Determination of Fitted Size Distribution for Atmospheric Aerosols
A synthetic set of aerosol optical depths (AODs) generated from a standard set of aerosol size distributions was analyzed by a parameter based particle swarm optimization (PBPSO) routine in order to test the reproducibility of the results. Junge and lognormal size distributions were consistently reproduced. Gamma and bimodal distributions showed large variability in solutions. values were used to determine the best subset of possible solutions allowing quantification of parameters with uncertainties when using PBPSO. AODs measured by a sun photometer on a clear day (20160413) and a foggy day (20160508) were then processed by the PBPSO program for both bimodal and lognormal distributions. Results showed that in general the foggy day has smaller values indicating that the PBPSO algorithm is better able to match AODs when there is a larger aerosol load in the atmosphere. The bimodal distribution from the clear day best describes the aerosol size distribution since the values are lower. The lognormal distribution best describes the aerosol size distribution on the foggy day (20160508). KEYWORDS: Atmospheric Aerosols; Size Distributions; Junge; Bimodal; Gamma; Lognormal; Particle Swarm Optimization; Inverse Problem; Aerosol Optical Depth