Improved aerosol retrieval algorithm using Landsat images and its application for PM10 monitoring over urban areas

Aerosol retrieval using MODerate resolution Imaging Spectroradiometer (MODIS) has been well researched over the past decade. However, the application is limited to global- and regional-scale studies, which may not be applicable for urban areas due to its low spatial resolution. To overcome the limitation, this paper proposed an improved aerosol retrieval algorithm for Landsat images (ImAero-Landsat) at spatial resolution of 30 m. This ImAero-Landsat algorithm has been improved in the following two aspects: (i) it does not require a comprehensive look up table and thus it is more efficient in AOT retrieval; and (ii) it can be operated in both bright and dark surfaces. The derived aerosol optical thickness (AOT)images were validated with AErosol RObotic NETwork (AERONET) measurements as well as MODIS MOD04 AOT products. Small root mean square errors (RMSEs) of 0.11 and 0.14 and mean absolute difference (MAD) of 0.07 and 0.11 between ImAero-Landsat AOT, with MODIS MOD04 and AERONET products were observed. By correlating with ground based PM10 concentrations, the ImAero-Landsat method outperforms (r(2) = 032) than MOD04 AOT products (r(2) = 023). In addition, the accuracy of estimating PM10 can be improved to r(2) = 0.55 when the derived AOT was integrated with meteorological parameters. The accuracy is similar to the results derived from AERONET AOT (r(2) = 0.62). This study offers a simple and accurate method to investigate aerosol optical thickness at detailed city-scale. Environmental authorities may use the derived methods for deriving aerosol distribution maps and pinpointing the sources of pollutants in urban areas. (C) 2014 Elsevier B.V. All rights reserved.