4.
IASI Level 2 Products Overview
|
|
|
4.
IASI Level 2 Products Overview |
The Infrared Atmospheric Sounding Interferometer is composed of a Fourier transform spectrometer (IASI) and an associated Integrated Imaging Subsystem (IIS). The Fourier transform spectrometer provides infrared spectra with high resolution between 645 and 2760 cm-1 (3.6 µm to 15.5 µm). The IIS consists of a broad band radiometer with a high spatial resolution. However, the IIS information is only used for co-registration with the Advanced Very High Resolution Radiometer (AVHRR).
The main goal of the IASI mission is to provide atmospheric emission spectra to derive temperature and humidity profiles with high vertical resolution and accuracy. Additionally it is used for the determination of trace gases such as ozone, nitrous oxide, carbon dioxide and methane, as well as land- and sea surface temperature and emissivity and cloud properties.
The IASI Level 2 processing foresees not only the usage of the IASI Level 1c radiance spectra but also of the ATOVS measurements namely AMSU-A L1b, AVHRR L1B, MHS Level 1b and ATOVS L2 products. Additionally the IASI L2 processing makes usage of NWP data. The mapping of ATOVS measurements onto the IASI Instantaneous Field Of Views (IFOVs) is therefore an essential part of the processing. In the following section the IASI sampling and the synchronisation and collocation with the ATOVS instruments is discussed before the L2 processing itself is introduced in Section 4.2.
IASI is an across track scanning system with scan range of ±48° 20´, symmetrically with respect to the nadir direction. A nominal scan line covers 30 scan positions towards the Earth and two calibration views. One calibration view is into deep space, the other is observing the internal black body. The scan starts on the left side with respect to the flight direction of the spacecraft.
The elementary field of view (EFOV) is the useful field of view at each scan position. Each EFOV consists of a 2 x 2 matrix of IFOVs. Each IFOV has a diameter of 14.65 mrad, which corresponds to a ground resolution of 12 km at nadir (and a satellite altitude of 819 km). The 2 x 2 matrix is centred on the viewing direction. The instrument points spread function (PSF) is defined as the horizontal sensitivity within an IFOV. The non-uniformity within the inner 80% of the IFOV is not larger than ±5%. The IIS field of view is defined by a square area of 59.63 x 59.63 mrad, consisting of 64 x 64 pixels and covering the same area as the IASI EFOV.
The instrument scans in a step and stare modus. Each interferogram is acquired within 151 ms. The 30 Earth interferograms per scan line are taken in equally spaced time intervals every 8/37 s so that a synchronisation with AMSU is reached. Figure 4.1 summarises the synchronisation of IASI with the ATOVS instruments AMSU and MHS.
Figure 4.1: Synchronisation of IASI, AMSU-A and MHS
| Characteristics | Value | Unit |
| Scan type | step and stare | – |
| Scan rate | 8 | s |
| Stare interval | 151 | ms |
| Step interval | 8/37 | s |
| Number of Earth scans / line - EFOV | 30 | – |
| Swath | ±48.333 | deg |
| Swath width | ±1100 | km |
| IFOV - shape at nadir | circular | – |
| IFOV - size at nadir | 12 | km |
| IFOV - size at edge of scan line across track | 39 | km |
| IFOV - size at edge of scan line along track | 20 | km |
Table 4.1: IASI scanning characteristics
Within the IASI Level 2 processing the measurements from the ATOVS instruments and the ATOVS L2 products are used. The collocation between IASI and the ATOVS instruments is shown in the following figures.
Figure 4.2: Collocation of IASI (yellow) and AMSU (red). The distance along and across track is given in km.
Figure 4.3: The collocation between IASI (yellow), AMSU (red), MHS (green) and HIRS (blue) is shown for four scan lines near nadir
Figure 4.4: The collocation between IASI (yellow), AMSU (red), MHS (green) and HIRS (blue) is shown at the end of the IASI scan line
The objectives of the IASI Level 2 ground processing is the derivation of geophysical parameters from the radiance measurements. The following parameters are derived during IASI Level 2 processing:
The processing is performed in synergy with the ATOVS instrument suite, AVHRR and forecast data from numerical weather prediction. The nominal input of IASI Level 2 processor consists of:
In the following sections an overview of the IASI Level 2 processing is given.
During the pre-processing phase the geolocation is extracted from L1C data for individual IASI IFOVs. The topography and surface type (land/water) within the IASI IFOV are extracted from a Land-Sea database and a Digital Elevation Model (DEM). A quality flag is generated indicating whether the geolocation of the IASI IFOV is valid (FLG_IASIBAD). The flag FLG_LANSEA summarises the topography and the surface type (land/water).
The validity of the input data is checked by comparing the data against valid bounds and evaluating the quality flags The input data consist of the ATOVS data Level 1 and Level 2, the AVHRR cloud mask and the surface and cloud top temperature, the IASI Level 1c radiance spectra, and the NWP forecast data.
The ATOVS Level 2, AMSU Level 1, MHS Level 1 and AVHRR cloud mask and surface or cloud top temperature (S/CTT) measurements are collocated with the IASI IFOV. The results from collocation of AVHRR and IASI are stored in a correlation mask which is later used for accessing the AVHRR pixel within a IASI IFOV. The flags FLG_ATOVINT and FLG_AVHAVL summarise the results from this task.
The NWP forecast data are interpolated on to the position of the IASI IFOV. Depending on the availability and quality the data are accepted and passed to the next processing step. The quality and processing flag FLG_NWPBAD is generated.
With the AVHRR correlation mask, by taking into account the IASI sounder point spread function and the AVHRR cloud mask, the weighted fractional cloud cover and the weighted S/CTT distributions are calculated for the individual IASI IFOV. Up to three different cloud top temperatures for up to three different cloud formations are calculated. The flag FLG_IASICLR summarises the results from this task.
4.2.2.1 Choice for pure IASI or combined IASI/ATOVS/AVHRR retrieval
Based on the availability of input data and its quality, the choice is made between four different combinations of input data for the retrieval. The retrieval can be based on a combination of IASI/AVHRR, IASI/ATOVS or IASI/AVHRR/ATOVS, or a IASI stand-alone retrieval. Finally the flag FLG_RETCHC specifies the retrieval choice.
The cloud detection is performed based on the combined information of the IASI L1c spectra and the AMSU-A data; it is foreseen to use ATOVS L2 data in addition. Eight different cloud detection tests are executed and the relevant flags are set accordingly (FLG_CLDTST and FLG_IASICLD). These include the following tests:
| Cloud detection | ||
| Test name | Type of test | Measurements used |
| Test A | Window channel test | IASI; NWP forecast, ATOVS L2, or climatology |
| Test B | IASI inter-channel regression test | IASI stand-alone |
| Test C | IASI-AMSU-A inter-channel regression test | IASI and AMSU-A |
| Test D | Horizontal coherency test | IASI stand-alone |
| Test E | Window cross-correlation test | IASI stand-alone |
| Test F | Window cross-correlation test | IASI stand-alone |
| Test G | Cloud test over elevated polar regions | IASI stand-alone |
| Test H | Detection of dust storms | IASI stand-alone |
Table 4.2: IASI Level 2 cloud detection tests
If ATOVS data are not available only the IASI stand-alone cloud detection tests will be executed.
4.2.2.3 Determination of cloud top height and fractional cloud cover
If clouds are detected in the field of view the CO2-slicing method is used to estimate the fractional cloud cover and the cloud top pressure for the IASI IFOVs.
If the IFOV has been declared cloudy by the AVHRR-derived fractional cloud cover the average cloud top height and cloud top temperatures are calculated for the number of identified cloud formations. The assignment of cloud top temperature (CTT) versus pressure is made by using a temperature profile from ATOVS, a NWP forecast or climatology depending on availability of data and configuration.
Finally the fractional cloud cover and the cloud top height derived from the CO2-slicing method and the weighted AVHRR measurements are compared. In the event of strong disagreements the higher fractional cloud cover and cloud top height are used. If only small differences occur the mean fractional cloud cover and cloud top height are calculated. The flag FLG_FRCSEL indicates the agreement or disagreement of the fractional cloud cover and the cloud top pressure derived from the different data sets and techniques. The fractional cloud cover, the cloud top height and the cloud top temperature, and the number of cloud formations are part of the Measurement Data Record (MDR) (FRACTIONAL_CLOUD_COVER, CLOUD_TOP_TEMPERATURE, CLOUD_TOP_PRESSURE and NUMBER_CLOUD_FORMATIONS).
Additionally the cloud phase is estimated for cloudy IASI IFOVs by evaluation of the infrared window regions between 8 - 9 and 11 - 12 µm. The cloud phase is part of the MDR (CLOUD_PHASE).
A summary based on information from the previous cloud detection tests is generated. The fractional cloud cover is evaluated against thresholds and used to determine whether a IFOV is clear, partly cloudy, or cloudy. Results are given in the cloud flag FLG_CLDSUM. Based on this information the retrieval selection flag FLG_RETCHC is updated as well.
The thin cirrus detection algorithm evaluates the IASI spectra near 11 and 12 µm. The thin cirrus flag FLG_THICIR summarises the result from this test.
4.2.2.4 Selection of clear, cloudy or cloud-clearing retrieval
The retrieval type is selected with respect to the results of the cloud processing. The options are the clear sky retrieval on an IFOV basis, cloudy retrieval in single IFOVs, and the cloud clearing option for partly cloudy field of views. The cloud clearing is specified to be performed using four IFOVs. That option is the subject of research work and not implemented in the operational processor yet. Also note that the iterative retrieval can produce a full profile as long as cloud fraction remains below a given threshold. Finally, a clear retrieval can be selected for completely cloudy IASI IFOVs where the profiles are derived only above the cloud top. The flag FLG_FINCHC is set accordingly.
For increasing cloud cover the choices are:
| clear retrieval | cloudy retrieval |
variational cloud clearing |
clear retrieval above cloud top |
| clear IFOV | ---> increasing
cloud cover ---> |
completely cloudy IFOV | |
4.2.3.1 Determination of climatological profiles and surface parameters
According to the geographic location and time of the IASI measurement, the climatological profiles and surface parameters are extracted from a database together with their covariance.
4.2.3.2 IASI channel or super channel selection
The iterative retrieval can be configured to use a sub-set of IASI channels or super channel clusters. The purpose of the super channel clustering is to reduce the number of channels by combining channels with redundant information. They consist of linear combinations of IASI channels with highly correlated radiances. The choice depends on the configuration and data availability. Super channels are only generated if the iterative retrieval is selected.
The first retrieval is generated using IASI alone or IASI in combination with ATOVS. The type of the first retrieval is specified by the flag FLG_FINCHC; either the EOF (Empirical Orthogonal Function) regression retrieval or the neural network retrieval can be used. Channels in the short-wave region of the IASI spectrum are excluded from the retrieval.
The super-saturation of the retrieved water vapour profile is examined. If super-saturation is found these values are set to the corresponding saturation value. The flag FLG_SUPSAT indicates whether super-saturation occurred and how many levels are affected. The temperature profile is examined with respect to super-adiabatic conditions. If a super-adiabatic temperature gradient is recognised the temperature is set to temperature representing an adiabatic temperature lapse rate. The flag FLG_SUPADI indicates whether super-adiabatic conditions occurred and how many layers are affected.
The results from the first retrieval can constitute the final product or be the first guess for the subsequent iterative retrieval. The state vector from this retrieval contains the temperature, humidity and ozone columns in thick layers, the surface temperature, surface emissivity, and the columnar amounts of carbon monoxide, methane, nitrous oxide and carbon dioxide.
The profiles and surface parameters from the first retrieval are checked for completeness and validated, results are summarised by FLG_FGCHECK.
4.2.3.4 IASI iterative retrieval
If profiles and surface parameters from the first retrieval are within valid bounds (FLG_FGCHECK) they are passed into the Fast Radiative Transfer Model (FRTM) to calculate the synthetic IASI brightness temperature spectra and the Jacobians (a matrix of the partial derivatives of the brightness temperature with respect to the profile and surface parameters). Out of bound values of the first retrieval are replaced by climatology data. Information which is not part of the retrievals but needed as input for the FRTM, e.g. trace gas profiles, is taken from the climatological background.
Based on the pre-selection and availability, the background state vector and its inverse covariance matrix are made available for the retrieval process. The background state vector does not contain information which is used within the iterative retrieval, e.g. ATOVS L2 data are not used as background information if the ATOVS data are used in the iterative retrieval. The flag FLG_SELBAC identifies the used background state vector.
A cost function is then minimised by iteration of the state vector and calculating new brightness temperature spectra and Jacobians until an acceptable solution is found according to predefined criteria.
The information about the convergence of the iteration is given by FLG_ITCONV. The number of iterations is in FLG_NUMIT. The results from the residual check are given by FLG_RESID.
The results of the iterative retrieval are passed to the quality control. A quality control flag (FLG_QUAL) is generated indicating whether the accuracy of the retrieved IASI Level product is in alignment with the EURD requirements.
Table 4.3 summarises the requirements for the IASI Level 2 products with respect to product accuracy, sampling and timeliness as they can be found in the EPS End User Requirements Document (EURD, [RD10]).
| Product | Accuracy | Sampling | Timeliness |
| Temperature | 1 K (2 K stratosphere) | IFOV | 3 h |
| Relative humidity | 10% | IFOV | 3 h |
| Cloud cover | 10% | IFOV | 3 h |
| Cloud top temperature | 2 K | IFOV | 3 h |
| Cloud top height | 300 m | IFOV | 3 h |
| Integrated CH4 | <20% | 250 km | 3 h |
| Integrated N2O | <20% | 250 km | 3 h |
| Integrated CO | <10% | 250 km | 3 h |
Table 4.3: Requirements for IASI Level 2 products
Users should consider carefully the definition of the IASI temperature and humidity profile products, before deciding to use them for their particular application. The IASI sounding products represent thermodynamic states of deep atmospheric layers of variable depths. The reason is the integrating nature of the radiation measurements at the top of the atmosphere. The number of independent pieces of information which are determined in the temperature and moisture profiles are of the order of 14 and 10 respectively, but the number varies with atmospheric situation. Profiles retrieved from radiance measurements are smoothed versions, where the smoothing functions are the so-called averaging kernels.
An example of a set of averaging kernels is shown in Figure 4.5 and Figure 4.6 for temperature and humidity, respectively. Two things can be seen: the vertical extent over which a particular kernel averages, and the amplitude, showing how sharply a kernel peaks at a particular height. Higher amplitudes indicate more information about the corresponding layer. An amplitude of 1 would indicate perfect measurements at a distinct level; however this is purely hypothetical and does not exist. Nevertheless, the retrieved profiles are represented on a fine vertical grid for the reason that the averaging kernels vary with atmospheric situation. Consequently, the vertical resolution and the centre altitudes of the resolved layers vary too. The actual variation is not known a priori, so the retrieval is performed on a fixed, fine pressure grid and the smoothing is represented by the a posteriori error covariance matrix, which is part of the product and represented on the same pressure grid. The off-diagonal elements of the covariance matrix describe the inter-relationship between the state-vector elements and provide information about the actual vertical resolution.
Figure 4.5: Averaging kernels for temperature profiles for a mid-latitude summer situation in Lindenberg 2007
Figure 4.6: Averaging kernels for humidity profiles for a mid-latitude summer situation in Lindenberg 2007
Quality and processing information
The quality of the retrieved profiles depends largely on the setting of the inversion for clear, partly cloudy or cloudy conditions, as shown in the validation report [RD26]. That setting is reflected for each IFOV in flag FLG_IASICLR. That flag can be used in the absence of additional information to screen the clear sky soundings. The FRACTIONAL_CLOUD_COVER parameter can be used to qualify the situation further.
Flag FLG_QUAL may indicate errors outside the EURD threshold. It is sufficient that the error for one of the levels exceeds the limit in order to have that flag set. IASI retrievals do not use much information about the uppermost levels included in the product, consequently the errors for these levels are larger – typically around 10 K. In such situations, the IFOVs are flagged as being outside the EURD threshold, whether or not the error in the lower levels remains within that threshold.
It is planned to provide values of the off-diagonal terms of the error covariance matrix for all levels in the field ERROR_DATA of the product.
The instrument mode is indicated by the IASI Level 1c flag GEPSIasiMode. The flag indicates whether the instrument is operated in normal scan mode or external calibration mode.
The navigation information is given at IFOV level. Information about the satellite's roll, pitch and yaw angles are given by ATITUDE_ANGLES, the time in seconds associated with it is given in TIME_ATTITUDE. The spacecraft's altitude is given in the SPACECRAFT_ALTITUDE variable (values are in km).
The bit string NAVIGATION_STATUS contains further detailed information about the navigational status.
A number of quality and processing information flags are generated during the Level 2 processing. These flags are distributed as part of the IASI Level 2 product and are associated with individual IFOVs. Table 4.4 contains the description of the IASI Level 2 processing and quality flags. The possible values and the conditions are described individually in Section 10. The flags do not only indicate the quality and completeness of the IASI L1c product but also inform about the choices made in the IASI Level 2 processing.
| Flag name | Description |
| FLG_ATOVCLR | Cloud flag for ATOVS products |
| FLG_ATOVCMP | Completeness of ATOVS products |
| FLG_ATOVINT | Interpolation of ATOVS to IASI |
| FLG_AVHAVL | Completeness flag for AVHRR products |
| FLG_AVHBAD | Validation flag for AVHRR products |
| FLG_CHNSEL | Channel selection |
| FLG_CLDAVH | Number of cloud formations analysed in AVHRR |
| FLG_CLDFRM | Number of cloud formations and origin of height assignment |
| FLG_CLDPHA | Cloud Phase |
| FLG_CLDSUM | Summary indicating which instruments see clouds |
| FLG_CLDTST | Cloud tests that are executed |
| FLG_DAYNIT | Discrimination between day and night |
| FLG_FGCHECK | Check whether first guess retrievals passed to iterative retrieval |
| FLG_FINCHC | Final choice of retrieval |
| FLG_FRCSEL | Selection of fractional cloud cover from IASI/ATOVS or AVHRR |
| FLG_IASIBAD | Validation flag for IASI Level 1 product |
| FLG_IASICLD | Results of cloud test |
| FLG_IASICLR | IASI IFOV clear, partly cloudy or cloudy |
| FLG_INITIA | Retrieval initialisation |
| FLG_ITCONV | Convergence of the iterative retrieval |
| FLG_ITRBOU | Validation of iterated state vector |
| FLG_LANSEA | Specifies surface type |
| FLG_NUMIT | Number of iterations used for retrieval |
| FLG_NWPBAD | Validation flag of NWP forecast |
| FLG_QUAL | Quality and completeness of the retrieval |
| FLG_RESID | Retrieval residual acceptance flag |
| FLG_RETBOU | Identification of out-of-bounds state vector elements |
| FLG_RETCHC | Choice of combined or IASI stand-alone retrieval |
| FLG_SATMAN | Indication of satellite manoeuvre |
| FLG_SELBAC | Selection of background state |
| FLG_SFCAVH | Indication of AVHRR derived surface temperature |
| FLG_SFCTOP | Validation flag for surface type and topography |
| FLG_SUNGLNT | Identification of sun glint |
| FLG_SUPADI | Indication of super-adiabatic conditions in final retrieval |
| FLG_SUPSAT | Indication of super-saturation in final retrieval |
| FLG_THICIR | Thin cirrus cloud test |
| FLG_THICOR | Thin cirrus has been corrected for |
| FLG_VARCLR | Cloud clearing by variational analysis |
| FLG_STER | Representation of retrieval errors. Determines content and size of data in MDR.ERROR_DATA. |
Table 4.4: IASI Level 2 processing and quality flags
A detailed explanation on the format is given later in the Level 2 EPS format description (Section 10 and Appendix C.1). Further details can be found in [RD13] and [RD17].
Not available for this issue of the user guide.