Positive and negative correlations between CCN concentrations and cloud droplet concentrations

 

Author

James G. Hudson — Desert Research Institute

Category

Cloud Properties

Description

Panel A displays correlation coefficients (R) between CCN concentrations at 1% S and cumulative droplet concentrations larger than the abscissa threshold sizes for averages of 11 Pacific Sulfate Experiment (PASE) and 16 Rain in Cumulus Over the Ocean (RICO) flights. PASE was done in very small clouds over the mid-Pacific. RICO was done in taller cumulus clouds over the eastern Caribbean. Panel B shows the average differential spectra from the two projects. Panel C shows a separate example of model predictions of cumulative droplet spectra grown on three CCN spectra that are exact multiples of each other (identical shapes). For sizes less than 8 µm, cumulative droplet concentrations are positively correlated with the CCN concentrations. For sizes between 9 and 13 µm, cumulative droplet concentrations are inversely (negatively) correlated with CCN concentrations. For sizes larger than 14 µm, cumulative droplet concentrations are again positively correlated with CCN concentrations.
Correlations between CCN concentrations and cumulative cloud droplet concentrations have shown consistent patterns in four aircraft field projects in diverse environments. Panel A illustrates the pattern of positive, then negative, and then less negative or positive correlation coefficients (R) with increasing size thresholds. The positive R between CCN concentrations and total cloud droplet concentrations (i.e., small size thresholds) is expected because droplet concentrations should be in proportion to the concentrations of preexisting aerosol that they condensed upon. The negative R between CCN and larger droplets concentrations arises because of the smaller droplet sizes when concentrations are higher. The maximum absolute value of negative R occurs just beyond the average mode of the droplet distributions (panel B), because this is where the effect of competition among droplets is the greatest, especially when concentrations are higher. For larger threshold sizes, R decreases in negative absolute value and can even become positive for larger droplet thresholds. This is explained by the diminished competition at larger droplet sizes where concentrations are lower. This causes the concentrations of larger droplets to revert to proportionality with the nuclei upon which they condensed. These are the CCN concentrations at lower S. If CCN concentrations at various S are in proportion, this can result in the positive or decreased negative R values that have been observed for larger threshold cumulative droplet concentrations. Intermediate R are due to the conflict between the original positive R between CCN and droplet concentrations and the negative R due to competition reducing droplet sizes. An adiabatic model that predicts droplet spectra from CCN spectra and updraft speed is consistent with these observations. This is especially the case for CCN spectra that are multiples of each other, i.e., identical concentration ratios at all S (identical spectral shapes) as in panel C. Variations in actual relative shapes of the CCN spectra cause variations in the relative droplet spectra that can result in differences in the R patterns. Real clouds are seldom adiabatic because entrainment effects on droplet sizes and concentrations are usually independent of CCN concentrations. Nevertheless, the general pattern of the R predictions was even found for considerably sub-adiabatic clouds, which indicates a persistent aerosol influence on cloud microphysics. These observations and model simulations indicate that droplet spectra are subject to greater CCN influence than merely determining total cloud droplet concentrations. The relative concentrations of large nuclei (CCN with low critical S; Sc) and giant nuclei (CCN with extremely low Sc) can influence the concentrations of large cloud droplets and drizzle drops, which then affect precipitation and thus the second indirect aerosol effect.