Wildfires release significant amounts of trace gas and aerosols into the atmosphere. Firefighters are exposed to wildland fire smoke with adverse health effects. At larger scale, depending on meteorological conditions and fire characteristics, fire emissions can efficiently reduce air quality and visibility, even far away from emission sources. Uncertainties in fire emissions and fire plume dynamics are two important factors which substantially limit the capability of current models to predict smoke exposure and air quality degradation. The approach developped here is to develop very high (50 meters) resolution models to explicitely resove these plumes and emission in order to better numerically understand and estimate these phenomenons.