HR: 1340h
AN: A23C-1486    [Abstracts]
TI: Inverse Modeling of Accidental Releases of Atmospheric Pollutants: New Developments.
AU: * Bocquet, M
EM: bocquet@cerea.enpc.fr
AF: CEREA Paris-Est University / ParisTech, Ecole Nationale des Ponts et Chaussées 6-8 avenue Blaise Pascal Champs sur Marne, Marne la Vallée, 77455, France
AB: An account is given on new data assimilation techniques that have recently been used to identify the source of an accidental release of pollutant into the atmosphere, and forecast (possibly in real time) the subsequent dispersion plume. It could be a chemical cloud from an industrial site or a release of radionuclides from a nuclear power plant (from minor accident to core meltdown). These methods are not necessarily based on Gaussian hypotheses, for instance the source may be given as positive or bounded. In particular, the usual least squares cost function (4D-Var) is replaced with purposely devised functionals that are not necessarily quadratic. These methods have been applied successfully to the reconstruction of the Chernobyl accident, the Algeciras incident or the ETEX experiment, and outperform previous approaches. As far as state assimilation is concerned, the techniques have been applied a posteriori to the reconstruction of the ETEX plume, after all observations are acquired. The method has also been applied to the reconstruction of the dispersion plume, in the context of an emergency situation: the data have been assimilated sequentially as they arrived and a forecast is performed after each analysis. This provides with a picture of what could be achieved (forecast, risk assessment) in case of a real emergency with a Chemistry Transport Model and advanced data assimilation techniques. A second-order sensitivity study that applies to the possibly non- quadratic cost functions has also been carried out in the context of these reconstructions.
DE: 0345 Pollution: urban and regional (0305, 0478, 4251)
DE: 3315 Data assimilation
DE: 3355 Regional modeling
SC: Atmospheric Sciences [A]
MN: 2007 Fall Meeting