Volatile Organic Compounds (VOC) play an important role in atmospheric chemistry and are involved in the formation of secondary atmospheric pollutants as ozone and organic aerosols. Biogenic emissions of volatile organic compounds are tenfold greater than anthropogenic emissions on a global scale but their characterization remains uncertain regionally. Advances in atmospheric chemistry modeling rely on better emission inventory which needs in situ flux measurement. This work focuses on biogenic VOC emissions from Mediterranean vegetation in the frame of ChArMEx (Chemistry and Aerosol Mediterranean Experiment) that aims to concentrate scientific effort to study atmospheric chemistry and aerosols in the Mediterranean region. The Eddy Covariance (EC) method allows direct measurement of trace gas exchange between the surface and the atmosphere. EC is a reference method for flux measurement but is adapted to only a limited number of trace gas species because it requires simultaneous fast measurement (~ 0.1 s) of the species concentration and vertical wind speed. In order to broaden the range of applicability of this method, several other methods derived from EC have been proposed. Among these, the Disjunct Eddy Covariance (DEC) method relaxes the constraint on fast concentration measurement while preserving good accuracy on the flux calculation. In the frame of this PhD work, a sampling system dedicated to the DEC method for VOC flux measurement was developed. This device called MEDEE (Mesures par Échantillonnage Disjoint des Échanges d'Espèces en trace) relies on a novel technology that allows the rapid capture of an air parcel and insures its transfer at a constant pressure towards an on-line analyzer. It is composed of two mechanical syringes moved by electric actuators with an alternated functioning to supply continuously the analyzer. Three solenoid valves are installed on each reservoir to drive the sample flow depending on the cycle. A micro controller chip is used to give the rhythm of the whole system with millisecond accuracy. This sampling system has been built in chemically inert materials to avoid sample contamination or destruction; this makes MEDEE fully compatible with reactive species. MEDEE has been tested and validated on ground and aboard an aircraft, during two field campaigns. It was coupled to a fast isoprene sensor for BVOC flux measurements during two field campaigns in summers 2010 and 2011 above a downy oak forest. A Neural Network (NN) approach has been used to derive a biogenic VOC emission algorithm from these canopy level measured fluxes and concurrent environmental parameters. Isoprene emission modeling has been validated for the observed environmental conditions. In the future, such emission parameterization will be implemented in a coupled chemistry-dynamics model to study the impact of biogenic VOC emissions on air quality
