An approach for determination of effective oxygen diffusivity within a polymer electrolyte membrane fuel cell cathode catalyst layer by combining tomography and geometric calculations is presented. Accurate determination of effective oxygen diffusivity is critical for modeling and characterization of cathode design architectures. In silico geometric characterization methods exploit the increasing availability, capability, and decreasing cost of tomographic imaging and high-performance computing. Geometric methods can be made simple and fast and could accelerate efforts to iteratively improve cathode design. The method described herein relied on mathematically exact and statistically significant definitions of geometric tortuosity and geometric constrictivity, in addition to porosity. To demonstrate generality and robustness, a correlation for obstruction factor underpinned by these parameters was statistically induced from hundreds of instantiations of stochastic cathode microstructures generated by a simulated annealing method and validated versus the obstruction factor of real cathode microstructures obtained separately by Star and Fuller in a previous publication and Ziegler et al. at the University of Freiburg. This work demonstrated the utility and viability of adopting a combined tomography/geomet ric-computational approach for determination of the effective diffusivity of reactant in the cathode catalyst layer and its underlying transport properties. This analysis was based on a conventional cathode microstructure with platinum catalyst, carbon black support, and perfluorosulfonic acid ionomer. However, the method could easily be extended to platinum group metal-free cathodes in which accurate characterization may be of even greater importance due to the thicker cathode layers required in these systems. This approach could be widely adopted and relied upon for characterization as the materials and processing conditions of cathode layers are iteratively tested and optimized. (C) 2019 Elsevier Ltd. All rights reserved.