REGULAR PAPER
Retrieval of particle size distribution from aerosol optical
thickness using an improved particle swarm optimization
algorithm
Jiandong Mao
1
•
Jinxuan Li
1
Received: 9 March 2015 / Accepted: 16 July 2015 / Published online: 1 August 2015
Ó The Optical Society of Japan 2015
Abstract Particle size distribution is essential for
describing direct and indirect radiation of aerosols.
Because the relationship between the aerosol size distri-
bution and optical thickness (AOT) is an ill-posed Fred-
holm integral equation of the first type, the traditional
techniques for determining such size distributions, such as
the Phillips–Twomey regularization method, are often
ambiguous. Here, we use an approach based on an
improved particle swarm optimization algorithm (IPSO) to
retrieve aerosol size distribution. Using AOT data mea-
sured by a CE318 sun photometer in Yinchuan, we com-
pared the aerosol size distributions retrieved using a simple
genetic algorithm, a basic particle swarm optimization
algorithm and the IPSO. Aerosol size distributions for
different weather conditions were analyzed, including
sunny, dusty and hazy conditions. Our results show that the
IPSO-based inversion method retrieved aerosol size dis-
tributions under all weather conditions, showing great
potential for similar size distribution inversions.
Keywords Aerosol particles Particle swarm
optimization algorithm Particle size distribution Simple
genetic algo rithm Aerosol optical thickness
1 Introduction
Aerosol is a multiphase system composed of liquid and
solid particles suspended in the atmosphere, such as dust,
soot, microorganisms, pollen, fog, ice, rain and snow par-
ticles. Aerosol plays an essential role in the environment
because it takes part in many physical and chemical pro-
cesses [1]. Aerosol has a significant influence on solar
radiation transfer through absorbing and scattering elec-
tromagnetic waves in the ultraviolet, visible and infrared
ranges. The effect of radiation plays a major role in climate
modeling, atmospheric transmission, application of remote
sensing and environmental monitoring.
Because of their wide variety of sources, the properties
of aerosol particles, such as size, shape, chemical compo-
sition and aerosol optical thi ckness (AOT), may be
heterogeneous. Their temporal and spatial variations also
are very large [2]. Knowledge of aerosol optical properties
(AOT, extinction, scattering cross section, phase function
and single-scattering albedo) and microphysical properties
(the particle size distribution and the complex refractive
index) is essential for determining the effect of aerosols on
climate and air quality. Among these properties, the par-
ticle size distribution and the AOT are two important ones
used for characterizing atmospheric conditions because
they play important roles in controlling climate [3].
Therefore, it is necessary to determine the size distribution
function of aerosol particles [4].
Aerosol particle size distribution can be derived from
aerosol optical properties, such as AOT, back scattering, or
a combination of angular scattering intensity and AOT.
Generally, the particle siz e distribution can be represented
as a mathematical functi on, namely n(r). In inversion
methods for particle size distribution, the relationship
between the extinction factor and the particle size
& Jiandong Mao
mao_jiandong@163.com
1
School of Electrical and Information Engineering, Beifang
University of Nationalities, North Wenchang Road,
Yinchuan 750021, China
123
Opt Rev (2015) 22:809–818
DOI 10.1007/s10043-015-0130-5