Atmospheric Temperature Climate Data Records from Satellite Microwave Sounders - Satellite-based Applications on Climate Change - page 116

spacecraft, and the Earth. The correction for this effect, the so-called antenna
pattern correction, results in brightness temperature differences of about 0.5-2 K.
Mo ( 1999 ) has developed a set of antenna pattern correction coefficients for the
AMSU-A instrument. For a specific channel, this correction adds a same constant
for a satellite. Thus, it does not affect the inter-satellite calibration results.
However, similar to the calibration offset, it influences the absolute values of the
brightness temperature and bias corrections in the reanalysis data assimilation.
This correction is currently set to be optional for implementation in the NOAA
MSU/AMSU FCDR and TCDR products, depending on needs for brightness
temperature validations against other observations.
8.3.2 Limb Adjustment
A limb correction adjusts different incident angles of the off-nadir footprints to the
nadir direction. This correction is necessary for use of the off-nadir footprints in the
time series to increase observation samples and reduce noise- and sampling-related
biases. Limb-correction algorithms and coefficients have been developed by
Goldberg et al. ( 2001 ) for both the MSU and AMSU satellites using statistical
methods. Zou et al. ( 2009 ) examined the impact of the limb correction on the MSU
time series and found robust trend results when different limb-corrected footprints
were included in the time series. Currently, the Goldberg et al. ( 2001 ) limb-
correction coefficients are used as part of the NOAA MSU/AMSU TCDR
processing system.
8.3.3 Diurnal Drift Correction
The diurnal drift errors are caused by satellite orbital drifts, which are measured by
the slow evolution in the local equator crossing time (LECT) as a sun-synchronous
satellite ages. The satellite orbital drifts result in a change of local observation time
that, if not corrected, may introduce a false climate trend by bringing the diurnal
trend into it. Its effect is particularly large for the land areas of the lower- and mid-
tropospheric temperature channels (e.g., MSU channel 2 and AMSU-A channel 5)
where diurnal amplitude is large. It is also important for stratospheric temperature
channels.
Two correction methods were developed by previous investigators: Christy
et al. ( 2000 ) corrected the diurnal drift effect using diurnal anomalies estimated
by accumulating the local MSU or AMSU-A observations from different scan
positions at different local times, and Mears et al. ( 2003 ) used diurnal anomaly
climatology generated from NCAR Community Climate Model (CCM) for the
correction. The two methods have caused larger trend differences in the MSU
observations (Mears and Wentz 2005 ). Currently, the diurnal anomalies developed