To mitigate the risk and uncertainty associated with turbulent flows in tidal channels, developers often use tank experiments and numerical simulations to assess the power and loading performance of a turbine. However, it remains unclear if these controlled flows can be accurately scaled up to represent the natural turbulence present in tidal channels. The difficulty in translating between model, tank and full scale turbulent effects motivated the In-Situ Turbulence Replication, Evaluation And Measurement (InSTREAM) project. The three-year project included the development of a sensor system that combined acoustic (Doppler), and non-acoustic (electro-magnetic and shear probe) technology to create a system that could be used in both laboratory and field applications. The system was successfully deployed at the FloWave Ocean Energy Research Facility and in the Minas Passage, Bay of Fundy. The measurements from both the lab and the field were then used to perform numerical simulations of turbine performance. A direct comparison between the “tank” and “ocean” conditions was obtained by implementing a scaling method to translate the length scales between the two flow regimes. The results – from both the measurements and the simulations – highlight that there are significant differences in the turbulence characteristics between the tank and the field. In particular, the larger relative size (by a factor of three) of the 3D eddies present in the ocean generated more complex wake dynamics, and lower power and thrust coefficients.