550

Sensitivity of mineral dust emissions over the Sahara to the surface wind velocity from different meteorological models

Beatrice Marticorena 1, Guillaume Siour1, David Poulet2, Gilles Bergametti1, Christel Bouet1 ,3, Fabien Brocheton2, Gilles Foret1, Benoit Laurent1, Yiwen Xu1
1LISA, CNRS, UPEC, UPD, IPSL, Creteil, France, 2Numtech, Clermont-Ferrand, France, 3iEES Paris, Bondy, France

 

Mineral dust emission are mainly driven by wind, the dust emission fluxes being a power function of this quantity. The simulation of dust emissions at the regional scale requires the use of surface wind fields simulated by meteorological models. It has been shown that both global and regional meteorological models can fail in reproducing the surface wind velocities  in specific dust source regions, and in particular the high wind velocities  responsible of the most intense dust emissions.

                In the framework of the DRUMS (DeseRt dUst Modeling: performance and Sensitivity evaluation) research project, a sensitivity study of the Saharan dust emission to the surface wind fields has been conducted using surface wind velocities  provided by  two operational meteorological centers (ECMWF and NCEP), and surface wind velocities simulated by the regional meteorological model WRF forced by different products from these centers (Re-analysis, analysis, forecast). In a first step, the modeled surface wind velocities were compared to the surface wind velocities measured in the North-African and Sahelian meteorological stations to evaluate their capability to reproduce some of the main characteristics of the measured wind velocities (seasonal patterns, wind speed higher than different thresholds ...). In a second step, the dust emissions simulated using the different meteorological fields as input for the Dust Production Model developed by Marticorena and Bergametti (1995) were discussed and compared. Aerosol satellite products were used to discuss the relevance of the simulated dust emissions in terms of spatial and temporal distributions.