Uncertainty factors in modelling dispersion of smoke from wild fires in a Mediterranean area

The Mediterranean Area is affected by forest fires which can burn thousands of hectares in a few days. Biomass burning produces and releases into the atmosphere gases and particulates, spreading smoke across thousands of kilometres, affecting the air quality and the regional and global climate. Inclusion of gas and particulate emissions from wildfires in atmospheric composition models is a challenging task because of the large uncertainties related to the detection of fires and their spatial and temporal evolution, the emission factors as a function of vegetation cover, and the determination of injection height of the smoke. In this work we present a pre-processor for the estimation of forest-fire emissions and injection height. Emission fluxes have been estimated following the methodology proposed by Seiler and Crutzen (1980). We considered the gas species CO, NOx, SO2 and NH3 and the particulate matter PM2.5 and PM10 mainly composted by black carbon (BC) and organic carbon (CO). The estimate of emission rates depends on fire data, such as burned area, and consequently by associated vegetation cover, fuel loading, carbon fraction of the fuel, and combustion fraction. Most of these factors have large uncertainties, and this has an impact on the modelled concentration field. Emission heights relative to the emitted species have been estimated by means of a lagrangian backward model and the methodology proposed in the WRAP study (2005). Numerical simulations have been performed using the atmospheric composition model BOLCHEM (Mircea et al, 2008). Results of simulations for the forest fire event relative to the summer of 2007, during a period in which fires were most severe in Greece, Albania and Algeria, will be shown and discussed. The estimated pollutants emission will be compared with emission of the data set FINNv1. Afterwards, since the height of injection of smoke has an important impact on the simulated concentration, different model runs were performed with smoke injected into grid columns between altitudes of different height and with different daily modulation.