The results of numerical models in fluvial hydraulics are affected byuncertainties in boundary conditions, river topography or roughnesscoefficients.The aim of this work is to setup methodology for modeloutput uncertainty estimation.  Actually a first study is in handwith automatic differentiation of free surface hydraulic codes MAGE(1D) and RUBAR20 (2D) developed at Cemagref Now we want to use the dualcode to study the question of how to pass from sensitivity analysis touncertainty propagation, in terms
of confidence interval or in a deterministic form which is to bedefined. Applicant will work on the following points, by aiming inparticular the flows in urban environment:
1.What sources of uncertainty have to be taken in account:
geometry, different kinds of flow resistance, boundary condition.h
2.What output variables are suitable for evaluating numerical model
uncertainty.
3.Definition of a measurement of uncertainty.
4.Outline of an automatic procedure of evaluation of uncertainties.

A PhD in fluid mechanics / hydraulics or in applied mathematics -scientific computation is required. The contract (restricted tonon-French citizen) will have a duration of
10 months with 25 000 Euros. The successful candidate will be a willingand effective communicator and committed to team collaboration.

Applications with full CV giving names and addresses of two referees
should be sent to: faure@lyon.cemagref.fr
Cemagref Lyon: http://www.cemagref.lyon.fr/hh/
Lyon city : http://www.mairie-lyon.fr/