Adam Whitten, Physics
A stochastic particle swarm optimization (SPSO) technique’s robustness is studied in regards to atmospheric aerosol size distribution estimations for a bimodal distribution that focuses on Aitken and accumulation mode aerosols. The SPSO method is used to calculate a set of 11 aerosol optical depth (AOD) values based on a size distribution and match them to an inputted set of AOD values. This method is tested using computer generated AOD values with fixed distribution parameters, generated AOD values with varying distribution parameters, two sets of AOD measurements in clear conditions, and one set of AOD values in hazy conditions. The SPSO method is found to be robust and consistent for generated AOD values when keeping mean radii and width distribution parameters constant, but less consistent when all parameters are allowed to vary. For real data, the SPSO program was not consistent from run to run, and a process of averaging the runs is suggested. An analysis of the Mie scattering coefficient’s large differences at small radii ranges and analysis of the consistency of the AOD values for the hazy conditions lead us to suggest that testing a bimodal distribution that focuses on accumulation and course mode aerosols should be done, as the major inconsistencies in the distributions occur in these small radii ranges.
Nault-Maurer, Benjamin D., "Retrival of Atmospheric Aerosol Size Distributions Using Stochastic Particle Swarm Optimization" (2016). All College Thesis Program, 2016-present. 5.