A basic precondition for obtaining successful measurements of the earth’s atmosphere from remote sensing satellites is being able to convert with high precision the recorded spectral signals into physical quantities and to maintain this process over the entire duration of the satellite mission. Which steps have to be taken into account for this purpose depends on the characteristics of the recording instrument. It can also be necessary to take into account the influence of the platform bearing the instrument. At present the Atmospheric Processors department is exclusively concerned with atmosphere missions supplying data in the UV to SWIR spectral ranges. So-called retrieval is a process by which the desired geophysical parameters are extracted from calibrated spectra (radiation densities), usually by analyzing the recorded absorption characteristics. Therefore, the calibration undertaken by the team is limited to absorption spectrometers, which operate in this spectral range.
On principle, calibration involves the transfer of electronic detector signals, which exist as binary units, into physical quantities. This is accomplished by sequential application of calibration steps as specified in a calibration equation. In general, this equation relates the incoming wavelength-dependent intensity to the measured signal for each spectrometer. The influences of transmission, quantum efficiency, scattered light, dark current and electronic effects need to be taken into account. The calibration task consists in solving this equation for each detector pixel and obtaining by inversion the intensity as a function of wavelength, in other words, the spectrum. The following steps are necessary for calibration:
- Correction for detector and electronic effects: these can be, for example, nonlinearity, dark signals, and differences in the response behavior of individual pixels
- Wavelength calibration: a wavelength is assigned to each individual pixel, which is accomplished either by analyzing special calibration lamps on the instrument or with the help of natural light sources, such as the Fraunhofer lines of the sun. The precondition is precise knowledge of the line positions of the light source
- Correction for light scattered by the optical system: a distinction is made between
- Light scattered by reflection(s) within the instrument, which diverts light of specific wavelengths to pixels which should not detect light of this wavelength (spectrally scattered light) and
- Light scattered from outside the direct field of view into the telescope (spatially scattered light)
- Correcting for polarization effects: the transmission behavior of optical instruments is also a function of the polarization unless special countermeasures are taken. In order to prevent polarization from distorting the spectrum it is usually measured, analyzed and corrected.
- Radiometric calibration, which converts the signal received at the detector into physical measurements, such as watts per square meter per wavelength per solid angle, with the help of data from measurements of calibrated light sources.
After all these steps have been taken, the wavelength-dependent reflectance can be calculated, in other words, the quotient of earth’s radiation density and solar radiation (irradiance). Reflectances are the foundation for trace gas retrieval.
The parameters required for calibration are determined during extensive ground testing as the instrument is developed. Thus all the information needed to calibrate the sensor is available at the time of launching. But these calibration parameters do not remain constant throughout the mission. Ambient conditions such as temperature or radiation load vary over the course of an orbit. In addition, optical and electronic components degrade in orbit. It is therefore necessary to carry out regular calibration and monitoring measurements during missions. Their analysis allows conclusions about changes in the calibration parameters and aging effects which have to be taken into account.
The final goal of a complete calibration procedure is to generate highly precise spectra from which geophysical parameters which are as exact as possible can be derived with the help of retrieval algorithms.