6.3.1 Cloud and Aerosol Parameters
The cloud parameter cloud fraction is determined by the threshold algorithm Optical Cloud Recognition Algorithm OCRA (Loyola 1998). Polarization Measurement Device (PMD) measurements identified to be free of clouds from the red, green, and blue PMD are stored as RGB composite. The distance between the white point in the RGB reflectance space, which is dedicated to the fully cloudy case, and the actually determined PMD reflectance is taken to determine the cloud fraction. Heuristically derived scaling factors allow the appropriate representation of cloud fraction between zero (cloud-free) and one (fully cloudy). In the initial processor versions cloud top height or cloud top pressure was extracted from the ISCCP (International Satellite Cloud Climatology Project) database but it is replace from version 3.0 onwards by an algorithm based on evaluation of the O2 A-band in the visible spectral region (SACURA, Rozanov and Kokhanovsky 2004). This yields improved cloud top height and cloud optical thickness parameters since in the former implementation only one value for optical thickness serves as fixed input.
The absorbing aerosol index AAI relies on the ratio of the reflectances between spectral bands around 340 nm and 380 nm. From the logarithmic difference to a model calculation (DAK) that does not contain aerosols the AAI is then computed. It indicates the presence of absorbing aerosols and depends strongly on an accurate calibration of the reflectance including degradation correction.
6.3.2 Nadir Trace Gas Retrievals
OL data processing uses the Differential Optical Absorption Spectroscopy (DOAS) approach for the retrieval of slant column densities from launch onwards. The DOAS concept together with the derivation of vertical column densities is described in chapter 5. Vertical column densities are deduced for the species O3, NO2, BrO,H2O, CO, and SO2, while only slant columns are provided for OClO. The current operational DOAS algorithm is based on the approach originally implemented for GOME data processor (GDP, Thomas and Spurr 1999). Since DOAS algorithm development undergoes permanent improvements – currently GDP 4.0 is operational (Spurr et al. 2004, Van Roozendael et al. 2006) – also SCIAMACHY’s processor SGP L12 OL follows these changes (Lerot et al. 2009). For O3 the GDP 4.0 implementation yields vertical columns which are derived in an iterative way by taking into account radiative transfer calculations for the AMF and the DOAS slant column density determination for each iteration step. For other species, the usual approach based on the distinction between AMF extraction, slant column density derivation, and the vertical column density calculation is used.