Owing to their predictability, tidal currents are an attractive source of renewable energy. However, data on the environmental impacts, especially at array scale, are lacking. We present a spatial analysis of telemetry data, identifying the effects of the presence and operations of the world's first commercial sized tidal turbine array on the movements of an acoustically sensitive marine mammal; the harbour seal (Phoca vitulina). No significant change in at sea distribution was detected between pre and post installation of the 4 turbine array. However, seals showed overt avoidance responses during turbine operations, with a significant decrease in predicted abundance (95% CIs: −11%, −49%) within ∼2 km of the array. These results provide important information for estimating collision rates between seals and tidal turbine arrays and demonstrate an analytical framework which can be used to assess the effect of arrays on other marine animal distributions. Collision risk models used to estimate collision rate between tidal turbines and marine mammals traditionally assume some non-empirical degree of avoidance, or no avoidance. The avoidance response elicited by seals to these devices could reduce the likelihood of direct collisions, with estimated avoidance rates measured here suggesting a decrease in total number of between 0.4% and 2% per annum. Further, given sustained barrier effects were not identified, avoidance may only be occurring during the potentially dangerous periods during operations thereby not restricting movement through the site at all times. However, future environmental impact assessments and spatial planning for array deployments should consider the potential for repulsion of acoustically sensitive top-predators away from ecologically important sites.