Am. Ignatov et Il. Dergileva, MODEL APPROXIMATION OF ANGULAR-DEPENDENCE IN AVHRR BRIGHTNESS TEMPERATURES OVER A BLACK SURFACE, International journal of remote sensing, 16(18), 1995, pp. 3687-3693
Brightness temperatures t(i)(m) in spectral channels i = 3 (3.7 mu m),
4 (10.8 mu m), and 5 (12 mu m) of the Advanced Very High Resolution R
adiometer (AVHRR) and the Along Track Scanning Radiometer (ATSR) were
calculated for a variety of viewing directions 1 less than or equal to
m less than or equal to 3 (m = sec theta; theta is the satellite zeni
th angle), and for a representative set of surface-atmosphere conditio
ns (-2 less than or equal to t(s) less than or equal to 35 degrees C,
0.2 less than or equal to W less than or equal to 5.4 g cm(-2); where
t(s) and W are the surface temperature and the column water vapour con
tent). The assumption that the surface is a blackbody is used in order
to analyse the atmospheric effect in detail, and to provide a backgro
und for further incorporation of the surface emissivity. t(i)(m) are a
pproximated with a polynomial t(i)(m) = a(0)(i) + a(1)(i)m + ... + a(j
)(i)m(i)(j = 1, 2, and 3, for linear, quadratic, and cubic fits, respe
ctively). Linear fit provides standard errors of sigma less than or eq
ual to 0.07 and sigma less than or equal to 0.4-0.5 degrees C at 3.7 a
nd 10.8-12 mu m, respectively. Using a quadratic fit reduces these err
ors to sigma less than or equal to 0.01 and sigma less than or equal t
o 0.04-0.06 degrees C. The information content of the fit parameters i
s analysed in this Letter. The value of a(0) is within a few degrees o
f t(s). The values of sigma, (a(0)-t(s)), a(j) strongly correlate with
the factor W Delta t = W(t(s) - (t) over bar(a)), where (t) over bar(
a) is the effective temperature of the atmosphere. These results are u
seful for better understanding recent empirical studies of the angular
effect in AVHRR brightness temperatures, and for developing algorithm
s for t(s) and W retrieval using spectral methods for the cross-scanni
ng AVHRR, and angular methods for the ATSR.