Observed and simulated variability of droplet spectral dispersion in convective clouds over the Amazon

Analyses are based on aircraft measurements of growing cumuli near the Amazon basin, and on numerical simulations of an idealized ice-free cumulus. In cleaner clouds, the relative dispersion epsilon, defined as the ratio of the standard deviation to the mean value of the droplet diameter, is negatively correlated with the ratio of the cloud water content (q(c)) to the adiabatic liquid water content (q(a)), while no strong correlation between epsilon and q(c)/q(a) is seen in polluted clouds.

Bin microphysics numerical simulations suggest that these contrasting behaviors are associated with the effect of collision-coalescence in cleaner clouds, and secondary droplet activation in polluted clouds, in addition to the turbulent mixing of parcels that experienced different paths within the cloud. Collision-coalescence simultaneously broadens the DSDs and decreases q(c), explaining the inverse relationship between epsilon and q(c)/q(a) in cleaner clouds.