4.2 Detector Corrections
Several corrections related to the electronics of the detectors and the detectors themselves, i.e. the terms Selec and DC in equ. 4-1 have to be applied (see fig. 4-1). The UV-VIS-NIR channels 1-5 and the SWIR channels 6-8 must be treated separately during the calibration due to their different detector material and readout electronics. Signals are described in terms of Binary Units. The Analogue-to-Digital Converter (ADC) of SCIAMACHY digitises the signal of the detector with 16 bit resolution, meaning that detector signals (also referred to as ‘fillings’) range from 0 BU to 65535 BU.
Channels 1-5 (UV-VIS-NIR)
The first correction to the data is the so-called Memory effect. The Memory effect was discovered in 1996 during an investigation of the linearity of channels 1-5. In a number of measurements covering the range from low detector fillings to saturation it was found that the signal deviated from a linear response which is defined by a linear fit for all points of up to 90% of the maximum detector fillings (see fig. 5-2). The deviation was independent of the actual signal level, but dependent on the signal level of the previous readout (hence the name Memory effect). Note that the effect depends on the signal level including the analogue offset (see below) and dark current. Thus it has to be applied before any other correction. In order to characterise the Memory effect, WLS measurements followed by several dark measurements were performed on-ground and in-flight. The difference between the first dark measurement after the WLS measurement and subsequent dark measurements gives a correction value as a function of detector filling. This value needs to be subtracted from the data to correct for the Memory effect which is the same for all pixels. The total correction for a single readout amounts from -0.61% to 0.21% of the detector filling of the previous readout with a maximum effect at fillings around 19000-21000 BU, depending on the channel. More information can be found in Lichtenberg (2003). (fig. 4-2)