Atmospheric Temperature Climate Data Records from Satellite Microwave Sounders - Satellite-based Applications on Climate Change - page 109

Table 8.1 Channel and scanning view numbers for each MSU and AMSU-A antenna systems

AMSU-A
A1-1

AMSU-A
A1-2

AMSU-A
A2 Instrument antenna systems MSU-1 MSU-2

Channels 1-2 3-4 6-7, 9-15 3-5, 8 1-2

Earth views per scan line 11 11 30 30 30

Blackbody and space views
per scan line

1 1 2 2 2

PRTs in each warm target 2 2 5 5 7
Note that the MSU has two antenna systems: 1 and 2. AMSU-A has two units, A1 and A2, where
A1 has two antenna systems, A1-1 and A1-2, and A2 has its own antenna system

8.2.2 Solar Heating-Related Instrument Temperature Variability

A major calibration error that needs to be removed from the radiance FCDR is the
instrument temperature variability induced by solar heating differences on the instru-
ment, which originates from seasonal solar angle changes relative to the satellite orbit
normal over a year and its yearly differences due to satellite orbital drifts. As an
example, Fig. 8.2 shows the warm target temperature time series for NOAA-10
through NOAA-15, which contains representative examples for both the MSU
(NOAA-10 through NOAA-14) and AMSU-A (NOAA-15) instruments. It is seen
that these warm target temperatures, which are at the ambient instrument tempera-
ture, all incurred large seasonal and interannual variability due to solar heating
variations of the instruments.
The instrument temperature variability as shown in Fig. 8.2 causes variations of
the spacecraft thermal emission, which are received as part of the signals in the
antenna side-lobe views of the Earth and the calibration targets. For instruments
having a perfectly linear radiometer transfer function, the instrument thermal
emission signals can be mostly removed by the linear calibration equation. In
reality, however, most instruments are slightly nonlinear. Inaccurate expressions
of the nonlinearities cause residual instrument temperature variability to manifest in
the radiance time series. Figure 8.3 shows the global ocean-mean inter-satellite
brightness temperature difference time series for MSU channel 2 onboard TIROS-N
through NOAA-14 based on the NOAA operational calibrated radiances. The time
series was generated by averaging seven near-nadir, limb-corrected radiances in
each scan line over the global ocean for each satellite. In the global difference time
series, dominate climate signals are canceled out by nearly identical observations
from two overlapping satellites so that only instrument calibration errors and
observation sampling errors such as those related to diurnal drift remain. Fortu-
nately, diurnal drift errors can be ignored in global ocean means for MSU channel
2 observations (Mears et al. 2003 ; Zou et al. 2009 ); thus, any errors in the inter-
satellite difference time series can be traced to the instrument calibration errors. As
shown in Fig. 8.3 , NOAA operational calibration results in inter-satellite biases that
are a few tenths of a degree (Kelvin), and these biases change with time with their
variability highly correlated to the instrument warm target temperature variations