Realistic oceanographic conditions are essential to consider in the design of resilient tidal-stream energy devices that can make meaningful contributions to global emissions targets. Depth-averaged or simplified velocity profiles are often used in studies of device performance, or device interaction with the environment. We improve representation of flow at tidal-stream energy regions by characterising the velocity profile. At two potential tidal-stream energy sites, the 1/7th power-law with a bed-roughness coefficient of 0.4 accurately described the observed velocity profile on average (>1 month ADCP deployments). Temporal variability in the power-law fit was found at both sites, and best characterised with Generalised Extreme Value distribution; with correlation of variability to tidal condition, wind speed and wave conditions found. The mean velocity profile was accurately simulated using a 3D hydrodynamic model (ROMS) of the Irish Sea (UK) but with temporal variability in accuracy of power-law fits. For all potential tidal sites, the spatial-mean velocity profile was also found to be similar (characterised with similar to 1/7th power-law and 0.4 bed-roughness value). Therefore realistic flow conditions can be characterised for tidal-energy research, but dynamically coupled wind-wave-tide models, or long-term observations, are needed to fully characterise velocity profile temporal variability. (C) 2017 The Author(s). Published by Elsevier Ltd.