The IASI instrument produces around 1 300 000 spectra every day. It takes around 8 seconds for IASI to acquire data from one complete across track and the onboard calibration. The former consists of 120 interferograms, each one corresponding to one pixel. Furthermore, IASI has an allocated data transmission rate of 1.5
Megabits (Mb) per second. However, the data production rate is 45 Mbit/s and therefore, a major part of the data processing is set to be on board. As such, the transmitted data is an encoded spectrum that is band merged and roughly calibrated. Additionally, there is an offline processing chain located at the
Technical Expertise Centre, also referred to as TEC. Its task is to monitor the instrument performance, to compute the level 0 and 1 initialisation parameters in relation to the preceding point and to compute the long-term varying IASI products, as well as to monitor the near real-time (NTR) processing (i.e. levels 0 and 1).
IASI processing levels There are three such processing levels for the IASI data, numbered from 0 to 2. First, Level 0 data gives the raw output of the detectors, which Level 1 transforms into spectra by applying FFT and the necessary calibrations, and finally, Level 2 executes retrieval techniques so as to describe the physical state of the atmosphere that was observed. The first two levels are dedicated to transforming the interferograms into spectra that are fully calibrated and independent of the state of the instrument at any given time. By contrast, the third is dedicated to the retrieval of meaningful parameters not only from IASI, but from other instruments from MetOp as well. For example, since the instrument is expected to be linear in energy, a non-linear correction is applied to the interferograms before the computation of the spectra. Next, the two reference views are used for the first step of radiometric calibration. A second step, performed on ground, is used to compensate for certain physical effects that have been ignored in the first (e.g., incidence correction for the scanning mirror, non-blackness effect etc.). A digital processing subsystem executes a radiometric calibration and an inverse Fourier transform in order to obtain the
raw spectra.
Level 0 The central objective of the Level 0 processing is to reduce the transmission rate by calibrating the spectra in terms of radiometry and merging the spectral bands. This is divided into three processing sub-chains: • Interferogram preprocessing that is concerned with: • the non-linearity correction • spike detection that prevents the use of corrupted interferograms during calibration • the computation of NZPD (Number sampler of the Zero Path Difference) which determines the pivot sample corresponding to the Fourier Transform • the algorithm that applies a Fourier Transform to the interferogram to give the spectrum corresponding to the measured interferogram. • The computation of the radiometric coefficients and filtering • The computation of atmospheric spectra involving applying the calibration coefficients, merging the bands and coding the spectra. • by applying a spectral scaling law, removing the offset and applying a bit mask to the merged spectra, the transmission is done at an average rate of 8.2 bits per spectral sample, without losing useful information
Level 1 Level 1 is divided into three sublevels. Its main aim is to give the best estimate of the geometry of the interferometer at the time of the measurement. Several of the parameters of the estimation model are computing by the TEC processing chain and serve as input for the Level 1 estimations. The estimation model is used as a basis to compute a more accurate model by calculating the corresponding spectral calibration and apodisation functions. This allows the removal of all spectral variability of the measurements.
Level 1a :The estimation model is used here to give the correct spectral positions of the spectra samples, since the positions are varying from one pixel to another. Moreover, certain errors ignored in Level 0 are now accounted for, such as the emissivity of the black body not being unity or the dependency of the scanning mirror on temperature. :Also, it estimates the geolocation of IASI using the results from the correlation of
AVHRR and the calibrated IIS image.
Level 1b :Here, the spectra are resampled. To perform this operation, the spectra from Level 1a are over-sampled by a factor of 5. These over-sampled spectra are finally interpolated on a new constant wave-number basis (0.25 cm−1), by using a cubic spline interpolation.
Level 1c :The estimated apodisation functions are applied. :It generates the radiance cluster analysis based on AVHRR within the IASI IFOV using the IASI
point spread function.
Level 2 This level is concerned with deriving geophysical parameters from the radiance measurements: levels around the 15th of August 2010. The high values over
Russia are due to
wildfires. By contrast, the high values over
China are main due to pollution and agricultural fires.
Copyright 2014 EUMETSAT • Temperature profiles • Humidity profiles • Columnar ozone amounts in thick layers •
Surface temperature • Surface emissivity • Fractional
cloud cover •
Cloud top temperature • Cloud top pressure • Cloud phase • Total column of N2O • Total column of CO • Total column of CH4 • Total column of CO2 • Error covariance • Processing and equality flags The processes here are performed synergically with the ATOVS instrument suite, AVHRR and forecast data from numerical weather prediction. == Methods of research ==