Verification of Sciamachy’s Polarisation Correction over the Sahara Desert

The justification for doing the work presented in this paper lies in the recent and disturbing rumours regarding the reflectivity of the current SCIAMACHY Level 1b product. The reflectivity was claimed to show an overall 20-percent loss in magnitude compared to GOME. We will show that the reported discrepancy in the reflectivity is indeed not a fictional one, but that it is certainly not caused by the failing polarisation correction. We also concentrate on the polarisation correction itself. Unfortunately, the current version of the polarisation correction algorithm is not reliable, and shows severe errors in the applied correction. 1 POLARISATION CORRECTION OF RADIANCES Light can be described by means of Stokes vectors , in which the component refers to the intensity of the radiation, and and characterize the degree of linear polarisation [1]. Circular polarisation, described by the Stokes parameter , is negligible for atmospheric polarisation [2]. The spectral channels of SCIAMACHY are sensitive to the polarisation of the light it detects. This is an unwanted property. Because of the sensitivity of the spectral channels to polarisation, the response to detected atmospheric light is unfortunately not just a linear function of the intensity , but also depends on the Stokes parameters and describing the (linear) polarisation of the radiation. More specifically, it is given by [3]