Satellite Remote Sensing of Evaporation over Sea
443
Before going into operational application each remote sensing method has to be assessed by
independent means of comparison. For sea surface temperatures this is not an easy task since
temperatures as measured in situ by the voluntary observing fleet are often inaccurate and
they are restricted to bulk temperatures measured at depths of decimetres to metres whereas
the temperature measured by infrared radiometers is representative of the upper 10 to 20 micrometres, i. e. the skin of the ocean. Ground measurements of the skin temperatures are
only taken from research vessels or platforms during dedicated field campaigns. Therefore, it
is almost impossible to get a statistically significant number of comparisons between satellite
and surface skin measurements (Schliissel et al., 1987). Hence, for a global validation of the
retrieval schemes quality controlled data from drifting buoys are usually employed (Emery et
al., 1994; Mutlow et al., 1994). Those validation studies confirm that the ATSR delivers superior accuracy when compared to the AVHRR retrievals and that the algorithms including
additional water-vapour information from SSMjI or using quadratic terms for the water-vapour
corrections show a better correlation with in-situ measurements.
A further augmentation of all retrieval schemes is necessary if volcanic aerosol reduces the atmospheric transparency as indicated in table 16.3. The additional correction that is needed can
be performed by altering the retrieval coefficients just after a volcanic outbreak or by including
additional information about the enhanced atmospheric opacity as retrieved from spectral measurements in the near-infrared at 1.6 jtm. The latter is part of the ATSR measurements and
will be utilized in future for correcting effects imposed by volcanic aerosol in the stratosphere
but also by tropospheric aerosol particles.
16.2.2 The cool skin and the diurnal thermocline of the ocean
The sea surface temperature as measured with infrared sensors stems from the upper 10 to 20
jtm of the ocean while the "surface temperature" defined by oceanographers is that measured
by ships or buoys at depths of several decimetres or metres. The difference between bulk and
skin temperature is determined by the thermal molecular sublayer of the ocean and, during
daytime, also by the diurnal thermocline (figure 16.3). Usually, the surface is cooled by sensible, and latent heat fluxes as well as by longwave radiation fluxes causing a lower temperature
at the surface when compared to the temperature at depths of millimetres. The solar radiation
which is not acting directly at the surface penetrates to greater depths where it is absorbed
depending on the actual turbidity of the ocean water. Consequently, it affects both the skin
layer and the mixed layer below. Normally, the skin cooling is greater than the solar heating
and the bulk versus skin temperature difference remains positive across the molecular sublayer.
However, the formation of a diurnal thermocline during calm situations can alter the sign of
the integrated effect. Under clear sky situations with a dry atmosphere aloft both effects can
reach their extremes: During night the radiative cooling as well as the latent heat flux lead to a
strong lowering of the surface skin temperature while during daytime the solar radiation is well
transmitted by the atmosphere. Hence, the variablitity of the bulk versus skin temperature,
t::.T, is maximized in situations where infrared remote sensing of sea surface temperature is
possible. These situations can usually be found in the subtropical oceans and in the middle
latitudes. The temperature differences caused by these two effects can range from -1 K to 1 K
in the open ocean. Even larger values occur in coastal waters and in lakes. Developments of
the parameterisation of t::.T in terms of energy and momentum fluxes through the sea surface
443
Before going into operational application each remote sensing method has to be assessed by
independent means of comparison. For sea surface temperatures this is not an easy task since
temperatures as measured in situ by the voluntary observing fleet are often inaccurate and
they are restricted to bulk temperatures measured at depths of decimetres to metres whereas
the temperature measured by infrared radiometers is representative of the upper 10 to 20 micrometres, i. e. the skin of the ocean. Ground measurements of the skin temperatures are
only taken from research vessels or platforms during dedicated field campaigns. Therefore, it
is almost impossible to get a statistically significant number of comparisons between satellite
and surface skin measurements (Schliissel et al., 1987). Hence, for a global validation of the
retrieval schemes quality controlled data from drifting buoys are usually employed (Emery et
al., 1994; Mutlow et al., 1994). Those validation studies confirm that the ATSR delivers superior accuracy when compared to the AVHRR retrievals and that the algorithms including
additional water-vapour information from SSMjI or using quadratic terms for the water-vapour
corrections show a better correlation with in-situ measurements.
A further augmentation of all retrieval schemes is necessary if volcanic aerosol reduces the atmospheric transparency as indicated in table 16.3. The additional correction that is needed can
be performed by altering the retrieval coefficients just after a volcanic outbreak or by including
additional information about the enhanced atmospheric opacity as retrieved from spectral measurements in the near-infrared at 1.6 jtm. The latter is part of the ATSR measurements and
will be utilized in future for correcting effects imposed by volcanic aerosol in the stratosphere
but also by tropospheric aerosol particles.
16.2.2 The cool skin and the diurnal thermocline of the ocean
The sea surface temperature as measured with infrared sensors stems from the upper 10 to 20
jtm of the ocean while the "surface temperature" defined by oceanographers is that measured
by ships or buoys at depths of several decimetres or metres. The difference between bulk and
skin temperature is determined by the thermal molecular sublayer of the ocean and, during
daytime, also by the diurnal thermocline (figure 16.3). Usually, the surface is cooled by sensible, and latent heat fluxes as well as by longwave radiation fluxes causing a lower temperature
at the surface when compared to the temperature at depths of millimetres. The solar radiation
which is not acting directly at the surface penetrates to greater depths where it is absorbed
depending on the actual turbidity of the ocean water. Consequently, it affects both the skin
layer and the mixed layer below. Normally, the skin cooling is greater than the solar heating
and the bulk versus skin temperature difference remains positive across the molecular sublayer.
However, the formation of a diurnal thermocline during calm situations can alter the sign of
the integrated effect. Under clear sky situations with a dry atmosphere aloft both effects can
reach their extremes: During night the radiative cooling as well as the latent heat flux lead to a
strong lowering of the surface skin temperature while during daytime the solar radiation is well
transmitted by the atmosphere. Hence, the variablitity of the bulk versus skin temperature,
t::.T, is maximized in situations where infrared remote sensing of sea surface temperature is
possible. These situations can usually be found in the subtropical oceans and in the middle
latitudes. The temperature differences caused by these two effects can range from -1 K to 1 K
in the open ocean. Even larger values occur in coastal waters and in lakes. Developments of
the parameterisation of t::.T in terms of energy and momentum fluxes through the sea surface