Drizzle in stratiform boundary layer clouds. Part II: Microphysical aspects

This is the second of two observational papers examining drizzle in stratiform boundary layer clouds. Part I details the vertical and horizontal structure of cloud and drizzle parameters, including some bulk microphysical variables. In this paper, the focus is on the in situ size-resolved microphysical measurements, particularly of drizzle drops (r > 20 mu m). Layer-averaged size distributions of drizzle drops within cloud are shown to be well represented using either a truncated exponential or a truncated lognormal size distribution. The size-resolved microphysical measurements are used to estimate autoconversion and accretion rates by integration of the stochastic collection equation (SCE). These rates are compared with a number of commonly used bulk parameterizations of warm rain formation. While parameterized accretion rates agree well with those derived from the SCE initialized with observed spectra, the autoconversion rates seriously disagree in some cases. These disagreements need to be addressed in order to bolster confidence in large-scale numerical model predictions of the aerosol second indirect effect. Cloud droplet coalescence removal rates and mass and number fall rate relationships used in the bulk microphysical schemes are also compared, revealing some potentially important discrepancies. The relative roles of autoconversion and accretion are estimated by examination of composite profiles from the 12 flights. Autoconversion, although necessary for the production of drizzle drops, is much less important than accretion throughout the lower 80% of the cloud layer in terms of the production of drizzle liquid water. The SCE calculations indicate that the autoconversion rate depends strongly upon the cloud droplet concentration N-d such that a doubling of N-d would lead to a reduction in autoconversion rate of between 2 and 4. Radar reflectivity-precipitation rate (Z-R) relationships suitable for radar use are derived and are shown to be significantly biased in some cases by the undersampling of large (r > 200 mu m) drops with the 2D-C probe. A correction based upon the extrapolation to larger sizes using the exponential size distribution changes the Z-R relationship, leading to the conclusion that consideration should be given to sampling issues when examining higher moments of the drop size distribution in drizzling stratiform boundary layer clouds.