SATELLITE MEASUREMENTS
181
4.2.3
The GHRSST data processing model
The GHRSST-PP project has approached the problems identified in the
previous section by adopting a particular model of the thermal structure of
the near surface of the ocean and with it defining clear terminology that is
intended to avoid confusion and misunderstanding in the use of the loose
term “SST” which can have several meanings. A distinction is made
between skin and sub-skin SST as outlined in 4.2.2, and a new term, the
“foundation SST” or SST fnd is used to describe the temperature on which the
diurnal warming (if any) is built each day. It is most clearly specified as the
temperature of the well mixed layer found just below the skin layer at dawn,
when any diurnal thermocline structure from the previous day has collapsed.
At this time of day it is equivalent to the subskin SST. The SST fnd is defined
on a daily basis. It also corresponds closely to the temperature of the upper
layer of the ocean most commonly represented in numerical ocean
circulation models that do not attempt to represent diurnal warming. The
use of this name is preferred instead of the term “bulk” SST. The latter is
very imprecise because it tends to be used also for any in situ measurement
made by a thermometer in contact with the water (as distinct from a
radiometric measurement which observes the skin SST) which may or may
not be influenced by the diurnal thermocline, depending on the depth of the
thermometer and the character of a particular day’s diurnal variability.
A key aspect of the GHRSST processing model is that satellite
measurements are recognised to be either skin or subskin temperatures,
while it is assumed that the temperature normally required by models is the
SST fnd . The relationships between each of the different SST quantities are
parameterised in terms of the controlling air-sea interaction quantities,
particularly the wind speed and the solar irradiance, and these are expected
to vary during the day. Note that the difference between skin and subskin
(the thermal skin deviation), and between subskin and foundation (the
diurnal warming) are different and controlled by largely independent
processes. It is envisaged that when SST values are needed to high accuracy
for ingestion into a numerical ocean model, adjustments will be applied to
convert the satellite measurement of skin or subskin SST into the appropriate
value of SST fnd . However, there are circumstances where a model may need
to use the skin temperature, for example when estimating air-sea fluxes, and
in this case the GHRSST-PP processing model specifies how conversion can
be made from SST fnd to SST skin .
The approach outlined in the previous two paragraphs provides a
framework in which problems (b) and (c) from 4.2.2 are handled. It is also
intended that errors caused by (b) and (c) should be isolated from the more
general biases and errors of (a). The latter need to be provided along with
the basic temperature data to allow the model assimilation system to weigh
the importance to be attached to a particular SST ingest. Therefore
181
4.2.3
The GHRSST data processing model
The GHRSST-PP project has approached the problems identified in the
previous section by adopting a particular model of the thermal structure of
the near surface of the ocean and with it defining clear terminology that is
intended to avoid confusion and misunderstanding in the use of the loose
term “SST” which can have several meanings. A distinction is made
between skin and sub-skin SST as outlined in 4.2.2, and a new term, the
“foundation SST” or SST fnd is used to describe the temperature on which the
diurnal warming (if any) is built each day. It is most clearly specified as the
temperature of the well mixed layer found just below the skin layer at dawn,
when any diurnal thermocline structure from the previous day has collapsed.
At this time of day it is equivalent to the subskin SST. The SST fnd is defined
on a daily basis. It also corresponds closely to the temperature of the upper
layer of the ocean most commonly represented in numerical ocean
circulation models that do not attempt to represent diurnal warming. The
use of this name is preferred instead of the term “bulk” SST. The latter is
very imprecise because it tends to be used also for any in situ measurement
made by a thermometer in contact with the water (as distinct from a
radiometric measurement which observes the skin SST) which may or may
not be influenced by the diurnal thermocline, depending on the depth of the
thermometer and the character of a particular day’s diurnal variability.
A key aspect of the GHRSST processing model is that satellite
measurements are recognised to be either skin or subskin temperatures,
while it is assumed that the temperature normally required by models is the
SST fnd . The relationships between each of the different SST quantities are
parameterised in terms of the controlling air-sea interaction quantities,
particularly the wind speed and the solar irradiance, and these are expected
to vary during the day. Note that the difference between skin and subskin
(the thermal skin deviation), and between subskin and foundation (the
diurnal warming) are different and controlled by largely independent
processes. It is envisaged that when SST values are needed to high accuracy
for ingestion into a numerical ocean model, adjustments will be applied to
convert the satellite measurement of skin or subskin SST into the appropriate
value of SST fnd . However, there are circumstances where a model may need
to use the skin temperature, for example when estimating air-sea fluxes, and
in this case the GHRSST-PP processing model specifies how conversion can
be made from SST fnd to SST skin .
The approach outlined in the previous two paragraphs provides a
framework in which problems (b) and (c) from 4.2.2 are handled. It is also
intended that errors caused by (b) and (c) should be isolated from the more
general biases and errors of (a). The latter need to be provided along with
the basic temperature data to allow the model assimilation system to weigh
the importance to be attached to a particular SST ingest. Therefore