Land Surface Temperature (LST) Retrieval from GOES Satellite Observations - Satellite-based Applications on Climate Change - page 280

Center (NCEP) mesoscale Eta model is coupled with Oregon State University
(OSU) land surface scheme and with the simple biosphere (SIB) scheme. The
inclusion of detailed land surface schemes into the general circulation models
(GCMs) and numerical weather prediction (NWP) models is motivated by the
realization that the surface plays an important role, in regulating the exchange of
heat, momentum, and energy between the Earth's surface and the atmosphere. Skin
temperature is a key parameter for land surface process parameterization. At
present, information on surface skin temperature is available only from few field
experiments, such as the First International Satellite Land Surface Climatology
Project (ISLSCP) Field Experiment (FIFE) (Sellers et al. 1992), Boreal Ecosystem-
Atmosphere Study (BOREAS) Experiment (Sellers et al. 1995, 1997), the Atmo-
spheric Radiation Measurement Program (ARM) Experiment ( http://www.arm.
gov ) , the MONSOON experiment (Kustas and Goodrich 1994), the Oklahoma
Mesonet Network ( http://okmesonet.ocs.ou.edu ) , the CASES experiment ( http://
www.mmm.ucar.edu/cases ) , and many more shorter field observations. Till now,
surface shelter temperature was used as proxy to skin surface temperature, even
though; these are known to differ. Observations from satellites have been proven to
be useful for inferring surface skin temperature. However, not all satellites have the
necessary capabilities to derive surface temperature at high accuracy; some do not
have sufficient number of channels to derive surface emissivity, while others do not
observe the Earth's surface frequently enough to represent the diurnal cycle.
Deriving accurate land surface temperature (LST) from satellites is both attrac-
tive and challenging. It is attractive because LST is a highly variable quantity in
both space and time. Satellites provide efficient and practical means of capturing
this variability. It is challenging because the land surface is very heterogeneous,
LST is generally not homogeneous within one pixel, and land surface emissivities
may be quite different from unity and spectrally variable (Lyon 1965; Nerry et al.
1990).
There are three main sources of error in the determination of LST from satellites.
First, the satellite instruments have sensor noise and calibration errors that transfer
into errors in brightness temperature. Next, algorithms may have errors in the
determination of the atmospheric effect, and lack of knowledge of spectral
emissivities of the land surface. The third source of error is the evaluation process
itself. Ground observations are limited in scope and measurements are taken
at a point, while satellites measure a pixel average. For example, for GOES, it is
a4km

4 km area, and it is well known that land surface temperature is not
homogeneous on such scale.
While surface temperature retrievals from satellites utilize atmospheric windows
(where absorption is minimum), the influence of atmospheric absorption and
emission is not negligible. Water vapor is the major absorbing gas in the window
channels; it varies with season and latitude.
The effect of surface emissivity is twofold. Since the land surface emissivity is
generally less than one, part of the atmospheric downward radiation is reflected by
the surface and has to be accounted for (Lorenz 1986). The emitted radiation by the
surface is modified in each channel, yielding different values for the brightness
temperature.