4. ATOVS Level 1b Products Overview

Table of Contents

4. ATOVS Level 1b Products Overview

4.1 The ATOVS instrument suite

The ATOVS (Advanced TIROS (Television and Infrared Observational Satellite) Operational Vertical Sounder) is a sounding instrument package first flown on the NOAA-KLM (-15, -16, -17) satellite series. It is composed of the Advanced Microwave Sounding Units A and B (AMSU-A, AMSU-B) and is complemented by the High Resolution InfraRed Sounder (HIRS/3).

For Metop and the NOAA-18 and -19 satellite series, the AMSU-B sounder has been replaced by the Microwave Humidity Sounder (MHS), and the infrared sounder has been upgraded to HIRS/4.

Although not considered formally part of the ATOVS package, the Advanced Very High Resolution Radiometer (AVHRR/3) is an imager also flying on Metop and on NOAA-18 and -19, which supports the ATOVS Level 1b processing. A separate AVHRR Level 1b Product Guide is provided and AVHRR/3 will be discussed here only in the context of its support to the ATOVS products processing.

A detailed account of the ATOVS instruments technical characteristics is given in [RD41], but we will give in the next sections the basic information necessary for product understanding and usage.

4.1.1 AMSU-A Technical description and spectral characteristics

The AMSU-A is a fifteen-channel microwave radiometer that is used for measuring global atmospheric temperature profiles and providing information on atmospheric water in all of its forms.

AMSU-A measures in 15 spectral bands, summarised in the table below, where the temperature sounding mainly exploits the oxygen band at 50 GHz.

Hardware for the two lowest frequencies is located in one module (AMSU-A2) and that for the remaining thirteen frequencies in the second module (AMSU-A1). This arrangement puts the two lower atmospheric moisture viewing channels into one module and the oxygen absorption channels into a second common module, in order to ensure commonality of viewing angle independent of any module and/or spacecraft misalignment due to structural or thermal distortions. The AMSU-A2 module has a single antenna assembly, providing data for channels 1 and 2. AMSU-A1 has two separate antenna assemblies: AMSU-A11 provides data for channels 6, 7 and 9-15, and AMSU-A12 provides data for channels 3, 4, 5 and 8.

The following table summarises the spectral characteristics of AMSU-A.

    Channel frequency
    No. of pass bands
    Nominal bandwidth (MHz)
    Temperature sensitivity
    Calibration accuracy (K)
    1 23.8 1 270 0.30 <2.0 90-θ
    2 31.4 1 180 0.30 <2.0 90-θ
    3 50.3 1 180 0.40 <1.5 90-θ
    4 52.8 1 400 0.25 <1.5 90-θ
    5 53.590.115 2 170 0.25 <1.5 θ
    6 54.40 1 400 0.25 <1.5 θ
    7 54.94 1 400 0.25 <1.5 90-θ
    8 55.50 1 330 0.25 <1.5 θ
    9 FLO= 57.290344 1 330 0.25 <1.5 θ
    10 FLO 0 .217 2 78 0.40 <1.5 θ
    11 FLO 0 .3222
    0 .048
    4 36 0.40 <1.5 θ
    12 FLO 0 .3222
    0 .022
    4 16 0.60 <1.5 θ
    13 FLO 0 .3222
    0 .010
    4 8 0.80 <1.5 θ
    14 FLO 0 .3222
    0 .0045
    4 3 1.20 <1.5 θ
    15 89.0 1 <6000 0.50 <2.0 90-θ
    The polarisation angle is defined as the angle from horizontal polarisation (electric field vector parallel to the satellite track) where θ is the scan angle from nadir. θ indicates horizontal polarisation and 90-θ indicates vertical polarisation.

    Table 4.1: Spectral characteristics of AMSU-A

Each antenna assembly contains a warm calibration target with a different number of Platinum Resistance Thermometers (PRTs), five for the AMSU-A1 modules and seven for the AMSU-2 module. Scanning geometry

AMSU-A is an across-track scanning system with a scan range of ±48.33° with respect to the nadir direction. The instantaneous field of view (IFOV) of each channel is approximately 57.6 milliradians (3.3°) leading to a circular instantaneous field of view size close to 47.63 km at nadir and a swath width of ±1026.31 km (sampling time of 200.0 ms) for a nominal altitude of 833 km. The sampling angular interval is close to 58.18 milliradians (3.3333°). The distance between two consecutive scans is approximately equal to 52.69 km.

There are 30 Earth views, two views of the internal warm target, and two views of cold space per scan line for each channel. Each scan takes 8.0 seconds to complete.

The following table summarises the scanning characteristics.

    Characteristics Value Unit
    Scan direction West to East (northbound) -
    Scan type step -
    Scan rate 8 s
    Sampling interval (duration) 200 ms
    Sampling interval 3.3333 deg
    Pixels/scan 30 -
    Swath ±48.33 deg
    Swath width ±1026.31 km
    IFOV 3.3 deg
    IFOV type circular -
    IFOV size (nadir) 47.63 km
    IFOV size (edge) - across track 146.89 km
    IFOV size (edge) - along track 78.79 km
    Scan separation (adjacent scan lines) 52.69 km

    Table 4.2: Scanning characteristics of AMSU-A Instrument calibration

During each in-orbit scan line, the AMSU-A views three different types of targets:

The accuracy of the warm calibration load brightness temperature is better than ±0.2 K.

The cold space views, together with the internal warm target views and PRT measurements, are used during the ground processing to calibrate the AMSU-A radiances.

4.1.2 HIRS/4 Technical description and spectral characteristics

The HIRS/4 instrument measures the incident radiation primarily in the infrared region of the spectrum in 19 channels, including both longwave (6.5 to 15 µm) and shortwave (3.7 to 4.6 µm) regions, and it also has one channel in the visible (0.69 µm).

The instrument is composed of a single telescope and a rotating wheel with 20 filters. The energy received by the telescope is separated by a dichroic beam splitter into longwave (>6.4 µm) and shortwave (<6.4 µm) energy, controlled by filter stops. The shortwave energy is passed through a second dichroic beam splitter to separate the visible channel. At each of the scan mirror positions, all 20 filter segments are sampled. There are separate sensors for the visible, shortwave and longwave IR energy. The shortwave and visible optical paths have a common field stop, the longwave path has an identical but separate field stop.

If the IR channels data appear to be affected by contamination of the sensors, the IR channel cooler can be heated up to 300 K. During such a decontamination campaign, HIRS/4 IR scans are not produced.

The following table summarises the spectral characteristics of HIRS/4.

    Centre wavenumber (cm-1)
    Centre frequency (µm)
    Half bandwidth (cm-1)
    Maximum anticipated scene temp (K)
    Specified NEΔN
    668.5 1.3
    3.0 +1/-5
    680.0 1.8
    10.0 +4/-1
    690.0 1.8
    12.0 +6/-0
    703.0 1.8
    16.0 +4/-2
    716.0 1.8
    16.0 +4/-2
    733.0 1.8
    16.0 +4/-2
    749.0 1.8
    16.0 +4/-2
    900.0 2.7
    35.0 5.0
    1030.0 4.0
    25.0 3.0
    802.0 2.0
    16.0 +4/-2
    1365.0 5.0
    40.0 5.0
    1533.0 +2/-6
    55.0 5.0
    2188.0 4.4
    23.0 3.0
    2210.0 4.4
    23.0 3.0
    2235.0 4.4
    23.0 3.0
    2245.0 4.4
    23.0 3.0
    2420.0 4.0
    28.0 3.0
    2515.0 5.0
    35.0 5.0
    2660.0 9.5
    100.0 15.0
    14500.0 220
    1000.00 150.0
    100% Albedo
    0.10% Albedo

    Table 4.3: Spectral characteristics of HIRS/4 Scanning geometry

HIRS/4 is an across-track scanning system with a rotating mirror and a scan range of 49.5° with respect to the nadir direction. The instantaneous field of view (IFOV) of each channel is approximately 0.69°, leading to a circular IFOV size close to 10.0 km at nadir for a nominal altitude of 833 19 km. The major difference between HIRS/3 and HIRS/4 is that HIRS/3 has an IFOV size close to 20 km. Each scan line takes 6.4 s to complete. At the end of the scan line, the mirror rapidly returns to its home position (8 retrace steps of 100 ms each) and the scanning pattern is repeated.

There are 56 Earth view samples per scan for a swath width of 1080.35 km (sampling time of 100.0 ms). The sampling angular interval is close to 31.42 milliradians (1.8°). The distance between two consecutive scans is approximately equal to 42.15 km.

The HIRS/4 instrument can be commanded by a Calibration Enable command to automatically enter a calibration mode every 256 seconds (i.e., every 40 scan cycles). This is the nominal instrument mode. If the instrument is commanded by the Calibration Disable command not to perform calibration scans, normal Earth view scans are produced instead of calibration scans.

When the instrument is in the calibration mode, the mirror rapidly slews to a space view position and performs measurements in all channels for the equivalent time of one complete scan line. Due to the time required to bring the mirror into its space view position, the first 8 scan steps are not usable, reducing the number of usable space scan steps to 48. Next, the mirror is moved to a position where it views the warm internal calibration target and data are taken for the equivalent time of 56 scan steps. Upon completion of the calibration mode, the mirror continues its motion to home position, where it begins a normal Earth scan. The total calibration is equivalent to two scan lines (no Earth data are acquired during this period). Therefore, there will be 2 lines of calibration data followed by 38 lines of Earth view data, forming a so-called calibration cycle.

The following table summarises the scanning characteristics.

    Characteristics Value Unit
    Scan direction West to East (northbound) -
    Scan type continuous -
    Scan rate 6.4 s
    Sampling interval (duration) 100 ms
    Sampling interval 1.8 deg
    Pixels/scan 56 -
    Swath ±49.5 deg
    Swath width ±1080.35 km
    IFOV 0.69 deg
    IFOV type circular -
    IFOV size (nadir) 10.0 km
    IFOV size (edge) - across track 33.27 km
    IFOV size (edge) - along track 17.03 km
    Scan separation 42.15 km

    Table 4.4: Scanning characteristics of HIRS/4 Instrument calibration

IR calibration of the HIRS/4 is provided by programmed views of two radiometric targets: the warm target mounted on the instrument baseplate, and a view of space. Data from these views provide sensitivity calibrations for each channel at 256-second intervals if commanded. Internally generated electronic signals provide calibration and stability monitoring of the detector amplifier and signal processing electronics.

During each calibration cycle, the HIRS/4 views three different types of targets:

The calibration repeatability is specified to be better than 0.3 K and the inter-channel accuracy better than 0.2 K.

The cold space views, together with the internal warm target views and PRT measurements, are used during the ground processing to calibrate the HIRS/4 radiances.

Note that two on-board targets at different temperatures are available on HIRS instruments and that temperature measurements are carried out for both targets. However, it was found on HIRS/2 that the cold on-board target did not improve the calibration and was largely not used in the ground processing. Therefore, only the warm target is used in the operational HIRS/4 calibration sequence. The view of the second cold on-board target is only selectable by command, and it is not planned to use this command for HIRS/4 operations.

Finally, it is important to mention that additional radiation coming from different parts of the instrument is also measured by the HIRS/4 detector. For that reason, the instrument temperature is carefully controlled. The exception is the baffle (secondary mirror/telescope), which due to its composition materials is subject to relatively high short-term temperature variations. The baffle temperature is measured every scan line and an additional correction for this effect can be applied in the ground calibration processing.

4.1.3 MHS Technical description and spectral characteristics

The MHS is the follow-on instrument to the Advanced Microwave Sounding Unit-B (AMSU-B) which flew as a part of ATOVS on the NOAA-KLM satellite series. It is procured by EUMETSAT for the Metop and NOAA satellites.

MHS is a five-channel microwave radiometer, which complements the Advanced Microwave Sounding Unit-A (AMSU-A) channels. In some MHS descriptions documents, MHS channels may be numbered as a continuation of the AMSU-A channels: 16, 17, 18, 19 and 20.

It is planned to derive from these frequencies humidity profiles and cloud liquid water content. Additionally, the instrument's sensitivity to large water droplets in precipitating clouds can provide a qualitative estimate of precipitation rates.

It is technically similar to the AMSU-B instrument, except for channel 20, where the AMSU-B side-band at 176.31 GHz is missing.

The following table summarises the spectral characteristics of MHS.

    Central frequency
    Temperature sensitivity
    Calibration accuracy

    Table 4.5: Spectral characteristics of MHS Scanning geometry

MHS is an across-track scanning system with a scan range of ±49.44° with respect to the nadir direction. The IFOV of each channel is approximately 19.2 milliradians (1.1°) leading to a circular instantaneous field of view size close to 15.88 km at nadir for a nominal altitude of 833 km. Each scan takes 2.667 seconds to complete.

The scan of the MHS instrument is synchronised with the AMSU-A scan, i.e. there are three scans of MHS for each scan of AMSU-A.

There are 90 Earth samples per scan and per channel for a swath width of ±1077.68 km (sampling time of 19.0 ms). The sampling angular interval is close to 19.39 milliradians (1.1111°), which is slightly larger than that of AMSU-B (1.1000°). The distance between two consecutive scans is approximately equal to 17.56 km.

The following table summarises the scanning characteristics.

    Characteristics Value Unit
    Scan direction West to East (northbound) -
    Scan type continuous -
    Scan rate 2.667 s
    Sampling interval (duration) 18.52 ms
    Sampling interval 1.1111 deg
    Pixels/scan 90 -
    Swath ±49.44 deg
    Swath width ±1077.68 km
    IFOV 1.1 deg
    IFOV type circular -
    IFOV size (nadir) 15.88 km
    IFOV size (edge) - across track 52.83 km
    IFOV size (edge) - along track 27.10 km
    Scan separation 17.56 km

    Table 4.6: Scanning characteristics of MHS Instrument calibration

The MHS instrument calibration is based upon the measurement of cold space and of an on-board black body target. This calibration sequence is performed once every 2.667 seconds for each scan line. During one scan, MHS observes

The warm target contains five platinum resistance thermometers (PRTs), as opposed to the seven PRTs for the older NOAA AMSU-B instrument.

The cold space views, together with the internal warm target views and PRT measurements, are used during the ground processing to calibrate the MHS radiances.

4.2 Overview of the ground processing and calibration

The Level 1 ground processing chains for the AMSU-A, MHS and HIRS/4 are illustrated in Figure 4.1 below.


Figure 4.1: Functional overview of the ATOVS Level 1 ground processing chain

The first objective of Level 1 ground processing for the ATOVS instruments data is the generation of the AMSU-A, MHS and HIRS/4 Level 1b products, containing radiances as the main geophysical parameter. This processing is data driven and is applied to every science source packet.

Additionally, ATOVS Level 1b products are considered an input to the EUMETSAT ATOVS and IASI Level 2 processors. Furthermore, AVHRR/3 Level 1b generated at EUMETSAT is used as an input for HIRS Level 1b processing. In Section 10, the context of the ATOVS, IASI and AVHRR processing chain interactions is provided for information.

The ATOVS ground processing at EUMETSAT is applied to data from the ATOVS instruments on both Metop and NOAA satellites.

4.2.1 Pre-processing

Pre-processing of ATOVS data has similar objectives for data coming from AMSU-A, MHS and HIRS/4, namely validation of input data, geolocation of each pixel, estimation of solar and satellite azimuth and zenith angles and finally, computation of calibration coefficients, which will be later applied in the Level 1b processing to calculate radiances.

Basic raw data validation checks are applied and the instruments' telemetry and other auxiliary data are also validated and related to the input raw data flow. Geolocation

Geolocation for individual pixels along the scan line is performed, as well as their satellite and solar zenith and azimuth angles. The navigation of individual pixels is carried out by means of a satellite ephemeris model and an instrument scanning model. This first step in the navigation is based on default attitude values, directly after data acquisition. Please refer to Appendix B for further information on satellite orbit and attitude models.

The calculation of the satellite zenith and azimuth is done by applying a transformation matrix to the Earth fixed satellite position coordinates previously obtained during the pixel navigation. The solar azimuth and zenith are obtained taking into account the actual solar declination, for which an accurate time stamp for the scan line is necessary.

Using a high resolution coastline data set and the geolocation information estimated above, a surface type is assigned to each pixel.

Full details on geolocation processing are provided in [RD20]. Computation of calibration coefficients for AMSU-A and MHS

In the case of AMSU-A, and MHS, each Earth scene radiance is related to the Earth view counts for that scene through a non-linear relationship involving a linear gain, to which a non-linearity correction is applied in terms of an additional quadratic term. Several steps are involved in the calibration processing:

The corrections of the internal warm target temperature are intended to compensate for expected instrument temperature effects. The instrument temperature is measured on board and reflected in the instrument telemetry.

The corrections of known fixed temperature of cold space are necessary to compensate for the effect of contamination by radiation coming from the spacecraft and the Earth limb. Additionally, a moon glint correction is applied. Computation of calibration coefficients for HIRS/4

Concerning HIRS/4, computation of the calibration coefficients of all thermal channels and the visible channel is also performed during the pre-processing.

The HIRS/4 IR channels calibration processing is also based on a two-point calibration scheme, using a warm on-board target of known temperature and deep space view. In this case however, calibration counts are only available every 40 scan lines. Instantaneous calibration coefficients are calculated every time a scan of calibration data is available, and those corresponding to the Earth view scans of that cycle are estimated by interpolation between the instantaneous calibration coefficients of the present and previous cycle.

In general, the relationship between Earth view counts and radiances is assumed to be quadratic. The coefficient of the quadratic term is assumed to be invariant and determined prior to launch as part of the instrument on-ground characterisation. It is usually set to 0 for all channels. The following steps are involved in obtaining the coefficients of the linear part of the equation.

The HIRS/4 visible channel calibration consists of a linear equation, whose coefficients are determined prior to launch as part of the instrument on-ground characterisation and rarely change. Determination of cloud coverage for a HIRS/4 scene

The determination of the cloud coverage consists in determining which AVHRR/3 pixels fall into the HIRS/4 IFOV and calculating from the cloudy/clear information in the AVHRR/3 Level 1b pixel data set, the percentage of the clear AVHRR/3 pixels. This method involves therefore two steps: AVHRR/3-HIRS/4 measurement collocation and AVHRR/3 clear pixel counting.

4.2.2 Level 1b processing

Calibration coefficients are applied to both visible (linear equation) and IR channels (quadratic equation, but with the quadratic term set to zero), in order to convert channel count values into reflectivity and radiances, respectively. These are the geophysical parameters which constitute the HIRS/4 Level 1b products. For AMSU-A and MHS, the quadratic calibration equation is applied to all channels, converting counts to radiances. Then, an antenna correction is applied.

4.2.3 Quality control

This function covers both the radiometric and the geometric quality assessment. The radiometric quality assessment consists of the production of a detailed set of radiometric characteristics of the data for each detector/channels, this for different imaged scenes during the dump (day/night sides, calibration viewing, etc.).

For AMSU-A and MHS, the geometric quality control consists of a check for every scan line on the Earth's view antenna position against a set of pre-defined thresholds. Geometric quality control is not performed operationally for the HIRS/4 instrument.

Finally, statistics produced by the quality control function are used to perform trend analysis and to derive information on the misalignments between instruments and mis-registration between channels. Updates of the model parameters for the platform/instrument being processed are then estimated and this information allows compensating for slow drifts and changes in these parameters.

4.3 ATOVS Level 1b product characteristics and use

4.3.1 General characteristics

Table 4.7 summarises the main characteristics of ATOVS Level 1b products available to users. All products contain quality control and other information about the retrieval and their use, which are important to know when you choose the product needed for your application. Two different types of Level 1b products are generated, from Metop and from NOAA data.

 Instrument  Product Main geophysical parameter Accuracy Resolution
/grid spacing (nadir)
Swath width Coverage Generated
Level 1b from Metop
Geolocated radiances
<1 K
47.3 km
30 pixels per scan, 52.69 km scan separation along track
2053 km Global and continuous EPS CGS
Level 1b from NOAA
Level 1b from Metop
Geolocated IR radiances and reflectivity for channel 20
<1 K for IR channels
10.0 km
56 pixels per scan, 42.15 km scan separation along track
2161 km
Level 1b from NOAA
MHS Level 1b from NOAA
Geolocated radiances
<1 K
15.88 km
90 pixels per scan, 17.56 km scan separation along track
2155 km

Table 4.7: Summary of the main characteristics of ATOVS Level 1b products

Apart from the main geophysical parameters given for each pixel, navigation information per scan line and geolocation information for every pixel is given, as well as angular relations for every navigation point. Calibration data are also appended to the product for all channels.

In the case of the HIRS/4 Level 1b product, the percentage of clear sky as derived from the analysis of geolocated AVHRR/3 data is also included.

4.3.2 Quality information in the products

A number of quality flags are generated during the Level 1b processing, associated with individual scan lines. The following are the most relevant with respect to data use. A full list and detailed explanation of all flags is given later in the record contents and format descriptions for the AMSU-A, HIRS/4 and MHS Level 1b products (Sections 11, 12 and 13 respectively).

4.4 Summary of ATOVS Level 1b product current and potential applications

The main internal use of the ATOVS Level 1b product is for further processing in the ATOVS and IASI Level 2 processors in the EPS CGS.

In particular, higher level products of direct application in meteorology and climatology are derived from ATOVS Level 1b, such as temperature and humidity profiles, cloud top temperature and pressure, effective cloud amount, total ozone column, cloud cover. Applications of those geophysical parameters are discussed in the ATOVS Level 2 Product Guide.

Direct applications of ATOVS Level 1b products consist of direct assimilation of radiances in NWP. In most major NWP centres, satellite data are introduced into the assimilation process, which is the first step in a forecast cycle. The satellite observations are used as radiances, i.e. directly after calibration, and not as derived parameters as temperature profiles or similar. In an assimilation system, a variable measured within a certain time window (cut-off window) is processed to correct a so-called background state (also called first guess), which is in most cases the result of a short-term forecast from the previous analysis step. The assimilation process provides a series of corrections of this background state closer to the observations, hence correcting the previous forecast trajectory in the phase space (see [RD42] for more information).

The assimilation of satellite data is particularly different from the assimilation of "conventional" data. Measured quantities (radiances) are not related directly to model quantities. Hence model variables have to be adjusted within the processing to simulate the measurements as well as possible, i.e., there is the need for a radiative transfer model, or better, an observation operator to simulate the radiances from the model variables. Furthermore, the incoming data are difficult to use and may require the clearing from cloud effects and similar. Bias corrections need to be introduced before the assimilation process can start. A good knowledge of the error characteristics of the measurements is needed as well. Strict monitoring and quality control is required.

This advance in data assimilation techniques, alongside other improvements in the quality and range of satellite observations and our understanding of how to model them, has led to a situation where satellite measurements are a vital and integral part of the global observing system in all regions, not just those where other observations are sparse. ECMWF, the Met Office and Météo-France assimilate ATOVS radiances operationally in their models.

Assimilation of raw (cloud cleared) radiances, rather than higher level products, allows more accurate representation of the scanning geometry and of the impact of cloudiness, better quality control and faster usage of data.