238
P.J. Minnett
(O’Carroll et al., 2008). The large numbers of the drifting buoy data set render it too
valuable a resource for SST validation for it to be neglected, but the quality control
of the buoy data sets has to be thorough.
Moored buoys in the tropical Pacific and Atlantic Oceans, and in US waters are
also used as a source of validating data. They generally have a thermometer at a
depth of about 3 m, and while these are recovered for post-deployment recalibration (Freitag et al., 1999) the disturbance to the upper ocean by the presence of the
mooring introduces a source of uncertainty into the comparison that is difficult to
quantify.
The distribution of collocated (within 0.1 ◦ of latitude and longitude) and coincident (within 30 min) measurements from the Aqua MODIS and drifting and moored
buoys is shown in Fig. 14.5 for the year 2003 (Minnett and Barton, 2010). There are
12,536 match-ups in the plot and only those satellite retrievals with the best quality
flag (i.e. confidently cloud-free and satellite zenith angle <45 ◦ ) are shown, but even
with this large number there are ocean areas that are poorly sampled.
14.4.3 Traceability to Temperature Standards
Given that CDRs of SST span several satellite missions, ensuring that the validating
measurements are themselves accurate over the CDR period is of prime importance.
Without this assurance, systematic changes in the characteristics of the data sets
used to validate satellite SST could be misinterpreted as systematic changes in the
upper ocean and the climate. The only way of ensuring stability in the calibration of
the sensors used to provide validation data is to have a traceable calibration chain
to a national SI temperature standard. Since buoys are seldom recovered for recalibration, the key to the generation of an SST CDRs lies in the calibration of the
ship-based radiometers. The path to national temperature standards, therefore, is
through the calibration of the radiometers used to validate the satellite retrievals,
and this requires, and provides, radiometric traceability to national standards. In the
USA, the national reference standards are maintained by the National Institute of
Standards and Technology (NIST).
As part of the pre-launch characterization of the satellite radiometers, they are
carefully calibrated in thermal-vacuum chambers to replicate the conditions on
orbit. The pre-launch calibration is traceable to national standards, but the satellite radiometers are not recovered at the end of the mission for recalibration and
re-characterization. To ensure traceability to NIST standards, an infrared calibration
facility has been set up at the Rosenstiel School of Marine and Atmospheric Science
at the University of Miami. Three international workshops have been held at which
many of the ship-board radiometers used to validate satellite-derived SSTs were
calibrated using a water-bath blackbody calibration target, built to a NIST design
(Fowler, 1995). This calibration target consists of a black-painted, thin-walled, hollow, tapered, copper cone surrounded by a water bath, the temperature of which
can be very accurately controlled. The water bath temperature is monitored by two
thermometers, which have calibrations traceable to NIST standards. The internal
P.J. Minnett
(O’Carroll et al., 2008). The large numbers of the drifting buoy data set render it too
valuable a resource for SST validation for it to be neglected, but the quality control
of the buoy data sets has to be thorough.
Moored buoys in the tropical Pacific and Atlantic Oceans, and in US waters are
also used as a source of validating data. They generally have a thermometer at a
depth of about 3 m, and while these are recovered for post-deployment recalibration (Freitag et al., 1999) the disturbance to the upper ocean by the presence of the
mooring introduces a source of uncertainty into the comparison that is difficult to
quantify.
The distribution of collocated (within 0.1 ◦ of latitude and longitude) and coincident (within 30 min) measurements from the Aqua MODIS and drifting and moored
buoys is shown in Fig. 14.5 for the year 2003 (Minnett and Barton, 2010). There are
12,536 match-ups in the plot and only those satellite retrievals with the best quality
flag (i.e. confidently cloud-free and satellite zenith angle <45 ◦ ) are shown, but even
with this large number there are ocean areas that are poorly sampled.
14.4.3 Traceability to Temperature Standards
Given that CDRs of SST span several satellite missions, ensuring that the validating
measurements are themselves accurate over the CDR period is of prime importance.
Without this assurance, systematic changes in the characteristics of the data sets
used to validate satellite SST could be misinterpreted as systematic changes in the
upper ocean and the climate. The only way of ensuring stability in the calibration of
the sensors used to provide validation data is to have a traceable calibration chain
to a national SI temperature standard. Since buoys are seldom recovered for recalibration, the key to the generation of an SST CDRs lies in the calibration of the
ship-based radiometers. The path to national temperature standards, therefore, is
through the calibration of the radiometers used to validate the satellite retrievals,
and this requires, and provides, radiometric traceability to national standards. In the
USA, the national reference standards are maintained by the National Institute of
Standards and Technology (NIST).
As part of the pre-launch characterization of the satellite radiometers, they are
carefully calibrated in thermal-vacuum chambers to replicate the conditions on
orbit. The pre-launch calibration is traceable to national standards, but the satellite radiometers are not recovered at the end of the mission for recalibration and
re-characterization. To ensure traceability to NIST standards, an infrared calibration
facility has been set up at the Rosenstiel School of Marine and Atmospheric Science
at the University of Miami. Three international workshops have been held at which
many of the ship-board radiometers used to validate satellite-derived SSTs were
calibrated using a water-bath blackbody calibration target, built to a NIST design
(Fowler, 1995). This calibration target consists of a black-painted, thin-walled, hollow, tapered, copper cone surrounded by a water bath, the temperature of which
can be very accurately controlled. The water bath temperature is monitored by two
thermometers, which have calibrations traceable to NIST standards. The internal