Exploring the modeling and site-ranking performance of Bayesian spatiotemporal crash frequency models with mixture components.

•A comprehensive evaluation was conducted for 9 spatiotemporal crash frequency models.•The model performance was evaluated based on both in-sample and out-of-sample errors.•The site-ranking performance of the proposed models was assessed using three criteria.•A flexible approach was proposed which accommodates the variations of time-trend across space.•The research findings indicated the advantage of the proposed mixture approach to accurately predict crash counts.