Calibrating a System of Satellite Instruments - Satellite-based Applications on Climate Change - page 26

2.5 Future Developments
It is a significant challenge to calibrate the operational satellite instruments for climate
change detection. As a result, further developments in many areas are needed. First, in
order to establish consistency through inter-satellite calibration, uncertainties in the
SNO method should be further reduced with SNO site characterization using highly
accurate spectral, spatial, BRDF (bidirectional reflectance distribution function), and
elevation models. This will be especially helpful for the window or surface channels.
The reduced uncertainty will allow us to better quantify the inter-satellite biases and
small trends in the satellite measurements. Long-term observation of vicarious sites
such as the Dome C, Libyan Desert, and other sites with stable instruments will
provide independent site stability and calibration accuracy assessments. International
collaboration under the WMO/GSICS and CEOS facilitate data sharing and allow us
to intercalibrate radiometers globally to establish a calibration reference network and a
quasi on-orbit standard.
Further improvements in onboard calibrators, that is, blackbody in the infrared
and microwave, and multiple solar diffusers in the visible/near-infrared will reduce
calibration uncertainties and facilitate the establishment of on-orbit calibration
standard. For instruments with onboard calibration, not only the biases between
satellites but also the root cause of the biases should be investigated. This is because
bias correction without knowing the root cause can be unreliable. Once the root
cause is identified, this information can be used as feedback to the instrument
development process to improve the calibration of future models.
For the microwave instruments, further improvement in the prelaunch nonline-
arity and sidelobe characterization is highly desirable. Knowledge of nonlinearity is
critical for decadal climate change detection. Further improvements in instrument
noise would significantly reduce the uncertainties in the SNO analysis, since sample
size is relatively small at the SNOs for these instruments. Also, the possibility of
long-term nonlinearity change, spectral response, and frequency drift should be
investigated.
For infrared instruments, intercalibrations between sounders and imagers on the
same satellite are very valuable. Since both types of instrument are on the same
spacecraft and simultaneous Earth views are available globally, accurate intercali-
bration, both radiometrically and spectrally, is possible (Tobin et al. 2006 ; Wang
and Cao 2008 ). Recent analysis of MetOp data show that the HIRS has a more
stable calibration than that of AVHRR when both are compared to IASI on MetOp
IASI. This will help us use HIRS to calibrate AVHRR historical data since both are
on the same spacecraft. As more and more hyperspectral sounders become avail-
able, this type of inter-sensor calibration will become more valuable.
Uncertainties in the spectral response functions (SRF) have been a major issue
for optical instruments, including the prelaunch measurement uncertainties, SRF
differences between instruments, and SRF changes over time. The lack of stringent
requirement in the prelaunch testing, the inability to make identical SRFs, and the
lack of on-orbit spectral calibration devices are the root causes of these problems. It