IGACtivities logo

Issue No. 17, May 1999

 

 

Aerosol-cloud interactions and indirect forcing

  

Contributed by
John H. Seinfeld and Richard C. Flagan, California Institute of Technology, USA

The indirect climatic effect of aerosols refers to the influence of aerosols on cloud optical depth and albedo, and cloud lifetime, as a result of anthropogenic changes in the number and composition of cloud condensation nuclei (CCN). Tropospheric aerosols currently pose one of the largest uncertainties in prediction of climate forcing from anthropogenically induced changes in the composition of the atmosphere, and the most significant component of that uncertainty is associated with indirect forcing. The large uncertainty attending indirect forcing exists because of the complex sequence of phenomena that connects aerosols with CCN, CCN with cloud droplets, and cloud droplets with cloud albedo and cloud lifetime.

Because of the difficulty in isolating the CCN-cloud relationship in ambient experiments, there are few studies that demonstrate clearly the cause-and-effect relationship between increased CCN and changes in cloud droplet number concentration. Circumstantial evidence can be found in the fact that CCN concentrations are greater in continental air masses (exceeding 1000 cm-3) than in the marine atmosphere (rarely exceeding 100 cm-3) and that continental clouds tend to exhibit greater cloud droplet number concentrations than do marine clouds. Cloud droplet number concentrations have been found to increase with increasing aerosol loading [Pueschel et al., 1986]. Satellite data on mean cloud droplet radii in the Northern versus the Southern Hemisphere appear to reflect the effect of higher anthropogenic emissions of aerosols in the Northern Hemisphere [Han et al., 1994]. The existence of ship tracks, linear features of high cloud reflectivity found in marine stratus clouds, result from aerosols originating from ship stack exhaust emissions [Durkee et al., 1999a,b]. Indeed, ship tracks offer the most promising opportunity to isolate the CCN-cloud droplet number concentration linkage. We'll return to this later.

Because of the challenge of quantifying the relationship between changes in aerosol number and composition and cloud properties, current efforts at representing indirect climatic effects of aerosols in global climate models rely on empirical relations, such as that between cloud droplet number and sulfate mass concentration [Boucher and Lohmann, 1995], or on approximate analytical expressions that calculate the fraction of particles activated based on adiabatic parcel model simulations [Ghan et al., 1993; Abdul-Razzak et al., 1998]. To proceed beyond the purely empirical approach to representing aerosol-cloud interactions, it is necessary to be able to relate changes in atmospheric aerosol properties to changes in cloud properties. Key aerosol properties include particle size, number, and composition. Clouds form when aerosols are subjected to water supersaturations sufficient to cause activation. A coupling exists between the particles themselves and the water vapor field such that simulation of cloud formation depends upon aerosol concentration and CCN properties.

The classical approach to aerosol activation is Köhler theory, which assumes instantaneous equilibrium between the particle and the local water vapor saturation field; dynamic effects are not accounted for. We now know that time scales can be such that activation can be kinetically limited [Chuang et al., 1997; Ghan et al., 1999]. Classical Köhler theory also assumes a completely soluble solute in the drop, a specified fraction of which may be insoluble. It does not include two potentially important effects that serve to alter the critical particle size for activation: (1) Soluble gases, such as HNO3, that can dissolve in the growing droplet, and (2) Partially soluble solutes, such as crustal species and organics, that continue to dissolve as more water is added to the drop [Kulmala et al., 1993; Laaksonen et al., 1998]. In addition, the chemical composition of the aerosol, and whether the aerosol is externally or internally mixed, affects the CCN behavior of the air mass [see, for example, Russell and Seinfeld, 1998 and Ghan et al., 1998]. The degree of mixing of the particle population is important because multiple aerosol types compete for existing water vapor, leading to a selective activation of those particles with the more facile activation properties. This effect is important if two populations are competing as CCN, one of which is anthropogenic in origin and one natural.

Coupling all this with a proper model of cloud formation and cloud microphysics is a daunting problem for a single cloud [see, for example, Kogan et al., 1996] let alone a global model; such a detailed model is out of the question for imbedding in a global climate model. Yet, it is necessary to predict cloud droplet distributions with respect to droplet size and height in the cloud, as these will govern the radiative properties of the cloud. It would seem that simulations that are capable of representing realistic in-cloud activation are essential for obtaining sound estimates of cloud radiative properties. While a large-eddy-simulation (LES) model represents the ideal from a physical point of view, such a model is simply too computationally intensive. One-dimensional parcel models, perhaps coupled with Lagrangian parcel trajectories in cloud generated from an LES model [Feingold et al., 1998], may represent an attractive approach to capture realistic cloud microphysics, while not incurring the extreme computational burden of an LES model simulation.

In short, the major issues that involve aerosol-cloud interactions in global models are: (1) Linking aerosol mass with aerosol number. Currently, GCMs handle aerosol on a mass basis, which is sufficient for direct forcing; (2) Accurately determining the relationship between aerosol number and cloud droplet number. This relationship must consider aerosol composition and the role of soluble gases; (3) Accounting for the degree of mixing in the aerosol population. This would allow multiple aerosol types to compete for existing water vapor; (4) Incorporating cloud processing of aerosols. Chemistry taking place within cloud droplets is influential in aerosol-cloud cycles.

Now, we all know that this is a tall order, and several groups are actively pursuing these issues. When we have such a model for inclusion in a global aerosol model, how do we know that it is "correct"? The type of field program aimed at this question has been termed an aerosol-cloud closure experiment. Such a program was carried out during IGAC's Second Aerosol Characterization Experiment (ACE-2) under the CLOUDYCOLUMN experiment. The goal of an aerosol-cloud closure experiment is to measure below-cloud aerosol size and composition, in-cloud residual aerosol size and composition, and cloud drop number concentration and size distribution more or less simultaneously and then compare what is measured with what would be predicted theoretically based on the best understanding of the microphysics and chemistry involved. Multiple aircraft were involved in these types of experiments in ACE-2, and results are still being evaluated. The difficulties in making simultaneous below-cloud and in-cloud measurements of aerosol and cloud properties with a time resolution appropriate for aircraft speeds pose serious measurement challenges.

A sub-experiment within the aerosol-cloud closure experiment is the CCN closure experiment. In this experiment, aerosol size and composition are measured together with the CCN activation properties of the aerosol using an appropriate CCN instrument. Then one tests whether the CCN measurements agree with what one would predict from Köhler theory based on the measured aerosol size and composition. Again, there was an attempt at such a closure experiment during ACE-2. Measurements of CCN using either single supersaturation or supersaturation spectrum devices have long been a staple of atmospheric aerosol field campaigns. A few of these instruments have been applied on aircraft platforms. Measuring particle activation at realistic ambient cloud supersaturations in a controlled and precise way remains a continuing challenge, as is proper calibration of the devices in which these measurements are carried out. Several types of CCN instruments have been employed; a few of these instruments have been applied on aircraft platforms. A number of groups employ the classical static, thermal diffusion cloud chamber. In this device, a sample is drawn into a chamber between two horizontal, wetted plates that are maintained at different temperatures [Alofs and Carstens, 1976]. Particles in the middle between the two plates are exposed to supersaturation as the steady state temperature and water vapor partial pressure profiles are established. The concentration of particles that activate is measured, either by light scattering from the ensemble of droplets in an illuminated volume, or by observations with a video camera. The limited growth time in such instruments restricts their use to relatively high critical supersaturations, i.e., 0.2%. Measurements of supersaturation spectra require changing the temperature difference, leading to long delays and, in flight experiments, poor spatial resolution.

Continuous-flow thermal diffusion cloud chambers have been constructed to overcome some of the complications that arise due to transient operation. The size and weight of these instruments has limited their use primarily to ground-based measurements, but they do introduce some interesting capabilities. Fukuta and Saxena [1979] developed a version in which a temperature gradient across the width of the flow channel was used to produce a transverse gradient in supersaturation. Measurements at different supersaturations could then be quickly made by shifting the observation point across the flow channel.

Hudson [1989] introduced a temperature profile in the flow direction to establish a streamwise gradient in the supersaturation. Particles that activate at low supersaturations have more time to grow than do those that require higher supersaturations to activate. Hudson infers the critical supersaturation from the activated droplet size measured with an optical particle counter at the outlet of the flow channel. Supersaturation spectra have been reported over the critical supersaturation range 0.01%<SC<1%.

A new instrument developed at Caltech [Chuang et al., 1999] employs an approach first proposed by Hoppel et al. [1979] to produce supersaturation in a tube flow. The supersaturation is produced by dividing a wet wall tube into a sequence of alternating heated and cooled segments. The hot segment provides water vapor while the cooled segment produces the desired supersaturation. Because of compact size, low weight, and simplicity it was possible to produce an autonomous instrument that could be flown aboard a small airplane without operator intervention. In ACE-2, this alternating gradient CCN counter provided measurements of the number concentration of CCN with critical supersaturations of 0.2% or below.

The relationship between CCN measurements obtained with the various instruments and activation in cloud-forming regions of the atmosphere remains unclear. None of the available instruments simulates the supersaturation-time histories experienced by atmospheric particles as they undergo activation. The thermal diffusion cloud chambers that operate at fixed supersaturation, i.e., the static thermal diffusion cloud chamber and the fixed supersaturation and transverse supersaturation gradient flow instruments, are generally regarded as absolute instruments, but this can only be true for supersaturations sufficiently high that activation occurs within the limited residence time of the particles in the supersaturation region. The streamwise gradient instrument does not attempt to provide an exact simulation of cloud forming conditions. Because the measurements depend on the kinetics of droplet growth, the instrument must be calibrated. Still, the possibility of simultaneously acquiring the entire supersaturation distribution, including that at low critical supersaturations, has made this technique very attractive and led to its use in a number of campaigns

Even with a well-planned and conducted field measurement program, achieving aerosol-cloud closure is not an easy task. As mentioned above, the ship track phenomenon provides an unusual target of opportunity for studying aerosol-cloud interactions, particularly the effect of perturbations on the background environment caused by anthropogenic aerosols. The Monterey Area Ship Tracks (MAST) experiment, carried out in the summer of 1994 off Monterey, CA under support of the U.S. Office of Naval Research, provided a wealth of data on the phenomenon [special issue of J. Atmos. Sci. forthcoming], but the measurements did not address aerosol-cloud and CCN closure in the detail ultimately required. IGAC might consider a future mission or subset of a future mission in which ship tracks can be probed with the specific intention of evaluating aerosol-cloud modules that are to be used in global climate models.

Where are the gaps? Certainly there is much to do in the realm of translating detailed knowledge of cloud microphysics and chemistry into practical modules for global climate models. But perhaps an even greater challenge exists in the realm of executing the types of airborne measurements that can unambiguously establish the extent to which the atmosphere is behaving the way we think it is.

References

  1. Abdul-Razzak, H., S.J. Ghan, and C. Rivera-Carpio, A parameterization of aerosol activation. Part I: Single aerosol type. J. Geophys. Res., 103, 6123-6132, 1998.
  2. Alofs, D.J. and J.C. Carstens, Numerical simulation of a widely used cloud nucleus counter, J. Appl. Meteorol., 15, 350-354, 1976.
  3. Boucher, O. and U. Lohmann, The sulfate-CCN-cloud albedo effect: A sensitivity study with two general circulation models, Tellus, 47B, 281-300, 1995.
  4. Chuang, P.Y., R.J. Charlson, and J.H. Seinfeld, Kinetic limitation on droplet formation in clouds. Nature, 390, 594-596, 1997.
  5. Chuang, P.Y. et al., CCN measurement during ACE-2 and their relation with cloud microphysical properties, Tellus, 1999 (submitted for publication).
  6. Durkee, P.A., K.J. Noone, and R.T. Bluth, The Monterey Area Ship Track (MAST) experiment, J. Atmos. Sci., in press, 1999a.
  7. Durkee, P.A. et al., The impact of ship produced aerosols on the microphysical characteristics of warm stratocumulus clouds: A test of Hypotheses 1.1a and 1.1b,
    J. Atmos. Sci.,     in press, 1999b.
  8. Feingold, G., S.M. Kreidenweis, and Y.P. Zhang, Stratocumulus processing of gases and CCN. I. Trajectory ensemble model, J. Geophys. Res., 103, 19,527-19,542, 1998.
  9. Fukuta, N. and V.K. Saxena, A horizontal thermal gradient cloud condensation nucleus spectrometer, J. Appl. Meteorol., 18, 1352-1362, 1979.
  10. Ghan, S., H. Abdul-Razzak, A. Nenes., P.Y. Chuang, and J.H. Seinfeld, Kinetic limitations on droplet formation, J. Geophys. Res., 1999 (submitted for publication).
  11. Ghan, S. J., C.C. Chuang, and J.E. Penner, A parameterization of cloud droplet nucleation. Part I: Single aerosol species. Atmos. Res., 30, 197-222, 1993.
  12. Ghan, S.J., G. Guzman, and H. Abdul-Razzak, Competition between sea-salt and sulfate particles as cloud condensation nuclei, J. Atmos. Sci., 55, 3340-3347, 1998.
  13. Hoppel, W.A., S. Twomey, and T.A. Wojciechowski, A segmented thermal diffusion chamber for continuous measurements of CN, J. Aerosol Sci., 10, 369-373, 1979.
  14. Hudson, J.G., An instantaneous CCN spectrometer.
    J. Atmos. Ocean Technol.     6, 1055-1065, 1989.
  15. Kogan, Z.N., Y.L. Kogan, and D.K. Lilly, Evaluation of sulfate aerosols indirect effect in marine stratocumulus clouds using observation-derived cloud climatology, Geophys. Res. Let., 23, 1937-1940, 1996.
  16. Kumala, M., A. Laaksonen, P. Korhonen, T. Vesala, and T. Ahonen, The effect of atmospheric nitric acid on cloud condensation nucleus activation, J. Geophys. Res., 98, 22,949- 22,958, 1993.
  17. Laaksonen, A., P. Korhonen, M. Kumala, and R. Charlson, Modification of the Köhler equation to include soluble trace gases and slightly soluble substances., J. Atmos. Sci., 55, 853-862, 1998.
  18. Pueschel, R.F., C.C. Van Valin, R.G. Castillo, J.A. Kadlecek, and E. Ganor, Aerosols in polluted versus nonpolluted air masses: Long range transport and effects on clouds, J. Climate Appl. Meteorol., 25, 1908-1918, 1986.
 

IGAC home
IGACtivities
IGACtivities #17: "Aerosols and Climate"