[1] We report on measurements that were specifically designed to determine iron oxides in mineral dust aerosols needed for improved optical modeling. Atmospheric dust samples as well as samples generated in a wind tunnel from soils were analyzed by a number of analytical techniques for their total and free iron content (bulk and size resolved), hematite and goethite, mineralogy, and size distribution. These samples are representative of several important dust sources in East Asia and northern Africa. A novel data set generated from these measurements enables us to perform an in-depth modeling study of dust optical properties in the solar spectrum. We modeled the iron oxide-clay aggregates, which are the key light-absorbing species, as well as their mixtures with nonabsorbing minerals. A volume fraction of iron oxide in aggregates was determined from measurements. Significant differences in the single-scattering albedo, ! 0 , were found between hematite-and goethite-clay aggregates, although these calculations involved several important assumptions about the partition of hematite and goethite in size-resolved aggregates. Furthermore, we found that variability of the free iron content is large enough to cause important differences in ! 0 of mineral dust originating from different sources. In contrast, this variability has little effect on the extinction coefficient and optical depth. We demonstrate that for the same size distribution, ! 0 calculated from data obtained for Chinese and Tunisian samples show higher values and more distinct wavelength dependence than those of Niger dust. All the above ! 0 differ from ones calculated using the refractive indices of Patterson et al. (1977) or the OPAC model (Hess et al., 1998), which are often used in radiative transfer studies. We conclude that information on a size-resolved content of free iron and a fraction of hematite and goethite in aggregates will need to be known on a regional basis to improve the prediction of the single-scattering albedo at solar wavelengths and hence the radiative impact of atmospheric mineral dust.

This paper presents data on elemental and mineralogical composition of mineral dust from various source regions of Africa collected during the African Monsoon Multidisciplinary Analyses (AMMA) SOP0/DABEX and Dust Ouflow and Deposition to the Ocean (DODO) DODO1 experiments (January–February 2006), and the DODO2 campaign (August 2006). Bulk filter samples were collected at the AMMA supersite of Banizoumbou, Niger, as well as on board the Facility for Airborne Atmospheric Measurements (FAAM) BAe‐146 aircraft. Both mineral dust and biomass burning in external mixing occurred in surface and elevated layers during the winter field phase of the campaign. However, mineral dust was overwhelming, accounting for 72% of the estimated aerosol mass in aged elevated biomass burning layers and up to 93% in plumes of mineral dust, which generally occurred in the boundary layer. A number of well‐defined episodes of advection of mineral dust could be identified both at the ground and on the aircraft. The elemental and mineralogical composition varied depending on source region. This variability could be well traced by the calcium content, which is enhanced in dust from North Africa but depleted in dust from the Sahel. Iron oxides in the form of hematite and goethite are enriched in dust emitted within Sahel and in Mauritania, whereas dust from the Bodélé depression is iron‐oxide depleted. Iron oxides represented between 2.4% and 4.5% of the total estimated dust oxide mass. This regional variability will have to be taken into account in estimating the optical properties of absorption of mineral dust from western Africa.

Abstract. In order to develop a model providing the mass size distribution of the dust raised from the ground by the sandblasting process, mechanical characteristics of 240 [tm saltating sand grains meant to be used as projectiles in wind tunnel sandblasting experiments were carefully determined. It was found that for values of the measured friction velocity less than about 55 cm/s, the constraint of the relatively small dimensions of the wind tunnel test section did not prevent saltation from developing freely. The kinetic energy of the sand grains was also determined. Aerosols were then produced in the wind tunnel by bombarding a clay target with the saltating 240 [tm quartz grains. The size distributions of these aerosols were measured for three wind speeds with an optical particle analyzer. For the lowest wind speed the size distribution of the aerosol was similar to that of the 8.6 [tm aggregates originally constituting the agglomerates of clay, but disaggregation into smaller particles became more important when wind speed increased. A theo• of sandblasting was then developed that gave theoretical results agreeing with the experimental ones. A consequence of this theo• was that submicron particles could be released from aggregates for high wind speeds. Experiments meant to check this implication were carried out and confirmed it. Cohesion energies of kaolin particles of three different sizes, 8.6, 2.8, and approximately 0.5 gm, were calculated and found to be a decreasing exponential function of the particle size. This explains (1) why the soil-derived aerosol size distributions present a common mode in the 1-10 [tm size range, these particles being readily released even in not particularly energetic conditions, and (2) why observation of a submicron mode in natural aerosols requires a higher friction veloci .ty than for the common 1-10 [tm mode.

No abstract