This study presents in situ aircraft measurements of Saharan mineral dust transported over the western Mediterranean basin in June–July 2013 during the ChArMEx/ADRIMED (the Chemistry-Aerosol Mediterranean Experiment/Aerosol Direct Radiative Impact on the regional climate in the MEDiterranean region) airborne campaign. Dust events differing in terms of source region (Algeria, Tunisia and Morocco), time of transport (1–5 days) and height of transport were sampled. Mineral dust were transported above the marine boundary layer, which conversely was dominated by pollution and marine aerosols. The dust vertical structure was extremely variable and characterized by either a single layer or a more complex and stratified structure with layers originating from different source regions. Mixing of mineral dust with pollution particles was observed depending on the height of transport of the dust layers. Dust layers carried a higher concentration of pollution particles below 3 km above sea level (a.s.l.) than above 3 km a.s.l., resulting in a scattering Ångström exponent up to 2.2 below 3 km a.s.l. However, the optical properties of the dust plumes remained practically unchanged with respect to values previously measured over source regions, regardless of the altitude. Moderate absorption of light by the dust plumes was observed with values of aerosol single scattering albedo at 530 nm ranging from 0.90 to 1.00. Concurrent calculations from the aerosol chemical composition revealed a negligible contribution of pollution particles to the absorption properties of the dust plumes that was due to a low contribution of refractory black carbon in regards to the fraction of dust and sulfate particles. This suggests that, even in the presence of moderate pollution, likely a persistent feature in the Mediterranean, the optical properties of the dust plumes could be assumed similar to those of native dust in radiative transfer simulations, modelling studies and satellite retrievals over the Mediterranean. Measurements also showed that the coarse mode of mineral dust was conserved even after 5 days of transport in the Mediterranean, which contrasts with the gravitational depletion of large particles observed during the transport of dust plumes over the Atlantic. Simulations with the WRF mesoscale meteorological model highlighted a strong vertical turbulence within the dust layers that could prevent deposition of large particles during their atmospheric transport. This has important implications for the dust radiative effects due to surface dimming, atmospheric heating and cloud formation. The results presented here add to the observational data set necessary for evaluating the role of mineral dust on the regional climate and rainfall patterns in the western Mediterranean basin and understanding their atmospheric transport at global scale.

The determination of the content of graphitic carbon (GC) in atmospheric aerosol samples deposited on glass fiber filters (Pallflex E70-2075W) is accomplished by means of a new Raman spectroscopic method. Raman spectra of atmospheric GC contain two main bands located at about 1600 and 1300cm−1. The GC mass loading of the filter samples is obtained by integrating the band at 1600cm−1, which scales with the amount of graphitic structures. The method is calibrated with the carbon black Monarch 71 (M71), because the location and shape of the Raman bands of M71 and atmospheric GC are very similar. Due to the specific sensitivity of the Raman method for GC a high precision of 2% is achieved (detection limit of 0.08μgcm−2). The special filter type was chosen for the Raman method, because it is employed inside the particle soot absorption photometer (PSAP), which measures the particle absorption coefficient σap by a change in light transmission during aerosol collection. This offers the opportunity to relate σap to the subsequent Raman measurement of the GC mass concentration, mGC, on one and the same aerosol sample in order to determine the GC mass absorption efficiency δGC. In a first attempt this approach is applied to atmospheric aerosol particle measurements carried out on a mountain site in central Germany yielding δGC values between 10 and 18m2g−1 with an average of (14.7±2.8)m2g−1. These results are interpreted by published model calculations where mass absorption efficiencies of light absorbing carbon inside internally mixed particles are computed. From the comparison it can be concluded that GC constitutes the dominant part of light absorbing carbon and is mainly located in the accumulation mode of aged continental aerosol particles prevailing at the observational site.

Aircraft inlets connect airborne instruments for particle microphysical and chemical measurements with the ambient atmosphere. These inlets may bias the measurements due to their potential to enhance or remove certain particle size fractions in the sample. The aircraft body itself may disturb the ambient air streamlines and, hence, the particle sampling. Also, anisokinetic sampling and transmission losses within the sampling lines may cause the sampled aerosol to differ from the ambient aerosol. In addition, inlets may change the particle composition and size through the evaporation of water and other volatile materials due to compressibility effects or heat transfer. These problems have been discussed at an international workshop that was held at the Leibniz-Institute for Tropospheric Research (IfT) in Leipzig, Germany, on 12–13 April 2002. The discussions, conclusions, and recommendations from this workshop are summarized here.

A ground-based seeding experiment using carbon dioxide and propane sprayed from pressurized bottles was carried out under supercooled cloud conditions on a small spatial and short time scale. Water vapor deposition on the artificially generated dry ice and propane ice germs as the main ice formation process (nucleation and growth) is consistent with the experimental results. After nucleation, diffusional growth of the ice particles, partly at the expense of evaporating small droplets, was identified during the mixing of the seeding line with the ambient supercooled cloud. Within the seeding plume, ice water contents up to 80% of the total condensed water are observed, although the size of the formed ice particles did not exceed 25 μm. From the changes of the ice and supercooled liquid phase with time under mixed-phase conditions, liquid water content (LWC) evaporation, ice water content (IWC) formation, and ice crystal growth rates are estimated, which are not affected by the artificial nucleation process. Thus, these rates are assessed to be applicable for a growing ice phase of small ice particles in a young mixed-phase cloud, where other growth mechanisms, like riming or aggregation, are negligible.

The paper focuses on the redistribution of aerosol particles (APs) during the artificial nucleation and subsequent growth of ice crystals in a supercooled cloud. A significant number of the supercooled cloud droplets during icing periods (seeding agents: C3H8, CO2) did not freeze as was presumed prior to the experiment but instead evaporated. The net mass flux of water vapour from the evaporating droplets to the nucleating ice crystals (Bergeron–Findeisen mechanism) led to the release of residual particles that simultaneously appeared in the interstitial phase. The strong decrease of the droplet residuals confirms the nucleation of ice particles on seeding germs without natural aerosol particles serving as ice nuclei. As the number of residual particles during the seedings did not drop to zero, other processes such as heterogeneous ice nucleation, spontaneous freezing, entrainment of supercooled droplets and diffusion to the created particle-free ice germs must have contributed to the experimental findings. During the icing periods, residual mass concentrations in the condensed phase dropped by a factor of 1.1–6.7, as compared to the unperturbed supercooled cloud. As the Bergeron–Findeisen process also occurs without artificial seeding in the atmosphere, this study demonstrated that the hydrometeors in mixed-phase clouds might be much cleaner than anticipated for the simple freezing process of supercooled droplets in tropospheric mid latitude clouds.

Due to the lack of in-situ measurements of multiphase processes in mixed clouds, the fate of the scavenged pollutant material in mixed phase clouds is poorly understood. In the few numerical models which treat the ice phase in clouds, and the associated chemical processes, it is assumed either that chemical species are completely incorporated into the ice phase upon freezing of droplets or that the incorporation process follows a Henry-like equilibrium. However, it is not known whether this is true or whether part of the material is expelled, redistributed or changed in any way during the transfer process to the ice phase. The scavenging of chemical species in mixed clouds was studied during the EU project CIME (Cloud Ice Mountain Experiment) at the Puy de Dôme, central France (1465 m a.s.l.).

The second field campaign of the Cloud Ice Mountain Experiment (CIME) project took place in February 1998 on the mountain Puy de Dome in the centre of France. The content of residual aerosol particles, of H2O2 and NH3 in cloud droplets was evaluated by evaporating the drops larger than 5 mum in a Counterflow Virtual Impactor (CVI) and by measuring the residual particle concentration and the released gas content. The same trace species were studied behind a round jet impactor for the complementary interstitial aerosol particles smaller than 5 mum diameter. In a second step of experiments, the ambient supercooled cloud was converted to a mixed phase cloud by seeding the cloud with ice particles by the gas release from pressurised gas bottles. A comparison between the physical and chemical characteristics of liquid drops and ice particles allows a study of the fate of the trace constituents during the presence of ice crystals in the cloud.

We investigated the partitioning of trace substances during the phase transition from supercooled to mixed-phase cloud induced by artificial seeding. Simultaneous determination of the concentrations of H2O2, NH3 and black carbon (BC) in both condensed and interstitial phases with high time resolution showed that the three species undergo different behaviour in the presence of a mixture of ice crystals and supercooled droplets. Both H2O2 and NH3 are efficiently scavenged by growing ice crystals, whereas BC stayed predominantly in the interstitial phase. In addition, the scavenging of H2O2 is driven by co-condensation with water vapour onto ice crystals while NH3 uptake into the ice phase is more efficient than co-condensation alone. The high solubility of NH4+ in the ice could explain this result. Finally, it appears that the H2O2–SO2 reaction is very slow in the ice phase with respect to the liquid phase. Our results are directly applicable for clouds undergoing limited riming.

A ground-based Counterflow Virtual Impactor (CVI) was optimized to achieve nearly complete in situ segregation of cloud droplets and ice crystals (with subsequent evaporation, releasing dissolved gaseous and non-volatile material) from their surrounding carrier gas and interstitial aerosol particles. With a one-dimensional numerical model, the CVI cut size D50 was reduced to 4μm from 7μm in an earlier design (Anderson et al., 1993). This could be achieved by a velocity increase to 225ms−1 inside the wind tunnel forming part of the ground-based CVI, and by minimizing all dimensions contributing to the stagnation length Lstag (distance from the wind intersection plane tunnel/CVI to the stagnation plane inside the CVI that cloud elements have to reach to be sampled). CVI and high-speed wind tunnel were designed and constructed according to the modeling results. Subsequent calibrations verified the calculated lower cut sizes D50 and quantified the slope of the collection efficiency curve in terms of cut sharpness Scut. With the new CVI lower cut sizes between 4 and 6μm can be achieved. A cloud chamber experiment was performed with CVI measurements supplemented by a Forward Scattering Spectrometer Probe (FSSP). It could be demonstrated that significant drop break-up is caused by wind tunnel velocities well beyond 150ms−1. For a reduced wind tunnel velocity of 150ms−1 a reasonable cut size of at least 5μm could be maintained, while avoiding break-up. The demonstration of break-up should have consequences for any cloud sampling technique featuring high relative velocities of cloudy air past the inlet. In particular, in-cloud retrieval of cloud nuclei concentrations on high-speed airborne platforms could be affected to a significant extent.

During February 1997, one of the 2 observational periods of CIME (cloud ice mountain experiment), a joint field experiment funded by the European Commission, took place on the summit of the Puy de Dôme in the centre of France. During this experiment the droplet spectra were measured with an FSSP and the aerosol particles in the drops and in the interstitial particle phase were measured with a counterflow virtual impactor and a round jet impactor inside a windtunnel. Very low aerosol particle and drop concentrations were observed and particles as small as 25 nm in diameter were found to activate. Two datasets obtained on 15 February and 17 February were used to study the activation of the small Aitken-mode particles and the spectral form of the droplet spectrum and the scavenging fraction. Numerous sensitivity studies were performed investigating the rôle of the number density and chemical composition of the aerosol particles. The rôle of mixing inside the orographic cloud was studied by using a new technique. It considers the fact that the air arriving on the summit of the Puy de Dôme is a mixture of air of different origins. Thus, it weighs the results of a spectral scavenging model (DESCAM or EXMIX) calculated along a number of individual trajectories. The weighing function is derived from tracer and trajectory studies with a 3-dimensional mesoscale model. The model was able to reproduce the activation of aerosol particles as small as 25 nm. It was caused by the low aerosol particle number concentrations. In general, we can conclude that the variability found in the sensitivity tests of the dynamical and chemical factors allows to reproduce the shape of the observed results. As too many free parameters exit at the moment we cannot quantify the contribution of each factor studied to the observed scavenging fraction, however, it seems that dynamics dominates.