Optimization of a Forecast Model for PM10 Based on Remote Sensing Aerosol Optical Depth from MODIS

Abstract

The atmospheric aerosol is a highly dynamic system that consists of tiny floating particles and affects our lives in multiple ways. Over the past several years, the remote sensing of trace gases and aerosols from space has improved dramatically. In the present study, regression models were established to monitor PM10 (inhalable particulate matter) with the derivation of Aerosol Optical Depth (AOD) from remote sensing data of the Moderate Resolution Imaging Spectroradiometer (MODIS). Nanjing City, China was taken as the study region. Besides the aerosol-vertical-distribution-modified AOD and relative-humidity-modified PM10, the wind speed and atmospheric pressure were also included to conduct multivariable regression. The experimental result shows that the multivariable regression model is better than one variable regression model in fitting AOD and PM10. In addition, different seasons’ regression analysis shows that the multivariable regression model is more proper for data of the summer and autumn than that of the winter and spring.