The Impact of Mineral Dust Aerosol Particles on Cloud Formation

Abstract

This investigation examines the role of selected mineral dust samples in heterogeneous ice nucleation. The experiments were done by seeding artificial clouds in the large aerosol chamber AIDA at temperatures between 273 and 190 K. Five different dust samples were collected for this purpose: Two airborne mineral dust samples (denoted Cairo dust 1:CD1 and Cairo dust 2: CD2) were collected during dust storm events at a location about 50 km North of Cairo city. The source regions of these dust samples were identified by altitude-resolved back-trajectory calculations using the FLEXTRA trajectory model and comparing with aerosol index data from the EP/TOMS, MODIS/Terra, and MODIS/Aqua satellite images. The third dust sample (denoted Egyptian Sahara mineral dust: SD) was collected from a hole of 1.5 m depth in the desert 70 km northeast of Cairo city. The fourth dust sample (denoted Asian dust sample: AD) was collected from the ground in the easterly part of the Takla Makan desert in northwest China. The fifth dust sample (denoted Arizona test dust sample: ATD) consists of crushed Arizona desert sand and was purchased from Powder technology Inc. (Minnesota, USA). It is used as a reference material for comparison. All dust samples were analysed on their elemental composition by X-ray fluorescence and on their water soluble ion contents by ion chromatography. The reported data reveals that the fraction of gypsum is the largest in CD1 and CD2, followed by SD. Furthermore, it could be shown that the airborne dust samples (CD1 & CD2) had accumulated soluble coatings during their transportation by interaction with air pollutants.<br /> In addition to the surface-chemical analysis, the mineral dust seed particles were collected during the experiments and analysed on their surface morphology by Scanning Electron Microscopy (SEM). It was found that spherical particles are slightly more efficient ice nuclei than non-spherical particles.<br /> The size distribution of the ice particles was measured by an optical particle counter (OPC, PCS-2000). The output of this instrument was successfully corrected for the evaporation of ice crystals or water droplets in the sampling line. The corrected count median diameters show a good agreement with the values determined from the optical particle counter Welas and FTIR measurements. The measured ice water content is in good agreement between the OPC-based and FTIR-retrieved. Generally, the observed ice water contents compare well with the results from upper tropospheric measurements that were obtained during the INCA campaign (Gayet et al., 2006), although both sets of measurements are totally different.<br /> In the temperature range between 273 and 240 K where mixed clouds form, all dust particles formed liquid water clouds for temperature between 273 and 255 K; very few ice crystals were formed either by condensation or immersion freezing. Between 255 and 240 K, SD and AD formed liquid water droplets, whereas ATD particles are already efficient ice nuclei by deposition freezing. The airborne Sahara dust samples CD1 and CD2 are also very efficient ice nuclei even in comparison with ATD in the temperature range from 255 to 240 K. Coating ATD particles with sulphuric acid had no significant influence on their ice nucleating ability.<br /> In the temperature range between 240 and 200 K where cirrus clouds form, ATD, SD, and AD show a reduction of the critical ice saturation ratio with decreasing temperature from 1.25 for ATD and 1.35 for SD and AD at 240 K to about 1.1 and less at the lower temperature of the cirrus cloud regime. This is in agreement with critical saturation ratios reported by Bailey and Hallett 2002 and the parameterisation line of cirrus cloud formation by Heymsfield and Miloshevich 1995. Coating ATD with sulphuric acid reduces its ice nucleation efficiency to values predicted for the freezing of haze particles after Koop et al., 2000. Coated SD, on the other hand, shows also reduced nucleation efficiencies, but the effect becomes less important at the lower temperatures.<br /> Finally, the measurements show that the nucleation rates for deposition freezing in the cirrus cloud regime increase as function of the saturation ratio, but decrease towards lower temperatures. The data reported by Archuleta et al. 2005 is in good agreement with our measurement. The results from this work can be used to improve the description of cirrus cloud formation in microphysical cloud models, but there still remain uncertainties about the ice formation in the mixed cloud regime.