The proposed research is focused on improving fundamental understanding of non-equilibrium turbulent flows close to rough surfaces, and their modeling within the framework of a structure-based, single-point turbulence closure. Using roughness-resolved numerical simulations, we will identify the role of roughness and its topography in modifying local and non-local turbulent structure as a result of both wall-blocking and mean-distortion effects. These effects will be modeled and integrated in a structure-based RANS model of turbulence which will capture the variation of turbulent statistics and anisotropies near smooth or rough walls, with arbitrary departures from flow equilibrium. To ensure that the predictive theory accounts for various roughness regimes (e.g., “waviness” and “roughness”), topographical or near-wall flow parameters that link with turbulence production and redistribution will be identified; a database of equilibrium and non-equilibrium wall-bounded flows will be established on systematically varied, mathematically generated topographies based on these parameters. The final year of the proposed effort will be focused on particular non-equilibrium flows chosen in consultation with ONR, such as unsteady gust flows or flows over curved surfaces of varying roughness. This research will contribute to a consistent theory for near-wall flows subject to various surface and flow disturbances, including roughness, longitudinal pressure gradients (in space or time) and surface curvature, and to predictive modeling strategies for Naval application.