Abstract
Dust particles mixed in the free troposphere have longer lifetimes than airborne particles near the surface. Their cumulative radiative impact on earth's meteorological processes and climate might be significant despite their relatively small contribution to total dust abundance. One example is the elevated dust-laden Saharan Air Layer (SAL) over the tropical and subtropical North Atlantic, which cools the sea surface. To understand the formation mechanisms of a dust layer in the free troposphere, this study combines model simulations and dust observations collected during the first stage of the Saharan Mineral Dust Experiment (SAMUM-I), which sampled dust events that extended from Morocco to Portugal, and investigated the spatial distribution and the microphysical, optical, chemical, and radiative properties of Saharan mineral dust. The Weather Research Forecast model coupled with the Chemistry/Aerosol module (WRF-Chem) is employed to reproduce the meteorological environment and spatial and size distributions of dust. The model domain covers northwest Africa and adjacent water with 5 km horizontal grid spacing and 51 vertical layers. The experiments were run from 20 May to 9 June 2006, covering the period of the most intensive dust outbreaks. Comparisons of model results with available airborne and ground-based observations show that WRF-Chem reproduces observed meteorological fields as well as aerosol distribution across the entire region and along the airplane's tracks. Several mechanisms that cause aerosol entrainment into the free troposphere are evaluated and it is found that orographic lifting, and interaction of sea breeze with the continental outflow are key mechanisms that form a surface-detached aerosol plume over the ocean. The model dust emission scheme is tuned to simultaneously fit the observed total optical depth and the ratio of aerosol optical depths generated by fine and coarse dust modes. Comparisons of simulated dust size distributions with airplane and ground-based observations are good for optically important 0.4-0.7 mu m particles, but suggest that more detailed treatment of microphysics in the model is required to capture the full-scale effect of large and very small aerosol particles beyond the above range.