Author(s): Jean-Paul Travert; Cedric Goeury; Vito Bacchi; Fabrice Zaoui; Sebastien Boyaval
Linked Author(s): Cedric GOEURY
Keywords: No Keywords
Abstract: Modelling geophysical dynamics as floods leads to constructing large systems of Partial Differential Equations (PDEs) models with uncertain parameters. In the Shallow-Water Equations (SWE), one of the uncertain input parameters is the roughness (often assumed time-independent for flood modelling, but which varies in space). Our objective is to find the best parameter structure (assigning a single coefficient to different subdomains) depending on the water depths observations. With the growing number of observations on floodplains, like satellite images e. g., one can consider data assimilation techniques to improve the numerical value of the roughness parameter. First, we resort to sensitivity analysis for reducing the dimensionality of the roughness parameter to identify, i. e. for reducing the number of subdomains with uniform roughness here.
Year: 2024