Abstract: Combining with the existing experimental data of water entrainment for turbidity currents, uncertainties of empirical coefficients, E1 and E2, in empirical formula were investigated using the method of Gibbs sampling based on Bayesian theory. A total of 2×105 sample values for E1 and E2 were obtained. Through statistics, the maximum probability (E1,E2) values are similar to the original formula. By inputting these sample values of (E1,E2) into a fully coupled mathematical model, the evolution of the turbidity currents over a steep slope is simulated. The results indicates that the range of the computed current thickness, velocity, sediment’s volume fraction and bottom deformation increases with distance. If the sample values of (E1,E2) with the highest frequency were used, the model may underestimate the current thickness and sediment volume fraction, and over-estimate the current velocity and bottom deformation. The range of hydraulic parameters calculated with 95% (E1,E2) sample values inputted into turbidity current models is much larger than that with 25% (E1,E2) sample values, which indicates that the model is very sensitive to the value of (E1,E2).

Key words: turbidity currents； empirical relation for water entrainment； Bayesian method； uncertainty； layer-averaged numerical modeling