Critical wind speeds for tornado-induced vehicle movements

Little is known about how tornadoes move (parked) vehicles, although they are one of the most commonly observed examples of tornado damage. In this study, models of vehicles were tested in a laboratory tornado simulator and in a boundary layer wind tunnel for sliding, flipping and lofting behaviors. Results were compared to full-scale observations to study whether vehicles might be a tool to estimate wind speeds and serve as an EF-scale damage indicator. This study's results could be viewed as indicative of onset wind speeds for vehicle motion since the surroundings that would play a role in the field, such as houses and trees, were not included. Such surroundings might reduce the probability of movement because they could reduce local wind speeds. Comparisons to field observations support this idea. Laboratory data predicted sliding at EF-1 velocities while the field observations of Paulikas et al. (2016) show only 24% of vehicles slide in EF-1 tornadoes. Results also predicted lofting at EF-4 velocities but field observations show only 15% of vehicles are flipped or lofted in this range. Future work to resolve these differences should examine how storm duration is related to gust factors and averaging times and the role of vertical velocity components in initiating vehicle motion.