ESA Earth Home Missions Data Products Resources Applications
    24-Jul-2014
EO Data Access
How to Apply
How to Access
Index
Credits
AATSR Data Formats Products
Records
SST record 50 km cell MDS
BT/TOA Sea record 17 km cell MDS
ATS_TOA_1P_MDSR_conf
ATS_TOA_1P_MDSR_cl
ATS_TOA_1P_ADSR_pix
ATS_SST_AX_GADSR
Vegetation fraction for Land Surface Temperature Retrieval GADS
Topographic Variance data for Land Surface Temperature Retrieval GADS
Land Surface Temperature retrieval coefficients GADS
General Parameters for Land Surface Temperature Retrieval GADS
Climatology Variance Data for Land Surface Temperature Retrieval GADS
Level 0 SPH
Level 0 MDSR
Auxilliary Data SPH with N = 1
1.6 micron nadir view MDS
Summary Quality ADS
Scan pixel x and y ADS
Grid pixel latitude and longtitude topographic corrections ADS
Across-track Band Mapping Look-up Table
Configuration Data GADS
Processor configuration GADS
LST record 50 km cell MDS
Distributed product MDS
Level 2 SPH
SPH
10-arcminute mds
Limits GADS
Validation Parameters GADS
BT/TOA Land record 17 km cell MDS
General Parameters GADS
Temperature to Radiance LUT GADS
Radiance to Brightness Temperature LUT GADS
Medium/High Level Test LUT GADS
Infrared Histogram Test LUT GADS
11 Micron Spatial Coherence Test LUT GADS
11/3.7 Micron Nadir/Forward Test LUT GADS
11/12 Micron Nadir/Forward Test LUT GADS
Characterisation GADS
Browse Day_Time Colour LUT GADS
Browse SPH
Grid pixel latitude and longtitude topographic correction ADS
Level 2 SPH
Auxilliary Products
ATS_VC1_AX: Visible Calibration data
ATS_SST_AX: SST Retrieval Coeficients data
ATS_PC1_AX: Level-1B Processing configuration data
ATS_INS_AX: AATSR Instrument data
ATS_GC1_AX: General Calibration data
ATS_CH1_AX: Level-1B Characterization data
ATS_BRW_AX: Browse Product LUT data
Level 0 Products
ATS_NL__0P: AATSR Level 0 product
Browse Products
ATS_AST_BP: AATSR browse image
Level 1 Products
ATS_TOA_1P: AATSR Gridded brightness temperature and reflectance
Level 2 Products
ATS_NR__2P: AATSR geophysical product (full resolution)
ATS_MET_2P: AATSR Spatially Averaged Sea Surface Temperature for Meteo Users
ATS_AR__2P: AATSR averaged geophysical product
Frequently Asked Questions
The AATSR Instrument
Instrument Characteristics and Performance
In-flight performance verification
Instrument Description
Internal Data Flow
Instrument Functionality
AATSR Products and Algorithms
Common Auxiliary data sets
Auxiliary Data Sets for Level 2 processing
Instrument Specific Topics
Level 2 Products
Level 1B Products and Algorithms
Level 1B Products
Algorithms
Instrument Pixel Geolocation
Availability
The Level 0 Product
Differences Between ATSR-2 and AATSR Source Packets
Definitions and Conventions
Conventions
Organisation of Products
Relationship Between AATSR and ATSR Products
AATSR Product Organisation
Data Handling Cookbook
Characterisation and Calibration
Monitoring of AATSR VISCAL Parameters
Latency, Throughput and Data Volume
Throughput
Introduction
Heritage
Data Processing Software
Data Processing Centres
The AATSR Products User Guide
Image Gallery
Breakup of the Ross Ice Shelf
Land cover in the Middle East
Typhoon Saomai
Mutsu Bay, Japan
Deforestation in Brazil
Spatially Averaged Global SST, September 1993
Further Reading
How to use AATSR data
Why Choose AATSR Data?
Why Choose AATSR Data?
Special Features of AATSR
Principles of Measurement
Scientific Background
The AATSR Handbook
SST record 17 km cell MDS
Surface Vegetation class for Land Surface Temperature Retrieval GADS
1.6 micron forward view MDS
12 micron nadir view MDS
12 micron forward view MDS
Summary Quality ADS
Surveillance Limits GADS
Master Unpacking Definition Table GADS
1.6 micron Non-Linearity Correction LUT GADS
General Parameters GADS
Thin Cirrus Test LUT GADS
Fog/low Stratus Test LUT GADS
1.6 Micron Histogram
Browse MDS
ATS_CL1_AX: Cloud LUT data
Glossary
Pre-flight characteristics and expected performance
Payload description, position on the platform
Auxiliary products
Auxiliary Data Sets for Level 1B processing
Summary of auxiliary data sets
Calculate Solar Angles
Image Pixel Geolocation
Level 0 Products
Acquisition and On-Board Data Processing
Product Evolution History
Hints and Algorithms for Higher Level Processing
Data Volume
Software tools
Summary of Applications vs Products
Geophysical Coverage
Geophysical Measurements
ATS_TOA_1P_ADSR_sa
Visible calibration coefficients GADS
Level 1B SPH
LST record 17 km cell MDS
Conversion Parameters GADS
12 Micron Gross Cloud Test LUT GADS
ATS_PC2_AX: Level-2 Processor Configuration data
Level 2 Products
Hints and Algorithms for Data Use
BT/TOA Sea record 50 km cell MDS
BT/TOA Land record 50 km cell MDS
Level 2 Algorithms
Signal Calibration
Services
Site Map
Frequently asked questions
Glossary
Credits
Terms of use
Contact us
Search


 
 
 


2.6.1 Algorithms

The Level 1B algorithms are described in detail in the following sections:

2.6.1.1 ATS_TOA_1P

Figure2.3 shows an overview of the Level 1B processing.

Figure 2.3 Overview of the Level 1B processing.

Level 1B processing flow diagram (17K)

The AATSR Level 1B processor will operate on segments of AATSR Level 0 data of duration up to slightly more than one orbit.

The Level 0 data product comprises a series of records presented in chronological sequence, each of which contains a single instrument source packet. Each source packet represents a single instrument scan. For the most part (with one significant exception), the data processing up to but not including the re-gridding stage treats each source packet independently. Thus each processing module can be regarded as a looping over a series of source packets, performing the same operations on (or in connection with) each.

The exception noted above refers to the derivation of the channel calibration coefficients. In the case of the infra-red channels, the calibration coefficients are determined by averaging calibration target data over a number of consecutive scans; the time interval containing these scans will be termed a calibration period or calibration interval. The recommended duration of the calibration interval is 10 scans; thus a new calibration interval will start at the end of the previous interval, and will continue for 10 scans or until a change in the pixel selection map is detected, whichever is the earlier.

For the visible channels, the channel offsets are derived by averaging over a calibration interval as above. The channel gains, however, are determined once per orbit, by averaging over the block of scans for which the VISCAL target is visible.

2.6.1.1.1 Source packet Processing

Initial processing must convert the data into the appropriate form for the higher levels of processing that geolocate the data and generate the products. The main functions can be summarised as follows:

  • Perform basic quality checks on each raw packet, ensuring that only those that pass the checks continue on to the calibration process.
  • Unpack all the auxiliary and housekeeping data containing the temperatures of the on-board black bodies and instrument health and status information.
  • Validate the unpacked auxiliary and housekeeping data items.
  • Convert the auxiliary data to engineering units where applicable.
  • Validate the converted auxiliary data items, especially those which are vital to the calibration.
  • Unpack and validate the science data containing the earth view and black body view pixel counts for all available channels from each packet.

2.6.1.1.2 Infra-Red Channel calibration

Derivation of the calibration parameters requires the calculation of the gain and offset values which define a linear relationship between pixel count and radiance. The gain and offset are derived for each channel, from the unpacked and validated auxiliary and pixel data averaged over a number of contiguous data packets.

The IR calibration module is a framework in which all the modules described in earlier sections are used. Up until this point all modules have been concerned with processing component data from a single source packet. The IR calibration calculations are concerned with using the unpacked, converted and validated auxiliary and pixel data generated by the modules previously described, to calculate average black body temperatures and average black body pixel counts over a series of contiguous packets. The number of packets over which these values are averaged is called the "calibration period".

Once the average values for both black body temperatures have been calculated from the data available in the processed source packets in a calibration period, and the corresponding black body pixel counts have been similarly averaged, it is then possible to use these values to derive the gain and offset parameters which can then be used to convert pixel count to radiance for all valid pixels in all valid packets within the calibration period.

The look-up table for conversion from temperature to radiance contains a correction for detector non-linearity to ensure a linear relationship between radiance and detector counts. It is therefore possible to derive the calibration parameters for each IR channel from the straight line formula y = m.x + c. The mean (odd and even) black body pixel counts for the two black bodies are the two x axis values, and the two black body radiance values make up the corresponding y axis values.

The gain and offset values are derived for both odd and even pixel numbers, because AATSR uses two integrators; one for odd pixels and one for even pixels. The response time of a single integrator is too slow to use the same one for all pixels.

Details of the calibration algorithm are contained in the following section:

2.6.1.1.2.1 Infra-Red Calibration

2.6.1.1.2.1.1 Physical Justification

2.6.1.1.2.1.1.1 Overview

The calibration of the thermal channels aims to characterise, for each channel separately, the relationship between the radiation incident on the detector and the detector output. The signal in counts from a radiometer channel whose spectral passband is Delta (1K) nu (1K) observing a blackbody target at temperature Tbb is

Equation 1 (1K) eq 2.1

where G is the radiometric gain, L(Tbb) is the radiance from a target, and S 0 is the radiometric offset of the channel. Thus radiometric calibration of the instrument consists of determining the linear relationship between the radiance and detector counts from each channel.

One way to do this might be to allow the instrument to view a zero radiance target, such as a cold space view, to determine the radiometric offset S 0 (i.e., L(Tbb) = 0 ). Then the radiometer views a hot calibration target to determine the radiometric gain of the channel, given by

Equation 2 (1K) eq 2.2

The major limitation of this approach comes from the assumption that the radiometer's response is linear over a wide range of scene temperatures, 0 to 350K in the case of an Earth viewing instrument. In practice there is always some non-linearity that, if not treated properly in the ground processing algorithms, results in errors in calibration.

A different approach is therefore adopted for AATSR to minimise the sensitivity of the calibration to any non-linearity in the radiometer characteristics. This has been done both by careful design of the signal processing electronics and careful pre-flight determination of the non-linearity for beginning of life and end of life conditions on the satellite, and also through designing the calibration system so that the on-board calibration is optimised over the limited range of temperatures that span the expected range of SST observations. This is done by the use of two blackbody calibration targets, rather than a single target hot target and space view. One of these targets operates at a temperature lower than the lowest expected SST and the one other warmer than the highest. With this arrangement the calibration is most precise over the normal range of observed temperatures, and the effects of any non-linearity in the system are minimised because linearity is only assumed over a small range of measurement space. Outside this range the calibration is no worse than using the space view and single target approach, but the precision is concentrated into the portion of the measurement space where the most accurate measurements are required. Outside this range the precision of the observations is less critical, so the larger calibration errors resulting from extrapolation can be tolerated.

The signal from the cold blackbody is given by,

Equation 3 (1K) eq 2.3

and that from the hot blackbody is given by,

Equation 4 (1K) eq 2.4

Hence, the radiometric gain G is

Equation 5 (1K) eq 2.5

and by substituting G back into the equations the offset S 0 is

Equation 6 (1K) eq 2.6

Thus, by using this two blackbody calibration approach we can determine the radiometric gain and offset of each of the AATSR channels, in a way that allows us to achieve the highest accuracy in SST with minimal correction for signal channel non-linearity.

In practice the situation is slightly more complex than this. Firstly, the blackbody targets used for calibration cannot be made perfectly black so they also reflect radiation and this component of the signal must be included in the calibration. Secondly, despite the care taken in designing the signal channel electronics in AATSR there is still some non-linearity in the response of the detectors to variations in scene radiance. These problems are discussed in more detail below, where the theoretical treatment follows that developed for ATSR by Mason (1991) Ref. [1.12 ] .

2.6.1.1.2.1.1.2 Detector Response to source radiation

The instrument fore-optics is designed to focus an image of the scene on the field stop. The dimensions of the field stop define the angular resolution, or instantaneous field of view, of the instrument (the angular width of the pixel); thus the solid angle subtended by the field stop at the primary mirror determines the solid angle Omega (1K) within which energy is accepted by the instrument. Then the energy falling on the nominal aperture A of the instrument from a scene of brightness temperature T is

Equation 7 (1K) eq 2.7

within a wavelength interval d lambda (1K), where B( lambda (1K), T) is the Planck function representing the radiance per unit wavelength interval of black body radiation of temperature T. In wavelength units the Planck function is

Equation 8 (1K) eq 2.8

Ideally all this energy should be focused on the detector; in practice the energy reaching the detector is reduced by absorption at the surfaces of the scan and primary mirrors, by diffraction at the field stop, and by absorption within the focal plane assembly (FPA). The energy entering the FPA, in the wavelength interval d lambda (1K), is

Equation 9 (1K) eq 2.9

where tau Au squared (1K) is the square of the reflectivity of the Rhodium coating of the mirrors and where the product throughput equation (1K) represents the throughput of the spectral channel lambda (1K). Note that throughout this discussion the use of lambda (1K) as a suffix labels the AATSR channel; the use in any other context denotes a frequency dependence. epsilon lambda (1K) represents the factor by which the instrument throughput in the channel is reduced below its nominal value (A Omega (1K)) owing to diffraction at the field stop and to any reduction in the aperture of the parabolic mirror below the nominal value A. A corresponding amount of radiation (1 - throughput (1K)) originating from the paraboloid surround thus enters the FPA, and contributes to the radiometric offset.

We now introduce the quantity spectra response (1K) representing the spectral response of the channel lambda (1K); spextral response (1K) is the detector response to monochromatic incident energy at wavelength lambda (1K), normalised to a peak value of unity. It is determined for each channel by laboratory measurements on the FPA during ground characterisation of the instrument. Then the response of the detector to the incident scene energy is proportional to

Equation 10 (1K) eq 2.10

where tau fpa (1K) is the transmission of the FPA at the peak of the spectral response of the channel.

The expression (10) is proportional to the photon flux incident on the detectors. For a given channel the integral

Equation 11 (1K) eq 2.11

is a function of the scene brightness temperature only; we may term it the scene radiance. [Strictly speaking it is not a true radiance, because the detectors act as photon counters rather than as thermal detectors, and our definition of R takes account of this. If the detectors responded to incident energy, R should represent the optical transmission of the components of the FPA, and the expression would be a true radiance. As it is, R also incorporates a term proportional to wavelength, to transform the energy in an elementary wavelength interval into the number of photons. Nevertheless, the effect of this is very small over the narrow bandwidths of the AATSR channels; moreover as we have defined it, with the normalisation of R, it has the dimensions of a radiance, and following Mason (1991) Ref. [1.12 ] we shall refer to it as such.]

The total photon flux falling on the detectors can thus be written

Equation 12 (1K) eq 2.12

where background signal (1K) represents the background signal that results from energy originating within the instrument reaching the detectors, including emissions from the scan and primary mirrors and from the paraboloid surround. It depends on the temperature of the instrument components and can be assumed to be constant over the period of validity of a calibration.

2.6.1.1.2.1.1.3 Effect of signal channel processing circuits

The signal output from each detector is amplified within a preamplifier stage and passed to a signal channel processing (SCP) subsystem located within the Instrument Electronics Unit 3.1.2.2. (IEU). The input stage of each SCP channel incorporates a digitally controlled variable offset, and is followed by further amplification the gain of which is also under digital control. These variable gain and offset functions permit the signal levels entering the digitisation stage to be adjusted to permit the full dynamic range of the digitisers to be used. The control words that determine the gain and offset are provided by the DEU. The gain and offset are generally controlled by a feedback loop (the so-called 'autocal loop') operating within the DEU, which monitors the black-body temperatures and operates to maintain the output counts in each channel at predetermined optimum levels. Alternatively the loop may be disabled and the gain and offset set by command. In either case the output signal is a linear function of the input.

After amplification the signal in each channel is integrated for 75 µs; owing to the short integration times used two integrators are provided for each channel; these operate alternately so that one is read out and reset while the other is integrating. The integrator outputs are then digitised with 12 bit resolution, and the resulting samples passed to the telemetry. The sampled signal (in counts) in channel lambda (1K) is given by

Equation 13 (1K) eq 2.13

where gain (1K) and offset (1K) represent the SCP gain and offset of channel lambda (1K) respectively. Introducing equation eq. 2.12 and gathering together the constant terms we can write

Equation 14 (1K) eq 2.14

where channel constant (1K) and channel constant (1K) are constants characteristic of the channel.

2.6.1.1.2.1.1.4 Determination of the calibration coefficients

Equation eq. 2.14 can be inverted to

Equation 15 (1K) eq 2.15

where constant (1K) and constant (1K) are constants. If these constants are known, then equation eq. 2.15 can be used to convert the measured pixel counts into the scene radiance L(T); T can then be recovered by inverting equation eq. 2.11 . Note that once the spectral response function R has been determined for a given channel, the radiance function ( eq. 2.11 ( and its inverse can be computed as look-up tables that can be used to convert from radiance to brightness temperature and vice versa.

The linear relationship eq. 2.15 is be defined if two points on the line are known. These points in turn are defined by the two black body calibration targets.

Figure 2.4 Radiance versus Counts

Figure (2K)

In the case of AATSR each calibration black body has 5 platinum resistance thermometers (PRT) mounted on the base, and a additional PRT on the shroud. The temperature of each black body can be determined using these. Given the temperatures of the two black bodies, their radiances can be determined; hence given a simultaneous measurement of the detector counts from the black body pixels, we can plot the points on the characteristic curve. The two black bodies taken together allow us to determine the bias and slope of the calibration characteristic.

Suppose that the radiances of the cold and hot black bodies are L1 and L2 respectively, and the corresponding counts are S1 and S2. Then

Equation 16 (1K) eq 2.16

and

Equation 17 (1K) eq 2.17

Note that these expressions are unchanged if the suffixes 1 and 2 are interchanged; it is thus not necessary to specify explicitly which is the hot black body and which is the cold. The same expressions would have been obtained if we had defined L1 to be the radiance of the hot black body, and L2 that of the cold black body. In practice the choice of which black body is to be the cold and which the hot target can be determined by macro-command; however, because of the above invariance we do not need to know which selection has been made at the time of the analysis. We can redefine L1 as the radiance of (say) the +X black body, and L2 that of the -X black body1, and still use these equations to determine the calibration coefficients. This is the approach that is adopted in the analysis.

(1 The two black bodies are conventionally known as the +X black body and the -X black body. This identifies them by their position in the satellite reference frame.)

If the calibration target were an ideal black body, its radiance would be given by evaluating equation eq. 2.11 at the physical temperature of the target. In practice this is not so; if the emissivity of the black body j (j = 1, 2) in the channel of wavelength lambda (1K) is emissivity (1K), assumed constant across the bandwidth of the channel, then the corresponding calibration signal is

Equation 18 (1K) eq 2.18

where BB temperature (1K) and instrument temperature (1K) represent the temperatures of the black body j and the instrument environment respectively.

A number of practical points arise. Firstly, the odd and even pixels are measured with different amplifier chains. Since the calibration includes the gain of the amplifier electronics, this has the consequence that separate calibration coefficients are required for the odd and even channels. Thus odd and even black body pixels are analysed separately to give the two calibrations.

Both the pixel counts and the PRT temperatures are averaged over a suitable calibration interval to reduce their respective noise levels; this is currently 10 scans. Thus for each black body and each pixel parity (that is, for the odd and even pixels separately) the pixel counts are averaged over the scans in the chosen calibration interval to give a mean count. The platinum resistance thermometer values are averaged over the same interval to give the corresponding temperatures BB temperature (1K).

2.6.1.1.2.1.1.5 Signal Channel Calibration

Once the slope and intercept (the coefficients of equation eq. 2.15 ) are known, pixel calibration for the corresponding channel is accomplished by the use of equation eq. 2.15 . Given the measured the counts for each channel, we derive the incident energy (radiance) from this equation directly. Then we derive the corresponding brightness temperature by inverting equation eq. 2.11 . In practice this is accomplished with a look-up table that inverts the equation.

2.6.1.1.2.1.1.6 Detector non-linearity

In the description of the calibration scheme above, we have assumed that each detector output depends linearly on the scene radiance within its operating range. This is not strictly true. The 11 and 12 µm channels of AATSR use photoconductive cadmium mercury teluride (CdHgTe) detectors, and these show small but detectable non-linearity. This is accommodated within the calibration scheme by modifying the look-up tables using pre-flight characterisation measurements of the detector response. We have assumed that the detector output is proportional to the scene radiance, whereas in fact it depends on a non-linear function f{L}. If the look-up table representing the radiance function L(T) is replaced by one representing the composite function f{L(T)} then it is evident that the scheme outlined above would work equally well.

A complication is introduced when the non-zero reflectivity of the actual calibration targets is taken into account. In this case the calibration signal from the black body target j is given by

eqn_19.gif (1K) eq 2.19

in place of equation eq. 2.18 , and, because the function f is by definition non-linear, it is not strictly possible to decompose this expression into a form that can be represented exactly by a simple look-up table. The equation eq. 2.19 is a function of two arguments BB temperature (1K) and instrument temperature (1K), and to evaluate it exactly it would be necessary either to use a table with two arguments, or to have two tables to represent the functions f and T separately, rather than a single table. In practice this complication is not necessary, because of the small value of emissivity (1K) and the relatively slight non-linearity in question; to a sufficient approximation we can write

Equation 20 (1K) eq 2.20

as the counterpart of equation eq. 2.18 . In other words we can still base the calculation on the single modified look-up table. This is discussed in more detail below.

2.6.1.1.2.1.1.7 Derivation of the look-up tables

The derivation of a table representing the radiance as a function of brightness temperature is straightforward; given measured values of the spectral response function R, for a channel the integral eq. 2.18 can be evaluated for any temperature T using standard techniques of numerical quadrature. In order to correct the resulting tables for detector non-linearity in the case of the 11 and 12 µm channels, the following approach, originally developed for ATSR by (Mason 1991 Ref. [1.12 ] ), is adopted.

The detector output is measured during ground characterisation in an environmental test facility with reference to external black bodies of known temperature and emissivity. The (normalised) detector output is plotted against the calculated radiance of the external targets. A straight line is then fitted to the data points of low radiance and this fitted line is regarded as the ideal detector characteristic, it being assumed that the detector response is linear at low incident radiance. The ratio of the actual detector response to the response predicted by using this fitted line is then calculated as a function of the source radiance; this quantity, the 'fractional fall-off' of the detector characteristic, is a measure of the non-linearity of the response.

In order to give an analytic representation of the fractional fall-off, a quadratic function of the normalised source radiance

Equation 21 (1K) eq 2.21

is fitted to the fractional fall-off data. The radiance is normalised to that at 320 K for the purposes of determining the coefficients. Table 2.8 gives the numerical values of the coefficients for the 11 and 12 micron channels for AATSR.

Table 2.8 Coefficients for the calculation of fractional fall-off for AATSR
Coefficient 11 µm 12 µm
coefficient 1 (1K) TBD TBD
coefficient 2 (1K) TBD TBD
coefficient 3.gif (1K) TBD TBD

It is now a simple matter to apply the non-linearity correction to the look-up table; given the true radiance calculated for the tabular value T, Llambda (1K)(T), the fractional fall-off can be calculated using equation eq. 2.21 above. The corrected radiance for the same tabular temperature is then given by the product

Equation 22 (1K) eq 2.22

[Note to be added on non-linearity of 3.7 µm channel.]

2.6.1.1.2.1.2 Algorithm Description

The calculation of the calibration parameters for the infra-red channels involves the calculation of the slope and offset for each channel separately, from the unpacked and validated data in the processed instrument source packets. It averages the necessary values over the number of packets in a calibration period. A calibration period comprises a block of consecutive source packets, each of which corresponds to an instrument scan; the number of source packets that make up a calibration period is defined by a parameter in the auxiliary Processor Configuration File 6.5.6. . The nominal value is 10 scans.

In the event that a change to the 'autocal' loop or the Pixel Selection Map is detected within the calibration period, the calibration period is terminated and the derived coefficients are based on the data preceding the change.

The following five steps are applied to all packets in a calibration period, provided that the user's option for calibration while other instruments are operational is compatible with the state of the blanking pulse flags for selected pixels (see note below).

Step 1. Derive mean BB temperatures.

The weighted mean PRT values are calculated for both hot and cold black bodies, from all valid BB PRT values from one packet. This is repeated for all packets in the calibration period. The component means are summed, and a single mean is derived for each BB over the calibration period.

The mean background (fore-optics) temperature is also derived from valid data over the same period.

Step 2. Convert the mean BB temperatures to radiance.

The weighted mean PRT temperatures for each BB are converted to a radiance value for each IR channel. This is achieved by means of a look up table converting temperature to radiance, with one look up table for each IR channel. The look-up tables are found in the auxiliary file of General Calibration Data 6.5.4. . They are sampled to allow linear interpolation between points.

Step 3. Correct the BB radiance values.

Each black body radiance value is now corrected, using the mean background temperature derived in step 1, and the emissivity constants for each IR channel, according to equation eq. 2.18 and equation eq. 2.22 . (Note that the non-linearity corrections are included in the look-up tables.)

Step 4. Derive the mean BB pixel count for each BB over the calibration period.

These values are calculated for odd and even pixels, for both hot and cold black bodies, and for all channels, from all valid BB pixels in the calibration period. The mean (odd and even) pixel counts for each channel is derived from the valid BB pixel data in each packet. This is repeated for all packets in the calibration period, and a mean BB pixel count over the calibration period is derived, for odd and even pixels, for each BB, and for each IR channel. For a pixel to count as valid in this step, its associated blanking pulse flags must be compatible with the selected option as described in the note below.

Step 5. Calculate the gains and offsets from BB counts and BB radiance.

The gains and offsets for the calibration period, for odd and even pixels, for each black body, and for each IR channel, are derived using equation eq. 2.16 and equation eq. 2.17 .

Note: Treatment of blanking pulse flags.

Two blanking pulse (BP) flags are associated with each pixel in the AATSR source packet. Each flag is associated with one of the radar instruments aboard ENVISAT (RA-2 and ASAR), and is set if the corresponding radar was transmitting at the time the pixel was measured. The purpose of these flags is to guard against the possibility of cross-coupling or interference between the radar transmitters and the AATSR signal channel processing electronics; the flags permit pixels to be excluded from the calibration if one or other radar is transmitting, in case there is any interference from the radar.

Whether or not pixels are included in the calibration depends upon a parameter, the blanking pulse calibration flag (pulse_cal) in the Level 1B Processor Configuration File (ATS_PC1_AX). Pixels are always included in the calibration if both BP flags are off. Otherwise, pixels are included or not according to the setting of this parameter as follows.

Case 101 (pulse_cal = flag_off_bthbp): Both BP flags off: Pixel is included in the calibration only if both flags are off.

Case 102 (pulse_cal = flag_off_asar): ASAR only flag off: Pixel is included in the calibration only if the ASAR flag is off. (The RA flag is ignored).

Case 103 (pulse_cal = flag_off_onlyra): RA only flag off: Pixel is included in the calibration only if the RA flag is off. (The ASAR flag is ignored).

Case 104 (pulse_cal = flag_on_bthbp): Both BP flags on: Pixel is included irrespective of the flag setting. (Either or both flags may be on.) This is the nominal setting.

Thus in Step 4 above, only pixels for which the associated blanking pulse flags are be compatible with the selected option are treated as valid and included in the averages.

2.6.1.1.2.1.3 Accuracies

TBD

2.6.1.1.3 Visible Channel Calibration

This module unpacks the VISCAL data once per orbit (if present), when it is detected that the VISCAL unit is in sunlight. (The VISCAL data is used as a reference for the calculation of the calibration parameters for the visible channels.) In addition, it calculates calibration parameters for all visible channels. These calibration parameters are not used to calibrate the science data in the visible channels for the current orbit, but are written to the Visible Calibration Coefficients ADS in the GBTR product.

Determination of the calibration parameters for the AATSR visible channels involves two main stages. The offset value for each channel is derived by averaging the black body pixel counts in the channel over a calibration period in a similar way to the derivation of the mean black body counts for the infra-red channels during the infra-red channel calibration. The slope value for the channel is derived from the on-board visible calibration system as described below.

Two on-board sources are used to calibrate the AATSR visible/near infra-red channels. The upper reflectance measurement is provided by the visible calibration unit, VISCAL, which gives a signal corresponding to ~ 15% spectral albedo at full solar illumination. The reflectance_factor for each channel are obtained from the pre-launch calibration of the VISCAL.

The zero reflectance signal is derived from one of the on-board black bodies, used also for the thermal calibration. The data from these sources are used to derive the calibration slopes for each channel using (in the case that the MXBB (Minus X Black Body) is used as the reference black body).

Calibration_Slope[chan] = Reflectance_Factor[chan]*Gain[chan] /
(Average_VISCAL_Pixel_Counts[chan]- Average_MXBB_Pixel_Counts[chan])

where the VISCAL and MXBB pixel counts are averaged only during the period when the VISCAL is at full solar illumination. The pixel counts are normalised to unit detector gain to allow for gain changes during an orbit. The Calibration_Slope is then used to convert pixel counts to top-of-atmosphere reflectance by

Pixel_Reflectance[chan] = Calibration_Slope[chan]*
(Pixel_Count[chan] - Average_MXBB_Pixel_Counts[chan])/ Gain[chan]

where the MXBB pixel counts are taken from the same scan as the pixel counts being calibrated. This scheme is modified in the case of the 1.6 µm channel to correct for measured detector nonlinearity.

More detail on the visible calibration will be added in a future issue.

2.6.1.1.4 Satellite Time Calibration

For each scan, the Satellite Binary Time (taken from the source packet) is converted to UTC. The conversion uses a linear relationship between SBT and UTC, defined by parameters taken from the SBT to UTC Conversion Information section of the Main Product Header (MPH) of the product model.

2.6.1.1.5 Geolocation

This heading covers a number of different steps as follows.

  • Instrument pixel geolocation: The latitude and longitude co-ordinates of the instrument pixels are determined. This is an intermediate step the output of which is used by the next stage.
  • Mapping to a cartesian image grid: Image co-ordinates are determined for each instrument pixel. These co-ordinates are used in the regridding stage. The tie point x and y co-ordinates appear in the output product, in the Scan Pixel x and y ADS 6.6.44. .
  • Image pixel geolocation: The latitude and longitude co-ordinates of the image pixels are determined. This process is distinct from the instrument pixel geolocation; the co-ordinates of the image pixels appear in the product in the Geolocation (Grid Pixel Latitude and Longitude) ADS 6.6.43. .

The following parameters are also derived as part of Geolocation:

  • Solar and viewing angles;
  • Topographic correction.

These topics are covered in the following sections.