Soil biogeochemical models (SBMs) are needed to predict future soil CO₂ emissions levels, but we presently lack statistically rigorous frameworks for assessing the predictive utility of SBMs. In this study, we demonstrate one possible approach to evaluating SBMs by comparing the fits of two models to soil CO₂ respiration data with recently developed Bayesian statistical goodness-of-fit metrics. Our results demonstrate that our approach is a viable one for continued development and refinement.