The indirect effect of aerosols on clouds is determined not only by the instantaneous mean droplet concentration change (i.e., the first indirect or Twomey effect; Twomey, 1977) but is also strongly associated with the development of precipitation and thus the cloud liquid-water path, lifetime of individual clouds and the consequent geographic extent of cloudiness (second indirect effect). These processes are well illustrated by marine stratiform clouds.
The longevity of marine stratiform clouds, a key cloud type for climate forcing in the lower troposphere, is dictated by a delicate balance between a number of source and sink terms for condensed water, including turbulent latent and sensible heat fluxes from the ocean surface, radiative cooling and heating rates, entrainment of dry air from above the cloud top inversion, and the precipitation flux out of the cloud. Unfortunately, changes to cloud extent induced by plausible changes in cloud droplet number concentration due to aerosol modulation can be slight and difficult to characterise (cf., Hignett, 1991). For example, Pincus and Baker (1994) point out that changes in short-wave absorption induced by changes in drop number act primarily to change cloud thickness - and cloud thickness is also strongly modulated by non-radiative processes. Depending on the cloud type, feedbacks involving cloud thickness can substantially reduce or enhance changes in cloud albedo due to change in droplet concentration (Boers and Mitchell 1994; Pincus and Baker, 1994). Feingold et al. (1997) have more recently examined the impact of drizzle modulation by aerosol on cloud optical depth.
Nevertheless, precipitation processes are extremely important to marine cloud
fraction, with varying precipitation efficiency leading to varying fractional
cloudiness and liquid water content (Albrecht, 1989). The postulated mechanism
is as follows. The activation of a larger number of aerosol particles limits
the size to which drops can grow for an available cooling rate. Hence, the number
of drops which grow large enough to initiate the collision-coalescence process
(the dominant precipitation process in warm clouds) is decreased and precipitation
rates are attenuated. With precipitation attenuated, a major sink for cloud
drops is removed and cloud lifetime is enhanced. Liou and Cheng (1989) first
estimated the potential global significance of this process. Further studies
with more sophisticated models have supported the significance of the modulation
of precipitation by aerosols and led to the consideration of several different
processes that contribute to the effects of aerosols on clouds.
One such process is the modulation of cloud optical depth by precipitation. Pincus and Baker (1994) showed, using a simple mixed-layer model, that the cloud optical depth in marine stratiform clouds was a strong function of the initial aerosol concentration upon which the cloud formed, the dependence being close to exponential. Boers (1995) subsequently demonstrated, through modelling calculations coupled with field observations, that a substantial part of the seasonal cycle in cloud albedo at Cape Grim could be due to the modulation of cloud optical depth by aerosols and their effects on the efficiency of precipitation.
A second important process which may be affected by onset of precipitation is that of decoupling of the cloud layer from the surface. Precipitation may sometimes produce a sub-cloud stable layer that cuts off the moisture flux to the cloud. However, decoupling is not an inevitable consequence of precipitation formation and, under some circumstances, a balance between the moisture flux from the surface and precipitation sinks determine the cloud extent. Pincus et al. (1997) observed no difference between precipitating and non-precipitating stratocumulus with respect to cloud fraction and both Austin et al. (1995) and Stevens et al. (1998) found that observed stratocumulus seemed able to maintain themselves despite a considerable precipitation rate.
While the effect of precipitation modulation on cloud amount is supported by a number of studies, several others have argued that external thermodynamic factors such as sea surface temperature (SST) are the main factors determining the formation and dissipation of marine stratocumulus (cf., Wyant et al., 1997). Such an analysis is also supported by the relationship of satellite-derived aerosol number concentration with cloud droplet number concentration and with liquid-water path. The former shows a positive correlation while the latter shows no particular relationship (Nakajima et al., 2001). Thus, the climatological significance of this aspect of the indirect effect needs a great deal more investigation.
Finally, it is important to note that the impact of anthropogenic aerosols on precipitation modulation will be dependent on both the natural and anthropogenic aerosol size distributions. For example, a number of studies have suggested that both natural (e.g., Feingold et al., 1999a) and anthropogenic (Eagen et al., 1974) giant CCN have a great impact on precipitation and will influence the effect of smaller, anthropogenic CCN on precipitation.
Other reports in this collection