Cloud phase structure: Experimental data analysis and parameterization

In this article, the data collected over 6 yr of daily observations at a network of aircraft sounding (31 stations) in the former Soviet Union, and the data collected by Canadian researchers in field campaigns in the 1990s, are reanalyzed and compared with each other. To describe the cloud phase structure (CPS), the notion of the cloud phase index (CPI)(3) is used; that is, the local mass fraction of the ice particles in the total (water + ice) water content. It is concluded that the average distribution of the (CPI)(3) values in clouds depends mainly on the temperature, the cloud types, and the scale of averagging. If these characteristics remain unchanged the geographic and seasonal variations of the phase structure are small. It is shown that for averaging scales of the order of 100 m, the frequency of occurrence of liquid clouds [(CPI)(3) = 0] varies from approximately 60% at 0 degrees C to 5% at -35 degrees C, and that of the ice clouds from about 5% to 60%. The frequency of occurrence of the mixed clouds only weakly depends on temperature, varying within 30%-40%. The dependence of the cumulative (CPI)(3) distribution oil temperature in the interval 0.1 < (CPI)(3) < 0.7 is close to linear. For stratiform clouds (without going into further details) the coefficients of the linear parameterization are found as a function of temperature. Knowing the (CPI)(3) distribution allows one to also estimate the humidity in clouds. The most urgent challenges for the experimental studies of the cloud phase structure are formulated.