Accueil > Actualités > Séminaires > Séminaire de Ronan Paugam au LMD-X

Séminaire

Titre : Parameterization of Fire Injection Height in Large Scale Transport Model
Nom du conférencier : Ronan Paugam
Son affiliation : King's College London
Laboratoire organisateur : LMD
Date et heure : 13-09-2012 11h00
Lieu : Ecole Polytechnique, salle de réunion du PMC (2e étage)
Résumé :

The parameterization of fire injection height in global chemistry transport model is currently a subject of debate in the atmospheric community. One approach proposed in the literature is based on
relationships linking injection height and remote sensing products like the Fire Radiative Power (FRP) which can measure active fire properties.

In this work we present an approach based on the Plume Rise Model (PRM) developed by Freitas et al (2007, 2010). This plume model is already used in different host models (e.g. WRF, BRAMS). In its original version, the fire is modeled by: a convective heat flux (CHF; pre-defined by the land cover and evaluated as a fixed part of the total heat released) and a plume radius (derived from the GOES Wildfire-ABBA product) which defines the fire extension where the CHF is homogeneously distributed. Here in our approach the Freitas model is modified; in particular we added (i) an equation for mass conservation, (ii) a scheme to parameterize horizontal entrainment/detrainment, and (iii) a new initialization module which estimates the sensible heat released by the fire on the basis of measured FRP rather than fuel cover type. FRP and Active Fire (AF) area necessary for the initialization of the model are directly derived from a modified version of the Dozier algorithm applied to the MODIS "MOD14" fire product.

Preliminary result of an optimization (using the simulating annealing method) of this new version of the PRM is then proposed based on fire plume characteristics derived from the official MISR plume height project and atmospheric profiles extracted from the ECMWF analysis. So far the data set covers the main Northern America and is set up to (i) retain fires where plume height and FRP can be easily linked (i.e. avoid large fire cluster where individual plume might interact), (ii) keep fire which show decrease of FRP and AF area after MISR overpass (i.e. to minimize effect of the time period needed for the plume to reach its top position) and (iii) split per fire land cover type to optimize parameters of the initialization module and the entrainment scheme for different fire regime. Result shows that the new PRM is more likely to predict fire variability, correcting the general overestimation of the original version.

Contact :

thibaud.thonat@lmd.polytechnique.fr