Investigation of panchromatic satellite imagery sensor low bias in shadow method aerosol optical depth retrieval

Abstract

A technique known as the Shadow Method was developed for calculating aerosol optical depths by measuring the radiance difference between shaded and unshaded regions in high-resolution satellite imagery by Vincent (2006). Previous research investigated use of the Shadow Method in regions of dust obscuration using buildings and clouds as shadow generation sources and a variety of background surfaces. A recurring low bias was seen when using the Shadow Method with the QuickBird satellite's panchromatic sensor. QuickBird and WorldView1 commercial imagery was examined using the Shadow Method at several sites co-located with AERONET observations sites. The results show that low bias is not attributed to sensor calibration, processing methods of the imagery, or water vapor content. The most likely source of low bias is the region of interest (ROI) selection geometry within the shadow regions.