Land Surface Temperature (LST) Retrieval from GOES Satellite Observations - Satellite-based Applications on Climate Change - page 292

Namely, the surface emissivity and atmospheric total column water vapor cause
the brightness temperature change and therefore, cause LST retrieval error.

19.3.2 Mathematical Description of the LST Algorithm

An analytic solution to Eq. 19.1 is not easy, because the integration of the terms
requires good knowledge of the atmospheric profiles which is not available in real
time. In addition, land surface emissivity is coupled with the surface temperature in
the equation, so the number of unknowns is always larger than the number of
equations and this is the so-called ill-posed problem, even multiple channels of
information are available. In the past 35 years, many approaches, by using the two
split-window (SW) channels (11.0 and 12.0

m) to the solution, have been
suggested (e.g., McMillin 1975 ; Walton et al. 1998 ) and widely used for producing
the LST product (e.g., Prata 1993 , 1994 ; Wan 1999 ; Caselles et al. 1997 ), as listed
in Table 19.2 .
Most split-window algorithms explicitly use land surface emissivity values,
while Sun and Pinker ( 2003 , 2004 ) and Sikorsky et al. (2002) proposed emissivity
information indirectly incorporated through the use of different coefficient sets
determined by different land surface types. The latter approach must be tolerant to
within-class emissivity variability and assume the land cover maps can be updated
frequently; however, most available land cover products can only be available
annually from EOS/MODIS and NOAA/AVHRR or seasonally from NPOESS/
VIIRS, meanwhile emissivity maps that accommodate within-class variability (Yu
et al. 2005 ) can be available more frequently from MODIS product. We expect that
other emissivity map developments will be significantly improved by the launch of
GOES-R in 2014.
Since the operational GOES-LST retrieval will be from current GOES sensors,
including GOES-12 and GOES-13, there will be no 12.0-

μ

m channel. Therefore,
split-window channels cannot be used to correct atmospheric effect. Due to the lack
of split-window channels for the imager of GOES M (12)-Q series, Sun and Pinker
( 2004 ) proposed a dual-window algorithm by using the characteristics of the mid-
infrared channel

μ

m) with less atmospheric (water vapor) absorption
(Fig. 19.1 ), and one-channel (11

(3.9 μ

μ m) plus water vapor correction algorithm.

19.3.2.1 GOES Split-Window Algorithm

For far-IR bands, terms R d , R s , and R r in Eq. 19.1 are negligible. Therefore, only the
first two terms on the right side are important. The first term represents the surface
contribution term, and it is the gray-body radiance emitted by the Earth's surface.
The second term is the atmospheric contribution term, referred to as path thermal
radiance in Eq. 19.5 , and is the vertically integrated effect of emission from every