clouds can significantly impact aerosols via scavenging by precipitation, hence contributing to the covariance between aerosol and the meteorology. These mutual interactions raise serious concerns about the numerous and contrasting satellite studies that claim correlations between aerosols and clouds, which are generally interpreted as evidence of aerosols impacting clouds. In fact, quite the opposite may actually be true.

To optimize the chances of distinguishing an aerosol signal from the background meteorological noise, we must first select a place where the aerosol variability is significant, while the variability of the meteorology is minimized. The covariance between the two, however, sets a limit to this strategy: reducing meteorological variability (e.g., by selecting specific weather situations) necessarily reduces the aerosol variability. The second aerosol characterization experiment (ACE-2), which took place in the North Atlantic in June 1997, illustrates this impediment. Within ACE-2, the cloudy column experiment was designed to test the Twomey hypothesis; namely, that the cloud optical thickness, r, scales with the LWP, W, and the droplet number concentration N as r <x W5/6 N13. The Atlantic Ocean, north of the Canary Islands, offers opportunities to sample air masses flowing around the Azores high that are generally pristine, except when they skim along the European continent (where they become polluted by anthropogenic aerosols). Droplet concentrations observed during the eight case studies ranged from less than 50 cm-3 in pristine air masses up to more than 250 cm-3 in the most polluted ones (i.e., a factor of five), whereas the LWP ranged from 33 g m-2 to 77 g m-2 (i.e., a factor of only two). Such a gap between aerosol and LWP variability allowed a precise validation of the Twomey hypothesis (discussed further below), but the dataset did not provide any evidence of an aerosol impact on the cloud life cycle. Indeed, aerosol and meteorology were closely correlated because polluted air masses had flown over the continent, and had hence experienced greater sensible and lower latent heat fluxes than the pristine oceanic air masses. Overall, the most polluted cases exhibited lower LWP than their pristine counterpart (Brenguier et al. 2003), and it was not feasible to determine whether clouds were thinner as a result of the reduced latent heat fluxes two days ahead over the continent, attributable to direct aerosol effects on the air mass when it moved over the ocean from the continent to the sampling area, or because of an indirect aerosol effect on the cloud layer.

Attempts were made to examine the climatology of precipitation downwind of large cities such as St. Louis, Missouri (Changnon et al. 1971). Careful analysis of the observations with a detailed numerical model indicated, however, that urban land use forced convergence downwind of the city, rather than the presence of greater aerosol concentrations, was the dominant control on the locations and amounts of precipitation in the vicinity of an urban complex (van den Heever and Cotton 2007).

An attractive alternative approach is to select situations where aerosol and meteorological variations are uncorrelated. This has served as the basis of weather modification control experiments for many years. Accidental biomass burning events offer such opportunities, but they are generally not frequent enough to build significant statistics. To improve the statistics, weekend effects have been examined climatologically (Forster and Solomon 2003; Gong et al. 2006). Evidence of a detectable weekly cycle of the diurnal temperature range has been shown, as well as the anthropogenic origin of this cycle. A potential aerosol effect on clouds was investigated by Bell et al. (2008), but may be masked by other atmospheric parameter changes (e.g., heat island effect) that are also correlated to the weekly cycle of anthropogenic activities. The identification of situations where aerosol variability is uncorrelated with meteorology thus remains to be resolved.

For shallow clouds, the issue of sensitivity is particularly acute. Indeed, their liquid water content is typically a few hundredths of the total water content. This explains why deriving their LWP and its temporal evolution from field observations of the thermodynamic fields (i.e. temperature and water vapor) remains beyond our grasp. The magnitude of the energy fluxes that govern the evolution of a cloud-topped boundary layer (e.g., surface fluxes including precipitation, cloud-top entrainment, and the flux divergences of short- and longwave radiation, 1M) are comparable in magnitude to their potential modulation by the aerosol, for example by suppressing precipitation (XA[). The susceptibility of marine stratocumulus clouds to aerosols is thus noticeable, as demonstrated by ship tracks.

Quantifying the susceptibilities of marine stratocumulus clouds to the meteorology and aerosol, respectively, is a challenge in the sense that small perturbations of the boundary layer state parameters need to be precisely measured. What facilitates the observation of these clouds, however, is their long synoptic lifetime (although, of course, the cloud element lifetime is only a few minutes), large spatial extension, relative statistical homogeneity at the mesoscale, and their reproducible diurnal cycle.

Beyond the susceptibilities, additional ingredients must also be carefully evaluated when designing a field experiment. This includes the spatial and temporal scales of interest, and the identifi cation of all physical processes which may interfere with the observations. Next we describe two basic categories of observational approach that are particularly useful for the study of the interactions between marine stratocumulus clouds and their environment. The principles involved, however, are applicable to other cloud systems.

0 0

Post a comment