water vapor rather quickly. Then the lowering of relative humidity forces the unnucleated cloud drops and aerosols to evaporate and become more concentrated, such that they are more difficult to freeze. So, the microphysical properties of high-altitude ice clouds depend very much on the chemical and physical properties of the aerosols.

Besides the freezing mechanism, hygroscopic aerosols may aid the formation of ice by taking on the role of INs. Abbatt et al. (2006) pointed out that dry ammonium sulfate particles may serve as INs to provide another pathway for cirrus formation. As indicated by Chen (1994), the deliquescence point becomes higher at lower temperatures, so at high altitudes aerosols more easily stay in a dry state and serve as INs. Note that the role of ammonium becomes important because if the aerosol loses its ammonium, the deliquescence point becomes lower and the aerosol will "get wet" more easily. Another possible mechanism for ice nucleation from hygroscopic aerosols involves the physical chemistry of solutions. Hazra et al. (2003) showed that hydrate crystals may form in concentrated ammonium sulfate solution, then trigger ice nucleation. This process could occur at rather high temperatures. They also pointed out that a solid mixture of ammonium sulfate and water may form below the eutectic temperature, which is around —18°C (see Fig. 5 and Chen, 1994). Such solids, just like the dry aerosols or the hydrate crystals, may initiate heterogeneous ice nucleation and formation of cirrus clouds.

Another effect that aerosols may have on ice clouds is the influence of solute coating on the evaporation of ice crystals. Chen and Crutzen (1994) hypothesized that nonvolatile solutes, which enter the ice through surface chemistry or the processes shown in Fig. 8, will remain on the surface of the ice when the water containing them evaporates. The solute interacts with the quasi-liquid layer on the surface of the ice and forms a solute coating that may reduce the surface vapor pressure and retard the evaporation of ice crystals. Such a mechanism has important implications for prolonging the lifetime of cirrus clouds. In addition, it may explain the unusually long survival time of cirrus crystal trails that fall a long distance to seed and glaciate low-level clouds, as observed by Braham (1967) and Braham and Spyers-Duran (1967).

The physical chemistry of aerosols is also important to the formation of polar stratospheric clouds (PSCs), which are responsible for the formation of the ozone hole (Turco et al., 1989). As shown in Fig. 10, it starts with the formation of stratospheric aerosols composed of sulfuric acid and water. Then, during the polar night, as the temperature in the stratosphere decreases to below 210 K, the solution droplet has the chance to freeze into a sulfuric acid tetrahydrate (SAT) crystal. Further cooling to temperatures less than about 200 K allows the simultaneous deposition of nitric acid vapor and water vapor to form nitric acid trihydrate (NAT). Continuous growth with NAT will result in Type I PSCs. If the cooling continues to reach the frost point (generally around —185 K in stratospheric conditions), then the crystals may grow by deposition of pure water vapor and form Type II PSCs.

Figure 10. Pathway for the formation of Type I and Type II polar stratospheric clouds.

2.1.4. Effects on cloud, chemistry

Soluble ingredients contained in aerosols are brought along into cloud water when they are activated into cloud drops. Some of them have already been dissociated into ions in the haze state, but some dissociate only when the drops are diluted under fast condensation growth after activation. Either way, they are involved in aqueous chemical reactions that may be important to many atmospheric phenomena. The most common soluble ingredients in aerosols are sulfate, ammonium, nitrate, sea salt (Na+ and Cl_), calcium, potassium, magnesium, organic acids, and transition metals. Because of the dominant role of sulfate and sometimes nitrate in controlling the ionic balance, most of the aerosols, as well as the cloud drops forming on them, are acidic, except over areas where the atmosphere has high loading of alkaline material such as mineral dust or sea salt.

Aerosol compositions may first affect cloud chemistry by controlling the acidity of the cloud water, because many aqueous phase reactions are strongly dependent on the pH value. For instance, the solubility of SO2 and many other species that dissociate upon dissolution in water is stronger at higher pH. Furthermore, the oxidation of dissolved SO2 by O3 or by O2 (with Fe3+ or Mn2+ as catalyst) also increases with pH under typical conditions (see Seinfeld and Pandis, 2006; p. 317). So the presence of alkaline material tends to enhance the conversion of atmospheric SO2 into sulfate, whereas the acidic materials do just the opposite.

Less noticed is that the size distribution of aerosols could also affect cloud chemistry. Cloud drops are formed from CCNs with different sizes, which means that they inherently contain different amounts of solutes after activation. With the same chemical compositions, the concentration of solutes in haze drops varies only slightly with the particle size, being higher in smaller drops due to the curvature effect (curves at time 0 in Fig. 11). But when the particles are activated into cloud drops, the solutes in them are quickly diluted because of the water uptake by condensation. Yet, the degree of dilution varies drastically with particle size, as shown in Fig. 11. Large drops dilute much slower than smaller drops, because the rate increase of water mass (i.e. dln m/dt) is roughly proportional to 1/r2, where m is the water mass and r is the

Figure 11. Time evolution of the concentration spectrum of aerosol and cloud droplets in an ascending air parcel considering only the condensation process. (a) Sulfate concentration; (b) pH value. Each curve represents the spectrum at a different time of cloud formation, and the labels indicate simulation time in seconds. At time zero, the air parcel is not saturated, so all particles are in their haze state. The air parcel becomes saturated after about 20 seconds, and a gap appears in the spectrum, which separates those activated into cloud drops and those remaining as interstitial haze particles.

Figure 11. Time evolution of the concentration spectrum of aerosol and cloud droplets in an ascending air parcel considering only the condensation process. (a) Sulfate concentration; (b) pH value. Each curve represents the spectrum at a different time of cloud formation, and the labels indicate simulation time in seconds. At time zero, the air parcel is not saturated, so all particles are in their haze state. The air parcel becomes saturated after about 20 seconds, and a gap appears in the spectrum, which separates those activated into cloud drops and those remaining as interstitial haze particles.

drop radius. From Fig. 11(a) one can see that the solute concentrations in the smallest cloud drops decrease by more than 4 orders of magnitude in less than 1 minute of cloud formation, whereas those in the largest drops (giant CCNs) hardly change at all. Similar changes are found for pH values, with those in the smallest cloud drops rising from less than 2 to near 6 in just 1 minute [Fig. 11(b)]. In a relatively narrow size range, the pH values may differ by 3 (3 orders of magnitude in hydronium concentrations) between large and small cloud drops. Such a drastic variation in solute concentration and pH values has profound influence on the cloud chemistry. Again, take SO2 chemistry as an example; most of the conversion into sulfate would occur in the smaller and diluted drops, which were formed from the smaller CCNs, unless the larger cloud drops were formed on alkaline aerosols such as sea salt or mineral dust.

The calculation shown in Fig. 11 considers only the condensation process. More complicated spectral distribution of solute concentration will result from the collision between cloud drops. In Fig. 12 is a two-dimensional particle framework for aerosols and cloud drops,

Figure 12. Evolution of the drop spectrum in the two-dimensional particle framework due to the condensation and coalescence processes. (From Chen and Lamb, 1992.)

which was developed by Chen and Lamb (1992) to describe the variation of drop composition in terms of both the water and solute contents. These aerosols are activated into cloud drops, then grow by vapor condensation and collision-coalescence, as simulated in a parcel model. Note that the first two solute bins are not activated due to a limited supersaturation acquired in the ascending air. One can see that the initially narrow spectrum evolved into a broad two-dimensional spectrum, in which droplets of the same water content may have very different solute concentrations and vice versa. Purdue and Beck (1988), Ogren and Charlson (1992), and as Pandis and Seinfeld (1991) all pointed out that the chemistry in mixed drops behaves very differently than while they are separated. Mixing of droplets with different solute concentrations will normally result in significant outgassing of dissolved volatile chemicals. This also means that droplets with solutes distributed nonuniformly among them may dissolve more trace gases than droplets of homogeneous concentrations, and thus relevant chemical reactions in the former proceed faster than in the latter. Traditional detailed (binned) cloud-microphysical and chemical models that do not consider the extra solute component must assume that droplets of the same size contain the same amount of solutes, and this leads to large errors in cloud chemistry calculation. How the final cloud drop spectrum is distributed in the two-dimensional framework as shown in Fig. 12 depends largely on the aerosol size distribution and composition. Thus one may conclude that aerosols have a strong influence on cloud chemistry through not only their chemical contents but also their effect on cloud microphysics.

2.2. Carbonaceous aerosols

Carbonaceous compounds constitute a significant fraction of aerosol mass, particularly in the area with heavy fossil fuel and biofuel burning (Nunes and Pio, 1993; Novakov et al., 2000; Andreae and Merlet, 2001). The two main classes of anthropogenic carbonaceous compounds are organic carbon (OC) and black carbon (BC). These carbonaceous ingredients play crucial roles in cloud-microphysical and radiative properties. BC has been know to influence atmospheric radiation either directly by absorbing shortwave radiation or indirectly by retarding the growth of droplets by the mechanism of radiative heating (Conant et al., 2002; Nenes et al., 2002), whereas organic chemicals may affect the capability of CCNs to activate into cloud drops in many ways. Soluble organic compounds decrease the water activity as well as the surface tension of droplets, both resulting in an increase of drop growth. (Fac-chini et al., 1999a; Abdul-Razzak and Ghan, 2004). Natural vegetation also emits an ample amount of organic gases into the atmosphere, some of which may turn into an aerosol or part of an aerosol. These organics generally behave similarly to the anthropogenic OC. Another set of carbonaceous aerosols are biogenic — those directly emitted by organisms such as pollen, fungi, or even the organisms themselves, as in bacteria and viruses. These bioaerosols may play a particular role in the formation of ice in clouds.

2.2.1. Soluble organic aerosols

Organic particulate matter can be represented as a complex mixture of OC of biogenic and/or nonbiogenic origin (Andrews et al., 1997). In the early 1980s, Likens et al. (1983) reported the existence of organic compounds in precipitation. Recently, a new class of macromole-cular polycarboxylic acids has been detected in aerosol samples (Mukai and Ambe, 1986), accounting for a significant fraction of the water-soluble organic carbon (WSOC) aerosols (Fac-chini et al., 1999b). This class of macromolecular compounds has physical and chemical properties similar to those of humic acids, the main constituent of dissolved OC in natural water (Stumm and Morgan, 1981), and for this reason they are sometimes referred to as HULIS (humic-like substances) in the literature. The possible sources of HULIS (Mukai and Ambe, 1986) are particularly the process of biomass combustion. It is noted that the molecular forms of the oxygenated, water-soluble organics (WSOC) and volatile organic carbon (VOC) are also produced in biomass combustion processes (Fal-kovich et al., 2004). Chemical characterization of atmospheric aerosols has revealed that OC is usually the second-most-abundant component of fine aerosols (0.01-1^m), after sulfates, around the globe (Novakov and Penner, 1993). Hence, organic aerosols are an important part of the global CCN budget.

Organic aerosols can be classified into directly emitted species (primary) and those formed by chemical conversion in the atmosphere (secondary). Both contribute to the atmospheric population of CCNs. The natural primary organic aerosol originates from a wide range of primary emissions of anthropogenic and biogenic activities, as well as the burning of open biomass due to natural fires and land-use practices. Based on the energy statistics for the year 1996, Bond et al. (2004) estimated that emissions of primary OC are 2.4, 5.8, 25 Tg yr_1 from combustion of fossil fuels, combustion of biofuels and open biomass burning, respectively. The main pathway of the anthropogenic emissions is vehicular exhaust, which is the primary source of dicarboxilic acid and monocarboxilic acid (Yao et al., 2003). The biosphere is another major source of primary organic aerosols. Primary bioaerosols play a very special role in cloud processes, so a separate subsection (2.2.4) is devoted to giving a more detailed introduction. Secondary organic aerosols are formed due to the oxidation of VOCs emitted from biological organisms. Bio-genic emissions are driven by temperature, light, and vegetation. Photo-oxidation products of monoterpenes (e.g. a- and ,3-pinene) (Hoffman et al., 1997), which are biogenic VOCs and are emitted mainly by terrestrial vegetation, also contribute to the aerosol budget (Kavouras et al., 1998).

Traditional cloud activation theory is commonly applied to CCNs that are composed of highly soluble inorganic salts. Yet, there are many highly or slightly soluble organic compounds that also can be considered as cloud-active nuclei (Kulmala et al., 1996). The importance of organics as CCNs was first noted during the field studies made by Desalmand et al. (1985), who found a positive correlation between the concentration of water-soluble organics and the number of CCNs. They suggested that vegetation can produce CCNs. Novakov and Penner (1993) suggested that organic aerosol particles may make up a significant portion of CCNs, comparable perhaps with the sulfate aerosol contribution to CCNs. They also indicated that 37% of the CCN number concentration measured at a marine site were sulfate particles, while the remaining 63% were attributed to organic aerosols. The CCN properties of highly soluble organic acids like oxalic, malonic and glutaric acids have also been identified (Kumar et al., 2003). Novakov and Penner (1993) found that organic aerosol mass dominated the total mass in the nucleation mode at a marine site in Puerto Rico. The presence of organics in small aerosols may be the result of nucleation from organic vapors. Therefore, identifying sources of the gas phase precursors is important to determining the origins of the organic CCNs.

Just as with the inorganic salts, there are two effects involved for organic aerosols to serve as CCNs: the solute effect, which lowers the activity of water, and the effect on surface tension. The modeling results of Anttila and Kerminen (2002) suggested that soluble organics indeed influence the activation process with both effects, but the slightly soluble compounds play only a minor role. Raymond and Pandis (2002) performed a laboratory study to show that organic species with solubility less than

0.01 g cm-3 can still be a good source of CCNs. They also found that the traditional Kohler theory (see Fig. 1) works well in predicting the activation of soluble organic aerosols, but needs modification when dealing with slightly soluble organic species. Note that the great majority of organic compounds identified in aerosols are semivolatile (Rogge et al., 1993), so when treating the activation and condensation growth of these particles one also needs to consider the mass transfer of these organics from the gas phase.

While aerosols of any composition reflect sunlight, only a few can also have absorption. These absorbing aerosols include BC or soot, desert dust (Sokolik and Toon, 1996) and some organic carbon species (Bond, 2001). They may have a warming effect, opposed to the cooling by scattering aerosols (Charlson and Pilat, 1969; Schneider, 1971). Soot, also called carbon black or black carbon, is residue from the combustion of carbon-rich organic fuels in the lack of sufficient oxygen. Fossil fuel, biofuel or biomass combustions often release soot particles in large quantities into the atmosphere; therefore, soot has become a major component of aerosols in polluted regions. Past studies of soot focused mostly on their direct effect on climate, such as the influence on large-scale circulation (Hansen et al., 1997), as well as atmospheric stability and convection activities (Ackerman et al., 2000). Much less attention has been paid to the effect of soot, also due to solar heating, on the growth of cloud drops. As demonstrated by Conant et al. (2002), when BC-containing CCNs are activated into cloud drops, the absorption of sunlight will raise droplet temperature and thus surface vapor pressure. Such heating is in effect increasing the Kohler curve saturation ratio Sd shown in Fig. 1. As derived by Chen and Hsieh (2004), Sd is modified by an exponential term under BC heating: Sd • exp

0 0

Post a comment