Abstract : Many works show it is important to study the phyical and chemical impacts of tropical deep convection. We used the 3D mesoscale non-hydrostatic model CATT-BRAMS to study the tracers transport in the tropical troposphere above South America. I validated the tool parallel to a study done in CPTEC (Brazil) for the dry season and in the lower troposphere. The results obtained in this work indicate a globaly correct meteorological behaviour. The associated transport show an over estimation of the chemical species transport in mid-troposphere and an under estimation in the upper troposphere.This comes from a to frequent triggering of shallow convection, from the deep convection scheme and from the representation of their interactions. An adaptation of the model for the wet season is necessary. At local scale, difficulties because of a high deep convection scheme sensitivity to the orography are encountered. The CATT-BRAMS model evolve to a model with chemistry, the C-CATT-BRAMS. The first results obtained indicate a strong impact of initialisation and boudary conditions on species NO and O3. Whatever be the initialisation or the boundary conditions, we observe an increase of the mixing ratio along time for these species. This can come from an over estimation of surface emissions in the model, especialy for megacities. It is important to continue the validation of this tool in order to be able to study the physical and chemical impacts of tropical deep convection with this model.