Physics-Guided Multi-Attribute Petrophysical Inversion for Quantifiying Acoustic Impedance and Porosity in the F3 Reservoir

Abstract

This study introduces a physics-guided multi-attribute inversion (PG-MAI) approach to estimate acoustic impedance and porosity from post-stack seismic data using Bayesian Ridge Regression. The method integrates ten seismic attributes—including amplitude, frequency, and geometric features—into a regularized linear regression framework with Bayesian formulation. Applied to the F3 Block in the Dutch North Sea, the workflow includes temporal upsampling, attribute extraction, and model calibration using two well logs. The inversion results demonstrate high spatial coherence and alignment with geological structures. Validation at both wells shows strong agreement between predicted and measured log values, with correlation coefficients exceeding 0.90 for both acoustic impedance and porosity. Zones of low impedance and high porosity correspond to interpreted deltaic sands and lobate geometries, reflecting the facies heterogeneity of the F3 depositional environment. These outcomes suggest that the inversion framework effectively captures lithological variability, making it valuable for reservoir delineation in similarly complex siliciclastic systems.