The possibility to derive the cloud base temperature by remote sensing
for convective cloudfields over
oceans with the spaceborne infrared radiometer AVHRR is investigated.
First, a set of atmospheric data for polar cold air outbreaks
is prescribed, containing different cloud base
temperatures, cloud droplet size distributions, and other parameters
influencing the radiative transfer properties.
Top of
the atmosphere radiances are then calculated with the help of a one-dimensional
adding and doubling model of radiative transfer in order to simulate the
expected signals in the thermal infrared channels of the AVHRR.
A method to detect pixels totally covered
with semitransparent clouds is presented.
The optical depth of those clouds can be estimated, because part of
the radiance of the warm ocean
can penetrate through the clouds in at least one channel. For such
clouds the base temperature is infered from top
of the atmosphere radiances.
The physical and statistical interpretation of the model results suggests two
different methods to derive cloud base temperatures for day and night
conditions. The night algorithm is based on all three thermal infrared
channels and yields a root mean square error (RMS)
of 2 K. For daytime scenes RMS increases to 4 K, because only
two channels can be used.
Differences of similar magnitude occur, if these methods are applied
to real satellite data and are compared with radiosoundings.
However, averaging over several satellite pixels is necessary.
Finally, the effect of cloud base temperature on the radiation budget of the
ocean surface is estimated. Results of radiative transfer calculations show an
increase of longwave downward flux densities by up to 90 W/m^2 for the
cloud covered sky as compared to cloudless conditions. This considerable
increase is due to the low water vapour amount in polar boundary layers. The
sensitivity of the longwave net flux density to cloud base temperature
is about 2 W/m^2 per Kelvin.
21. August 1996, Andreas Manschke