Upscaling Spectral Induced Polarization from Pore to Core Scale in Tight Reservoirs: A 3D Impedance Network Approach with an Improved Membrane Polarization Model

Tight reservoirs exhibit complex pore structures and strong heterogeneity, which complicate reservoir characterization. Although spectral induced polarization (SIP) is sensitive to pore structure, existing polarization models are formulated at the pore scale and are difficult to connect with measurements across different core volumes. We adopted an improved membrane polarization model and coupled it with a three-dimensional impedance network to investigate two-stage SIP upscaling: from pore systems to core-scale networks, and from local core volumes represented by subnetworks to larger parent cores. The framework was used to determine the representative elementary volume (REV), evaluate subnetwork representativeness, examine fracture-controlled heterogeneity, and compare simulations with laboratory measurements. The results show that REV size is controlled by the distribution ranges of pore radius and pore length. Homogeneous networks converge with sufficient sampling, whereas subnetworks from strongly heterogeneous fractured cores may exhibit relative deviations exceeding tenfold when fracture structures are not adequately captured. Impedance network inversion can reproduce measured spectra, but the recovered pore distributions are non-unique and represent electrically equivalent structures. Laboratory-scale SIP data may not fully capture large-scale heterogeneity; integrating logging or imaging information may, therefore, help constrain cross-scale non-local effects and improve the reliability of comprehensive reservoir evaluation.