Mecaflu team

Perception of modeling uncertainties

From Mecaflu team
Jump to navigation Jump to search


This theme focuses on an interdisciplinary approach to evaluate the uncertainties of flood modeling, the resulting risks and the perception of different social groups (scientists, decision-makers, population). Such information will allow to improve methods, tools and representations for maps and risk maps and indicators, for instance.


Floods often led to loss of human lives and important material damage. Numerical simulation is a classical tool for mapping flood hazard. Many models (commercial or research ones) exist to identify areas prone to flooding, to assist decision-makers in the design of operational management plans and in quick emergency decisions.

However, software based on numerical modeling and geographic information systems are not perfect, especially in the case of a local and sudden change in flow (log jam on a watercourse, mudslide , landslide...) or in the case of an extreme weather event.

These modeling imperfections can be studied and taken into account by calculating uncertainties in order to test the robustness of the flood hazard and risk simulation methods, and to deduce important information about the risk associated with the modelling errors.


Conceptual models

This theme notably focuses on the numerical modeling of watersheds (hydrology, hydraulics) and the modeling of socio-ecosystems (social sciences) of these watersheds in the framework of the project Idex-Emergence 2017-2019 PRIM'Eau, as an acronym for "Perception of the risk and uncertainties of modeling-Water".

The main case study is concerned with the catchment area of ​​Mutterbach (Moselle-Est) for the wealth of its hydraulic facilities and their uses over the last 100 years.

Numerical methods

In this theme, existing methods and software will be first evaluated by considering a diffusive wave model and a conceptual model used on the same watershed. Later higher-level methods will be proposed.

Best practices

These numerical methods can be used in other contexts. Here is a first example.


This study is concerned with the modeling of groudwater flows under an industrial platform in order to evaluate the implemented mitigation measures. These comprise two drains modeled as rivers, pink lines on the figure (a) and a almost impervious barrier represented in black. The piezometric heights calculated without perturbation are plotted in Figure (c). By a sensitivity analysis, the area enclosed by the impervious barrier and the variations of piezometric heights are accurately evaluated, see figures (c) and (d). The streamlines are drawn in pink in figure (c). To our knowledge, the sensitivity of the flux lines had never been studied. These results are published in [2-MCWM17].

Risk mitigation measures and sensitivity calculation