First, the GOES radiometers have a lower calibration accuracy than their polar
counterparts. For example, the current GOES imager calibration accuracy is 1 K,
compared to 0.5 K for AVHRR. The large diurnal variation in the GOES instrument
temperature, on the order of nearly 30 K, presents a challenge in intercalibration with
POES radiometers, which has an orbital temperature variation of 2–3 K. Second,
although the GOES nadir has a fixed location on the Earth, it does not necessarily
observe the nadir at the time of POES overpass because it takes ~30 min for GOES to
perform a complete scan of the Earth. As a result, the simultaneity between POES and
GOES is typically around 15 min, compared to 30 s in the SNO method. The fast
scanning capability on future GOES-R/ABI will significantly improve the simultaneity which will reduce the uncertainties.
Earth targets such as the Libyan Desert, Antarctica, and deep convective clouds
have been used for vicarious calibration of visible/near-infrared channels. However, this strategy for inter-satellite calibration is affected by observation time
differences and uncertainties introduced by bidirectional reflectance factors of the
surface and the intervening atmosphere. Vicarious calibration can achieve intersensor calibration with ~2% accuracy, but differences of 4–10% are not uncommon
(Green and Pavri 2002; Thome, 2006, personal communication). An extensive
inter-satellite calibration campaign using the Dome C site has been coordinated
under the CEOS/Working Group on Calibration/Validation (Cao et al. 2010) using
this approach.
It is possible to use the Moon for intercalibration of radiometers on different
satellites, as demonstrated in studies with MODIS, SeaWiFS and other instruments
(Barnes et al. 2004; Xiong et al. 2005). There are significant advantages with this
approach. The Moon has a stable reflectance (Kieffer 1997), and inter-satellite
calibration using the Moon is not affected by observation gaps between satellites if
the Moon is used as an absolute calibration standard. Consistency at the 1% level has
been demonstrated for intercalibration of the MODIS on Aqua and Terra using the
Moon. However, instrument design can impose a potential limitation to using the
Moon for calibration. For example, the AVHRR space clamp circuitry makes the lunar
calibration approach difficult (Cao et al. 2009). Spacecraft maneuvers to view the
Moon are possible for some missions but may be impractical for other missions.
Currently, the Moon is only used for the calibration of the visible/near-infrared
channels. There are significant challenges to using the Moon for calibrating infrared
instruments, and the feasibility of lunar calibration for microwave instruments has yet
to be investigated.
Satellite mission overlap is essential to most intercalibration techniques. This
requires not only time overlap and consistency in local observation time but also
assurance of spectral continuity in channel selection between satellites. For example, channel discontinuities were created when the HIRS channel 10 center
wavenumbers were changed over the course of the NOAA satellite series. The
small frequency change from MSU channel 2 to AMSU channel 5 also created
problems in climate change detection (Iacovazzi and Cao 2007). Such changes
should be avoided, if possible, in mission requirements.
2 Calibrating a System of Satellite Instruments
25
counterparts. For example, the current GOES imager calibration accuracy is 1 K,
compared to 0.5 K for AVHRR. The large diurnal variation in the GOES instrument
temperature, on the order of nearly 30 K, presents a challenge in intercalibration with
POES radiometers, which has an orbital temperature variation of 2–3 K. Second,
although the GOES nadir has a fixed location on the Earth, it does not necessarily
observe the nadir at the time of POES overpass because it takes ~30 min for GOES to
perform a complete scan of the Earth. As a result, the simultaneity between POES and
GOES is typically around 15 min, compared to 30 s in the SNO method. The fast
scanning capability on future GOES-R/ABI will significantly improve the simultaneity which will reduce the uncertainties.
Earth targets such as the Libyan Desert, Antarctica, and deep convective clouds
have been used for vicarious calibration of visible/near-infrared channels. However, this strategy for inter-satellite calibration is affected by observation time
differences and uncertainties introduced by bidirectional reflectance factors of the
surface and the intervening atmosphere. Vicarious calibration can achieve intersensor calibration with ~2% accuracy, but differences of 4–10% are not uncommon
(Green and Pavri 2002; Thome, 2006, personal communication). An extensive
inter-satellite calibration campaign using the Dome C site has been coordinated
under the CEOS/Working Group on Calibration/Validation (Cao et al. 2010) using
this approach.
It is possible to use the Moon for intercalibration of radiometers on different
satellites, as demonstrated in studies with MODIS, SeaWiFS and other instruments
(Barnes et al. 2004; Xiong et al. 2005). There are significant advantages with this
approach. The Moon has a stable reflectance (Kieffer 1997), and inter-satellite
calibration using the Moon is not affected by observation gaps between satellites if
the Moon is used as an absolute calibration standard. Consistency at the 1% level has
been demonstrated for intercalibration of the MODIS on Aqua and Terra using the
Moon. However, instrument design can impose a potential limitation to using the
Moon for calibration. For example, the AVHRR space clamp circuitry makes the lunar
calibration approach difficult (Cao et al. 2009). Spacecraft maneuvers to view the
Moon are possible for some missions but may be impractical for other missions.
Currently, the Moon is only used for the calibration of the visible/near-infrared
channels. There are significant challenges to using the Moon for calibrating infrared
instruments, and the feasibility of lunar calibration for microwave instruments has yet
to be investigated.
Satellite mission overlap is essential to most intercalibration techniques. This
requires not only time overlap and consistency in local observation time but also
assurance of spectral continuity in channel selection between satellites. For example, channel discontinuities were created when the HIRS channel 10 center
wavenumbers were changed over the course of the NOAA satellite series. The
small frequency change from MSU channel 2 to AMSU channel 5 also created
problems in climate change detection (Iacovazzi and Cao 2007). Such changes
should be avoided, if possible, in mission requirements.
2 Calibrating a System of Satellite Instruments
25