analysis, and is then fed into a Relevance Vector Regression algorithm to enable quantitative saturation prediction. Tests on both synthetic and field data demonstrate that the proposed Fluid Mobility Variation with Angle attribute effectively discriminates between high and low gas saturation, and the saturation estimates obtained using Relevance Vector Regression algorithm exhibit strong consistency with well log data, yielding a correlation coefficient of 0.898. The proposed workflow enables quantitative saturation prediction in complex reservoirs, providing critical geophysical constraints for optimized well placement and horizontal trajectory design.