Relationship between Pore Geometric Characteristics and SIP/NMR Parameters Observed for Mudstones

The reliable estimation of permeability remains one of the most challenging problems in hydrogeological characterization. Cost effective, non-invasive geophysical methods such as spectral induced polarization (SIP) and nuclear magnetic resonance (NMR) offer an alternative to traditional sampling methods as they are sensitive to the mineral surfaces and pore spaces that control permeability. We performed extensive physical characterization, SIP and NMR geophysical measurements on fractured rock cores extracted from a mudstone site in an effort to compare 1) the pore size characterization determined from traditional and geophysical methods and 2) the performance of permeability models based on these methods. We focus on two physical characterizations that are well-correlated with hydraulic properties: the pore volume normalized surface area (Spor) and an interconnected pore diameter (Λ). We find the SIP polarization magnitude and relaxation time are better correlated with Spor than Λ, the best correlation of these SIP measures for our sample dataset was found with Spor divided by the electrical formation factor (F). NMR parameters are, similarly, better correlated with Spor than Λ. We implement previously proposed mechanistic and empirical permeability models using SIP and NMR parameters. A sandstone-calibrated SIP model using a polarization magnitude does not perform well while a SIP model using a mean relaxation time performs better in part by more sufficiently accounting for the effects of fluid chemistry. A sandstone-calibrated NMR permeability model using an average measure of the relaxation time does not perform well, presumably due to small pore sizes which are either not connected or contain water of limited mobility. An NMR model based on the laboratory determined portions of the bound versus mobile portions of the relaxation distribution performed reasonably well. While limitations exist, there are many opportunities to use geophysical data to predict permeability in mudstone formations