Global sensitivity analysis of water vapour and visibility for atmospheric correction

Abstract

This study aims to quantify sensitivity of two atmospheric condition parameters: water vapour concentration (wv) and visibility in atmospheric correction processes in an operational processing chain. wv and visibility are important atmospheric condition parameters when retrieving surface reflectance from at-sensor radiance. To save cost, these parameters are often estimated using an image-based method. Their values are therefore uncertain, which in turn propagates to the surface reflectance in the atmospheric correction process. This study proposes an e-FAST based methodology that quantifies sensitivity of the two parameters in order to calculate their relative importance. The methodology is demonstrated with HyMap data. The results show that the two parameters have high sensitivity indices (> 0.8) and are dependent on the wavelength. This indicates that both parameters are important. The uncertainty in the parameters can possibly be further reduced by improving the estimation method.