Author(s): Leroux Romain; Goeury Cedric; El Kadi Abderrezzak Kamal; Tassi Pablo

Linked Author(s): Pablo TASSI, Cedric GOEURY, Kamal El Kadi Abderrezzak

Keywords: Bayesian method; Imprecise probability; Morphodynamic model; Nonlinear regression; Uncertainty propagation

Abstract: The purpose of uncertainty propagation is the quantification of input data uncertainties on the output results. This involves understanding (i) how uncertainty is represented in the model structure and the input data? (ii) how are uncertainties propagated in the model? (iii) Which uncertainties affect mostly the model outputs? The propagation analysis begins with the identification and characterization of the uncertainties of the input data. The aim of this work is to estimate the uncertainties pertaining the parameters of a 2D morphodynamic model so as to characterize the probability distribution P h (x, y, t) ≤of the water depth h (x, y, t) over the Gironde Estuary, wherea critical threshold of the water depth h (x, y, t) that allows navigation. To handle this purpose, we propose an original approache that includes sediment parameters and bathymetry data, through the use of probabilistic methods, imprecise probability and non-linear regression. The proposed strategy offers flexibility to handle the variability of these data are also suitable for data-driven applications since the uncertainty quantification can also be conducted from a small set of parameters of the 2D morphodynamic model.

DOI:

Year: 2018