NOAA KLM User's Guide

Section 7.3

Blue line drawn across page to separate text from document title.
Introduction Page, NOAA KLM TOC, Acronyms
Previous Section, Next Section

7.3 AMSU-A and AMSU-B

The Advanced Microwave Sounding Unit-A (AMSU-A) is afifteen-channel total power microwave radiometer in two separate units: A1 and A2. The AMSU-A2, which has its own antenna system, contains Channels 1 and 2 at 23.8 and 31.4 GHz, respectively. The AMSU-A1, which consists of the two antenna systems A1-1 and A1-2, contains Channels 3-14 in the range of 50.3 - 57.29 GHz and Channel 15 at 89.0 GHz. Between the two antenna systems, A1-2 provides Channels 3, 4, 5, and 8, whereas A1-1 furnishes Channels 6, 7, and 9 through 15. Each of the AMSU-A antenna systems is required to have a nominal field-of-view (FOV) of 3.3 degrees ± 10% at the half-power points and covers a crosstrack scan of ± 48 degrees 20 minutes (to beam centers) from the nadir direction with 30 Earth FOVs per scan line. The data received from the spacecraft will contain separate telemetry for each antenna system. Once every 8 seconds, the AMSU-A measures 30 Earth views, the space view twice and the internal blackbody target twice. Combination of Gunn diode cavity-stabilized and phase-locked loop oscillators (PLLO) are used to provide channel frequency stability. Channels 9 through 14 have both primary and secondary phase-locked loop oscillators (called PLLO#1 and PLLO#2, respectively) built-in. The PLLO#2 will be used for backup if the PLLO#1 fails. Gunn diode oscillators are used in other channels without backups. The output signals of these radiometric samples are digitized by 15-bit analog-to-digital converters.

The AMSU-B is a five-channel total power microwave radiometer with two channels centered nominally at 89 GHz and 150 GHz, and the other three centered around the 183.31 GHz water vapor line with double-sideband centers located at 183.31± 1, ± 3, and ± 7 GHz, respectively. AMSU-B has a FOV of 1.1 degrees ± 10%, and once every 8/3 seconds it measures 90 Earth views, four space views and four internal blackbody target views.

The calibration procedures for the AMSU-A and AMSU-B are the same, except a few minor differences to allow for the separate antenna systems. Table 7.3-1 lists the main differences between the AMSU-A and AMSU-B procedures. Radiances for both AMSU-A and -B Earth views are derived from the measured counts and the calibration coefficients inferred from the internal blackbody and space view data. Both AMSU-A and AMSU-B were tested and calibrated in T/V chambers before launch and the scannings of both instruments are synchronized to an 8-second pulse. Each instrument has four options for the viewing direction of the space view which can be selected by ground command, and one will be chosen immediately after launch. Note: The Calibration Parameters Input Data Sets (CPIDS) are included in the Header Records of Level 1b data according to the Header Record Format.

Two bias issues have been identified on NOAA-K, where the S band transmitters interfere with the AMSU-B instrument. These corrections must be applied to the instrument counts before calibration. See Appendix M for details on the NOAA-K AMSU-B correction algorithm.

Table 7.3-1. Differences between the AMSU-A and AMSU-Bprocedures.
Items AMSU-A AMSU-B Remarks
A1-1 A1-2 A2
Number of PRTs in each warm target 5 5 7 7 Internal black body targets
Number of Earth views per scan line 30 30 30 90 In-orbit
Blackbody and space samples 2 2 2 4 Per scan line
Definition of instrument temperature RF Shelf A1-1 RF Shelf A1-2 RF Shelf A2 Mixer temp. of Ch. 18-20 Available in housekeeping
Backup of instrument temperature RF Mux A1-1 RF Mux A1-2 RF Mux A2 Mixer temp of Ch. 16 (See Note 1.)
Secondary PLLO Ch. 9-14 NO NO NO Backup PLLO
1. These temperatures (available in the housekeeping data) will be used as backups if the primary ones fail.

7.3.1 Blackbody Temperature

The physical temperatures of the internal blackbody targets are measured by Platinum Resistance Thermometers (PRTs). The number of PRTs used to measure the physical temperatures of the internal blackbody targets in each antenna system is given in Table 7.3-1. These PRTs, which were calibrated by individual manufacturers against 'standard' ones traceable to NIST, measure temperatures of the internal blackbody targets with an accuracy of ± 0.1K. The outputs to the telemetry are PRT counts, which must be converted to PRT temperatures. The normal approach for deriving the PRT temperatures from counts is a two-step process, in which the resistance of each PRT (in ohms) is computed by a count-to-resistance look-up table provided by its manufacturer. Then, the individual PRT temperature (in degrees) is obtained from an analytic PRT equation. However, this can be compressed to a single step with negligible errors.

This single step process, which will be used with the NOAA KLM satellites, computes the PRT temperatures directly from the PRT counts, using a polynomial of the form

T sub k = sum from j=0 to 3 of f sub {kj} {C sub K}^j

where Tk and Ck represent the temperature and count of the PRT. The coefficients, fkj, will be provided for each PRT. Equation 7.3.1-1 is also used for other housekeeping temperature sensors, such as the mixers, the IF amplifiers and the local oscillators.

The mean blackbody temperature, Tw is a weighted average of all PRT temperatures:

T sub w = {sum from k=1 to m of w sub k times T sub k} over {sum from k=1 to m of w sub k} + Delta T sub w

where m represents the number of PRTs for each antenna system as listed in Table 7.3-1. The wk is the weight assigned to each PRT and ΔTw is the warm load correction factor for each channel derived from the T/V test data for three instrument temperatures (low, nominal, and high). Values for ΔTw will be provided for each instrument. For AMSU-A1-1, ΔTw values for both PLLO#1 and PLLO#2 will be provided. The wk value, which equals 1(0) if the PRT is determined good (bad) before launch, will be provided for each flight model. If any of the PRT temperatures, Tk, differs by more than 0.2K from its value in the previous scan line, then the Tk should be omitted from the average in Equation 7.3.1-2.

Similarly, a cold space temperature correction, ΔTc, is also provided for processing the in-orbit data. This is due to the fact that the space view is contaminated by radiation which originates from the spacecraft platform and the Earth's limb. Thus, the effective cold space temperature is given by,

T sub c=2.73 + Delta T sub c

where 2.73K is the cosmic background brightness temperature. The ΔTc, which represents the contribution from the antenna side lobe interference with the Earth limb and spacecraft, is estimated initially for individual channels, but its optimal value will be determined from post-launch data analysis.

7.3.2 Calibration Counts

For each scan, the blackbody counts Cw are the averages of two (four) samples of the internal black body in AMSU-A (AMSU-B). If any two samples differ by more than a preset limit of blackbody count variation ΔCw (the initial limit is set to 3σ, where the standard deviation, σ, is calculated from the pre-launch calibration data Cw for each channel), the data in the scan should not be used.

Similarly, the space counts Cc are the average of two (four) samples of the space view for AMSU-A (AMSU-B). If any two space view samples differ by more than the preset limit, the data in the scan should be excluded.

To reduce noise in the calibrations, the Cx (where x = w or c) for each scan line will be convoluted over several neighboring scan lines according to a weighting function:

mean c sub x={1 over{ n+1}} times {[ sum from i=-n to n of (1- abs{i}/{n+1}) times c sub x (t sub i)]} counts

where ti (when I ≠ 0) is the time of the scan line just before or after the current scan line. If t0 is the time of the current scan line, one can write ti=t0 + iΔt, where Δt=8 seconds for AMSU-A and 8/3 seconds for AMSU-B. The 2n+1 values are equally distributed about the scan line to be calibrated. For both AMSU-A and AMSU-B, the value of n=3 is recommended.

For the first and the last three scan lines in a file, the convolution of Cx should be omitted and the counts Cx from the individual scan line will replace mean of C sub x. In the case of missing scan lines in the 2n+1 interval, any one of the remaining scan lines can be selected to replace the missing one(s) in the convolution of Cx. If the gap of missing scans is larger than 2n+1 (i.e., 7), the convolution process must be terminated at the beginning of the gap and starts anew at the end of the gap.

7.3.3 Earth View Radiances

The following calibration algorithm, which takes into account any nonlinear contribution due to an imperfect square law detector, is employed to convert observed Earth-viewing counts to radiances:

R sub s=R sub w+{C sub s - mean C sub w}over G+Q in mW/m^2-sr-cm^-1

where Rs is the scene radiance and Rw and Rc are the Planck radiances corresponding to the blackbody temperature Tw and the effective cold space temperature Tc defined in Equations 7.3.1-2 and 7.3.1-3, respectively. The Cs is the radiometric count from the scene (Earth) target. The averaged blackbody and space counts, mean of C sub w and mean of C sub c, are defined by Equation 7.3.2-1. The channel gain G and the quantity Q, which contains the quadratic contributions, are given by:

G={mean C sub w - mean C sub C}over {R sub w - R sub c} in counts over {mW over (m^2 - sr - cm^-1)}


Q=u{(C sub s - mean C sub w)(C sub s - mean C sub c)}over G^2 in mW over {m^2-sr-cm^-1}

where u is a predetermined parameter which will be provided at three principal (or backup) instrument temperatures. The u values at other instrument temperatures will be interpolated from these three principal (or backup) values. For channels 9 through 14 (AMSU-A1-1) two sets of the u parameters are provided; one set is for the primary PLLO#1 and the other one for the secondary PLLO#2. The quantity G varies with instrument temperature, which is defined in Table 7.3-1.

For channels 19 and 20 of AMSU-B, the monochromatic assumption breaks down (e.g. channel 20 spans 16 GHz) and a band correction with two coefficients (b and c) has to be applied. These coefficients modify Tw to give an effective temperature T sub w prime

T sub w prime=b+cT sub w

which is then used in the Planck function B(T sub w prime) to calculate the radiances for channels 19 and 20. Radiances for all other channels (in both AMSU-A and -B) are computed from B(Tw). The application of Equation 7.3.3-4 is not necessary for the space temperature since the errors in the monochromatic assumption are negligible for such low radiances. For simplification of application, Equation 7.3.3-1 can be rewritten as,

R sub s=a sub 0 + a sub 1 C sub s + a sub 2{ C sub s}^2 in mW over {m^2-sr-cm^-1}

The coefficients ai (where i=0, 1 and 2) can beexpressed in terms of Rw, G, mean of C sub w and  mean of C sub c. This can be accomplished by rewriting the right-hand side of Equation 7.3.3-1 in powers of Cs and equates the ai's to the coefficients of same powers of Cs. The results are:

a sub 0=R sub w - {mean C sub w}over G + u times {mean C sub w times mean C sub c} over G^2 in mW over {m^2-sr-cm^-1}
a sub 1=1 over G - u times {mean C sub c + mean C sub w}over G^2 in mW over {m^2-sr-cm^-1}count


a sub 2=u times {1 over G^2} in mW over {m^2-sr-cm^-1} count^2

These calibration coefficients will be calculated for every scan line at each channel and appended to the Level 1b data. With these coefficients, Equation 7.3.3-5 can be used to obtain the scene radiance Rs. It should be noted that the coefficients defined in Equations 7.3.3-6 to 7.3.3-8 are functions of instrument temperature. Therefore, they are, in general, not constant and should be recalculated for each scan. Users who prefer brightness temperature instead of radiance, can make the simple conversion,

T sub s=B ^{-1} times (R sub s)

where B-1 (Rs) is the inverse of the Planck function for a radiance Rs. The Ts is the corresponding brightness temperature (or radiometric temperature). However, the conversion (Equation 7.3.3-9) is not performed in the NOAA Level 1b data. Note that for AMSU-B channels 19 and 20 the band correction coefficients must also be applied as follows:

T sub s prime= {T sub s - b}over c

where T sub s prime corresponds to the brightness temperature of channel 19 or 20.

7.3.4 Post-launch Calibration and Evaluation

The procedures and formulas described in the above sections are primarily derived from the pre-launch analysis of the AMSU-A and -B calibration data from T/V chambers. It is expected that these procedures are equally valid for processing the in-orbit data. However, both AMSU-A and AMSU-B are new instruments, which have not been flown in space before. In addition, the NOAA-K spacecraft is the first of the KLM series of Advanced TIROS-N spacecraft that differs significantly from the previous ones (such as NOAA-J). Therefore, a comprehensive test and verification of the instrument performance are warranted. NASA and NOAA plan to conduct a post-launch on-orbit verification and evaluation for NOAA-K. This post-launch checkout, which includes all instruments, will last approximately two months. Besides checking the post-launch instrument performance against its specifications, the following parameters from AMSU-A and -B will be monitored during the test period:

The long-term trending data of the above items will establish a valuable database for the new instruments. NOAA will continue to monitor the first three items during the lifetime of NOAA-K, in addition to the time series of channel gains and scan-position readings of both instruments.

Amended April 12, 2001

Previous Section Top of Page Next Section