A model-based approach is used to interpret equilibrium and transient conductivity measurements for 10% gadolinium-doped ceria: Ce0.9Gd0.1O1.95 − δ (GDC10). The measurements were carried out by AC impedance spectroscopy on slender extruded GDC10 rods. Although equilibrium conductivity measurements provide sufficient information from which to derive material properties, it is found that uniquely establishing properties is difficult. Augmenting equilibrium measurements with conductivity relaxation significantly improves the evaluation of needed physical properties. This paper develops and applies the computational implementation of a Nernst–Planck–Poisson (NPP) model to represent and interpret conductivity-relaxation measurements. Defect surface chemistry is represented with both equilibrium and finite-rate kinetic models. The experiments and the models are capable of representing relaxations from strongly reducing to strongly oxidizing gas-phase environments, and vice versa. Compared to alternative models, such as ambipolar models, the NPP approach enables the direct study of large variations in conductivity that are associated with large variations in gas-phase environments.