Understanding the potential effects of pile driving sounds on marine wildlife is essential for regulating offshore wind developments. Here, tracking data from 24 harbour seals were used to quantify effects and investigate sensitivity to the methods used to predict these. The Aquarius pile driving model was used to model source characteristics and acoustic propagation loss (16 Hz–20 kHz). Predicted cumulative sound exposure levels (SELcums) experienced by each seal were compared to different auditory weighting functions and damage thresholds to estimate temporary (TTS) and permanent (PTS) threshold shift occurrence. Each approach produced markedly different results; however, the most recent criteria established by Southall et al. [(2019) Aquat. Mamm. 45, 125–232] suggests that TTS occurrence was low (17% of seals). Predictions of seal density during pile driving made by Russell et al. [(2016) J. Appl. Ecol. 53, 1642–1652] were compared to distance from the wind farm and predicted single-strike sound exposure levels (SELss) by multiple approaches. Predicted seal density significantly decreased within 25 km or above SELss (averaged across depths and pile installations) of 145 dB re 1 μPa2⋅s. However, there was substantial variation in SELss with depth and installation, and thus in the predicted relationship with seal density. These results highlight uncertainty in estimated effects, which should be considered in future assessments.