Position details

Flood forecasting modeling using data assimilation
Training › Climate Modelling And Global Change - Data Assimilation

Required Education / Niveau requis

Master 2

From / Date de début

Fevrier 2013 approx

Duration / Durée

6 mois

Context / Contexte

With flood frequency likely to increase as a result of altered precipitation patterns tiggered by climate change there is a growing need for improved flood modeling. While significant advances have been made in recent years in hydraulic data assimilation (DA) for water level and discharge prediction, as well as for parameters correction, they are yet to be fully taken advantage of in the operational forecast of flood and inondation areas. In order to meet with the real-time constraints, the computational cost of the DA algorithm should be reduced. One classical algorithm is the Ensemble Kalman Filter (EnKF) that requires the computation and propagation of the background error covariance matrix (that represents the error statistics for the numerical model). Work is on going at CERFACS to implement the EnKF algorithm on top of the 1D Saint-Venant equations in MASCARET (developped by EDF) in the framework of a PhD. During this internship, we'd like to investigate the implementation of EnKF on top of a simplified 1D Saint-Venant model in Matlab and study the evolution of the covariance functions. A simplified model for the covariance model should be identified in order to emulate the EnKF at a low computational cost. This work has already been achieved on a flood wave propagation model and showed that with a steady observing network, the Emulated EnKF leads to similar results to those of the classical EnKF.

A second objective of the internship is to study the coupling of two hydrodynamics models when data are assimilated in only one model. Typically, over the Adour river, most of the river is modeled with a 1D model and data assimilation is used. Still, over the confluence area in Bayonne, a 2D model is used without data assimilation. The longitudinal coupling of the 2 models raises questions on how the boundary conditions should be set and how the data from the 2D modeling zone could be used.

Description / Description

This project is well-suited for students with an interest hydraulics,

data assimilation, numerical methods and

programming (Matlab, Fortran). The project is a collaboration between CERFACS, SCHAPI and EDF.

CERFACS has a long experience in climate modeling and data assimilation, SCHAPI is in charge of the production of flood risk map for most rivers in France, and EDF R&D is in charge of developing hydrodynamics modeling tools for operational use. A PhD project in collaboration with EDF and CNES may follow this initial internship in 2013.

Contacts / Contacts

Name: Thual Olivier
Phone: 05 61 28 58 08
Fax:
Email: olivier.thual@cerfacs.fr

Name: RICCI Sophie
Phone: +33(0)5 61 19 31 28
Fax: +33(0)5 61 19 30 00
Email: sophie.ricci@cerfacs.fr

Salary / Rémunération

580 euros/month