Offshore wind energy is an indispensable source of renewable energy for reducing greenhouse emissions. However, the underwater noise generated by impact piling during the installation of offshore wind turbine foundations poses a potential threat for the marine animals. Prediction of underwater noise is critical for the control and mitigation of the noise emission. To evaluate the significance of the uncertainties in the noise prediction and quantify its influence over range, a probabilistic quantification framework is established in this paper. First, a coupled vibroacoustic model for modelling the underwater noise generated by offshore pile driving is used to predict the noise levels. The model includes the pile-soil-water coupled system in which the seabed is modelled as a layered elastic continuum. Second, a sensitivity analysis is performed to indicate the influence of multiple factors on the sound pressure levels. Third, an investigation is performed to highlight correlation between the soil properties, blow energy and the sound levels obtained. Based on the probability distribution of single parameters, a multivariate probability distribution is derived. Considering the uncertainties through statistical approaches, the sound levels with the probabilities of occurrence are predicted using Monte-Carlo simulations. The method presented helps to quantify the uncertainties in the prediction of noise levels and the risk of the exceedance of the noise thresholds. Analysis results are compared with noise measured data collected during the construction of an offshore wind farm. ... Read more