Predicting the flow of air or water over a vehicle or other device is key to making engineering improvements. When a rough surface is exposed to the flow, it is assumed that the only effect of the roughness is to increase the friction without otherwise changing details about the flow. However, this assumption is unproven in all but the simplest situations. The purpose of this project is to definitively resolve this issue, using the novel approach of comparing flows over different surfaces that generate the same friction. This work will enable more reliable analysis and design of many systems and vehicles. The project will engage and train researchers at postdoctoral, doctorate, undergraduate, and high-school levels, enable interdisciplinary international exchanges, and provide unique research experiences for hundreds of undergraduate engineers. The approach will combine state-of-the-art experimental capabilities and experience at Virginia Tech with leading edge computational expertise at the University of Cambridge. The goal is to reveal the limits and mechanisms of rough wall similarity, by comparing the properties and mechanics of non-equilibrium boundary layers generated over rough walls with different geometry but with the same effective sand-grain roughness. This comparison will be conducted for a systematic set of two- and three-dimensional flow geometries that include many of the non-equilibrium complexities found in practical applications. At Virginia Tech, experim