Land Surface Temperature (LST) Retrieval from GOES Satellite Observations - Satellite-based Applications on Climate Change - page 319

Fig. 19.21 Spectral variation in surface emissivity for different surface types (From MOSART)

vapor along the view path increases with the increase of satellite zenith angle. For the
lower LST cases (LST less than 280 K) with dry atmospheric conditions, the STD and
bias errors are not significantly sensitive to the view zenith angle.
Stratifying our regressions by water vapor regime, we assume that water vapor
content can be well estimated as a priori. In practice, water vapor information is
usually available from satellite soundings, in situ radiosondes and/or operational
numerical weather prediction model forecasts. Nevertheless, two errors may occur.
First, the water vapor value may be mismeasured due to a variety of error sources.
Second, due to spatial resolution differences between the GOES observations and
water vapor data, both “dry” and “moist” atmospheric conditions may occur within
the unit spatial area over which the water vapor was estimated.

19.5.4.3 Emissivity Uncertainty

An emissivity error can cause errors in the simulated brightness temperature and
therefore an LST retrieval error. As shown in Fig. 19.21 , the brightness temperature
error increases with the increase in the emissivity error and is larger in the split-
window channels at 11.0 and 12.0

μ m and smaller in the middle-infrared (MIR)

3.9 m band. The brightness temperature error due to emissivity error in Fig. 19.10

is the average of all global data points over the entire temperature range. For a
specific temperature, the error may be larger. As shown in Fig. 19.21 , emissivity
variations are fairly small in the thermal IR bands (11.0 and 12.0 mm) but
somewhat larger in the MIR band (3.9 mm); however, an emissivity error causes
a smaller brightness temperature error in the MIR band than in the thermal IR bands
(Fig. 19.22 ). This is another reason for us to introduce the dual-window algorithm.
Analytically, the maximum LST uncertainty T s due to the emissivity uncertainty
can be described as

μ

p

δT 1 2

þ δT 2 2

δT s ¼ (19.58)