This study validates the continuum traffic flow model of capacity drop at sag and tunnel bottlenecks, as proposed by Jin (Transp. Res. B Methodol. 107, 41–56, 2018) and Wada et al. (Transp. Res. C Emerg. Technol. 113, 260–276, 2020), through empirical analysis. Specifically, after addressing the limitations in the existing studies, we calibrate the model using data from multiple congestion events at several expressway bottlenecks. We then demonstrate that the model can reproduce the observed speed recovery near the head of the queue, and assess whether both estimated bottleneck capacities and locations are consistent with observed traffic conditions. Finally, as an application of the calibration results, we examine the relationship between the spatial changes in the estimated traffic capacity and longitudinal gradients.
