Environmental Engineering Reference
In-Depth Information
Fig. 8.7
Daily averaged global mean total bias correction (Kelvin) for MSU channel 2 onboard
satellites NOAA-6 to NOAA-14. The time series from 1979 to 1988 in four different colors (with
larger seasonal variability) are for NOAA-6 through NOAA-9, and the smoother time series from
1987 to 2007 in other four different colors are for NOAA-10 through NOAA-14 (Plot from Saha
et al.
2010
)
(CFSR) and NASA Modern Era Retrospective analysis for Research and
Applications (MERRA) reanalysis (Saha et al.
2010
; Rienecker et al.
2011
).
Since the CFSR and MERRA reanalysis cover the entire period of MSU
observations from 1978 to 2006, they actually assimilated two different MSU
Level-1c data: the NOAA operational calibrated for TIROS-N through NOAA-9
and the postlaunch SNO calibrated for NOAA-10 through NOAA-14. The fact that
radiances from these different calibration procedures were assimilated into the
same system serves as an ideal experiment for evaluating the impact and perfor-
mance of the SNO inter-satellite calibration effort. Figure
8.7
shows the total bias
correction patterns of MSU channel 2 in the CFSR for the eight satellites from
NOAA-6 through NOAA-14. The total bias correction is a global mean difference
between the satellite observations and a background short-term forecast field. This
quantity is an indicator of how well the observations agree with the model. These
biases can also be used to determine the comparability of two different satellites,
since differencing of the bias corrections of two satellites gives inter-satellite
difference information similar
to the inter-satellite bias analysis shown in
Figs.
8.3
and
8.6
.
As seen in Fig.
8.7
, the bias corrections for NOAA-6 through NOAA-9 exhibit
obvious seasonal and interannual variability. As discussed earlier, this variability is
related to the solar heating-induced instrument temperature variations. In addition,
the bias correction values for NOAA-6 through NOAA-9 are different for different
satellites, indicating larger relative offsets between these satellites. This occurs
because each individual satellite was calibrated independently in prelaunch
calibrations. Similar bias correction patterns were also found for NOAA-10 through
NOAA-14 when operational calibrated radiances for these satellites were assimilated
into the ERA-Interim reanalysis (Dee and Uppala
2009
). In contrast, no instrument
temperature variability is observed in bias corrections of the SNO-calibrated MSU
observations for NOAA-10 through NOAA-14. Moreover, the bias correction values
for NOAA-10 through NOAA-14 are nearly the same, resulting in more consistent
Search WWH ::
Custom Search