Sedimentary architecture characterization by combining well logs and seismic data in river-dominated delta reservoirs: the Pearl River Mouth Basin, South China Sea

In offshore fields with limited well data, intricate geological configurations, and high reservoir heterogeneity, the accurate prediction of sand body distribution and characterization of sedimentary architecture pose significant challenges due to inherent geological uncertainties and data limitations. This study employs a comprehensive approach integrating three key methods to enhance prediction accuracy: (i) fusion of spectral-decomposed seismic attributes, (ii) seismic attribute fusion of target and neighboring zones, and (iii) colored seismic inversion. The first method leverages seismic information across various frequencies, yielding reliable results for sand bodies of different thicknesses. The second method mitigates the impact of seismic responses from adjacent zones on sand body predictions, making it particularly suitable for target intervals where neighboring zones significantly influence the seismic response. The third one, colored seismic inversion enhances the prediction of vertical distribution and the stacking relationships of sand bodies. These methods have been applied in an oilfield in the Pearl River Mouth Basin, southern China. Consequently, the sedimentary architecture of a braided river delta reservoir is successfully characterized, leading to the identification of four distributary channels within a depositional Zone 1 of the Zhujiang Formation. Additionally, a comprehensive workflow integrating well logs, seismic data, and depositional models significantly improves predictions of sand body distribution and sedimentary architecture in complex geological settings, providing critical geological insights for optimizing subsequent oilfield development strategies. • Based on seismic data and well data from the study area in the Pearl River Mouth Basin, this study utilizes machine learning-based fusion techniques for seismic attributes along with seismic colored inversion techniques to effectively depict sand bodies distribution and sedimentary architecture within the Neogene Zhujiang Formation's fluvial-deltaic system. • Guided by genetically related depositional models, this study employs two different attribute fusion methods and colored inversion to predict the distribution of the fifth-order and fourth-order hierarchical architectural elements in both map view and section view. By integrating various technical methods, reservoir distribution and reservoir characteristics in complicated geological formations can be identified more accurately.