Publications

Hojati, M; Boolorani, AD (2019). Using radiative transfer models in order to estimate PM10 concentration in south of Iran using MODIS images. INTERNATIONAL JOURNAL OF ENVIRONMENTAL SCIENCE AND TECHNOLOGY, 16(3), 1405-1420.

Abstract
Atmospheric particulate matters are the substances that are suspended in the atmosphere and consist of liquid and solid complex mixture. In recent decades, there have been a number of studies that focuses on the methods of estimation of PM10 concentrations using satellite images. Most of these modes use an index from those images, for example aerosol optical depth with a combination of meteorological parameters such as temperature and relative humidity. The main aim of this study is to use a satellite image index, extracted from MODIS images, in order to model PM10 concentration in Khuzestan province, Iran. In this study, a new index, aerosol reflectance, calculated from radiative transfer models is used beside some meteorological parameters. In order to model PM10 concentration, a regression-based mode and two artificial neural networks including multilayer perceptron as a feed-forward model and a radial basis function network are used. Results indicate that there is a correlation value of 0.432 between aerosol reflectance in band 3 (555nm) and PM10 values. After comparing the models in order to predict PM10 concentration, a value of R-2=0.765, with a root mean value of 79.01 is calculated for multilayer perceptron model which shows that multilayer perceptron networks results higher accuracy than others.

DOI:
10.1007/s13762-017-1591-9

ISSN:
1735-1472