When selecting suitable sites for tidal stream energy arrays a wide range of factors must be considered, from the magnitude of the tidal stream resource, to realistic oceanographic conditions. Previous computational and laboratory-scale investigations into the impact of waves upon tidal turbines (such as turbine blade loadings) and turbine arrays (such as array configuration) typically assume that waves propagate "inline" to the tidal current (waves following or waves opposing the tidal current with a 20 degrees tolerance limit). We investigated the wave climate at typical tidal stream energy sites across the British Isles. The wave climate was simulated at 18 sites using a 7-year (2005-2011) SWAN wave model simulation of the northwest European shelf seas. The principal semi-diurnal lunar constituent (M2) was also estimated at these sites using the three-dimensional ROMS tidal model. A significant proportion of the wave climate (between 49% and 93% of the time), including extreme wave events (>10 m wave heights), was found to be propagating in a direction which was "oblique" to the major axis of tidal flow (i.e. waves which propagate at an angle to the tidal current with a 20 degrees tolerance limit) at all 18 selected sites. Furthermore, the average "inline" wave climate was 2.25 m less in height and 2 s less in wave period in comparison to the oblique wave climate. To understand the direct effect of waves upon the tidal stream resource, the dynamically wave-tide coupled COAWST modelling system was applied to an idealized headland case study, which represented the typical tide and wave conditions expected at first generation tidal stream energy sites. Waves were found to alter the simulated tidal velocity profile, which, because tidal stream power is proportional to velocity cubed, reduced the theoretical resource by 10% for every metre increase in wave height (R-2 94% with 22 degrees of freedom) -depending upon wave period and direction. Our research indicates that wave angle should be considered when quantifying the impact of waves upon tidal turbines, such as computational fluid dynamic (CFD) studies, or laboratory-scale experiments of wake characteristics and turbine fatigue loading. Further, dynamically coupled tide-wave models may be necessary for a thorough resource assessment, since the complex wave-tide interaction affected the tidal resource; however, in situ observations of tidal velocity profiles during a range of wave events will be essential in validating such modelling approaches in the future. (C) 2014 The Authors. Published by Elsevier Ltd.