nents (e.g. veg-
er that depends
: work has been
structured scene
1980a; Kimes et
ordered crops or
strical structure
obtained over a
iy effects but in
ir variations on
s, Australia.
isurements cou-
»meter (ATSR).
found in Prata
; the same point
idir and at 55°.
s may be found
pose of deriving
¡wing ATSR 1.6
nighttime data,
lar flux and air
L data has been
( 12 )
( 13 )
is an angle-dependent parameter and,
Ae - e x (0i) - e x (d 2 ). (14)
7 a is a mean downward atmospheric radiance, and Ix(Oi) and /*($ 2 ) are radiances measured at
the two zenith angles 6\ and 6 2 . These angles vary across the swath of the ATSR. The last term
in ( 12 ) is generally small because the emissivity is usually close to unity and the atmospheric
radiance is small. This radiance algorithm can be used for comparison with temperatures by
inverting the Planck function at wavelength A to obtain T,. Since the ATSR has channels at 11
fim and 12 /xm, these can be used in ( 12 ) to obtain two estimates of the surface temperature.
Equation (12) is quite general and it is equally applicable for use over the sea surface.
The set of validation data consist of 31 “clear-sky” coincidences of ATSR 11 /xm and 12
jim nadir and forward temperatures collocated with areal averages of the in situ temperatures.
Figure 5 illustrates some of the features of these data.
Surface temperature (K)
11-12 pm temperature difference (K)
Figure 5: Aspects of ATSR 11 /xm and 12 /xm nadir and forward temperatures, (a)
Angular temperature difference (nadir -forward) as a fuction of the zenith angle dif
ference (forward-nadir), (b) Temperature difference's a function of in situ surface
temperature, (c) Angular temperature difference as a function of acquisition time
(local time), (d) Variation of the split-window temperature difference with the 11
/xm angular temperature difference.
The variation of the difference between the nadir and forward temperatures as a function
°f the angular difference between the forward zenith angle and the nadir angle is shown in Fig.
