8. Conclusions

Laboratory experiments are important in as far as they permit approach from several directions; to place boundaries on the possible, to be confirmed by a combination of new laboratory studies from wind tunnel to molecular scale. They are further of importance in suggesting aircraft studies which measure cloud properties at important places and numerical approaches to see if the numbers from laboratory and field make sense in the overall evolution of the cloud system. As important a concept is that just because a certain set of laboratory measurements ``fit'' the observations it may only be one of a set of physical processes. One needs to err on the side of rigor and caution in approaching this problem.

Desiderata may be summarized:

• Evidence for existence and frequency of occurrence, and chemical and physical properties of highly supercooled and supersaturated haze particles; the nature of nucleation of solute crystals; chemical reactions under conditions of high supersaturation. Application lies in the radiative properties of cloud layers, their importance in radiation balance particularly through the presence of large numbers of small particles. Such particles are also important in chemical reactions, which take place both on their surface and within their bulk for liquid particles. We know that a measure of CCN together with updraft leads to droplet dilution and freezing below --40C. All of these things will be nucleation sensitive. Frequency of suitable conditions for dominance of such small particles are not well characterized; it may well be that unstable supercooled or supersaturated haze particles are as common as supercooled water cloud.

• Different ice particle shapes are known to give rise to different radiative properties both in emission and absorption, which are wavelength dependent. Laboratory techniques are capable of this measurement and characterization; application to the atmosphere lies in knowing frequency of particle size and shape, and applying this in the setting of real cirrus, with an appropriate spread of sizes, shapes and temperatures.

• The difference of ice crystal shapes comes from interaction of growth processes with ice surface kinetics. This is still not well understood and new techniques should be investigated. The surfaces of ice particles are probably important in many chemical reactions and electrical charge separation processes; studies under controlled conditions are required.

• While the association of the mixed ice-super-cooled water phase and ice crystal/particle bounce with charge separation is well established, the characterization of the surface is not, and better controlled studies are needed.

Laboratory techniques have demonstrated sensitivity to microscale processes in several areas. This includes radiative properties of clouds, both cirrus and aerosol role in global radiation budgets; precipitation to remove water substance by ice/water processes to distribute latent heating/cooling as influenced by detail of the processes; electrification as a diagnostic tool for dynamical evolution. Uncertainty in introducing these ideas in the different scales of numerical models reflects the basic uncertainties of our knowledge of the processes discussed. Lack of understanding here is critical; numerical models without such insights will be of limited predictive utility.

Insight into preparation of this summary was obtained while working under Grant ATM-9021918, National Science Foundation, Meteorology Program.