770 
results axe shown in Figure 9. Also shown are the temperatures obtained from the in situ 
devices (solid circles). The limb darkening effect is noticeable as is the overall reduction in 
temperature sensed by TIMS, due principally to water vapour absorption. Most of the limb 
Zenith angle (degrees) 
Figure 9: Temperatures derived from the TIMS channel 
6 (11.7 fim) and from the in situ measurements as a 
function of viewing angle for the image shown in Fig. 8 . 
darkening effect is due to emissivity variations with angle. At the scale of a TIMS pixel it is 
likely that the radiometer is viewing completely bare ground toward the western side of the 
field site. Whereas for the ATSR, (whose pixels would fill the whole of the images shown in 
Figures 6 and 7), the angular emissivity effect is small for this field site, for the TIMS the 
emissivity effect is dominant. LOWTRAN-7 calculations have been performed using data from 
a radiosonde launched at the site at around 1000LT. At 30° from the zenith, the reduction 
in temperature is less than 1°C different to that for vertical viewing (see also Figs. 1 and 
2). By using all six channels of the TIMS it is noticeable that the baxe ground patches show 
considerable temperature variation with wavelength, whereas the vegetated axeas show little 
vaxiation. Also, the temperatures of the bare patches are generally lower than the surrounding 
vegetation, indicating a lower emissivity. 
4. DISCUSSION 
The angular effects on the radiances detected by satellite-borne and airborne radiometers have 
been investigated. Some parametrisations have been developed and used with data from the 
ATSR. The comparisons of ATSR dual angle estimates with spatially compatible in situ tem 
peratures shows that accuracies of around 1.2—1.4°C in LST are achievable. It has been argued 
that the angular variation of the emissivity is important for LST determination. The impact o 
this vaxiation on the LST estimates obtained here from the ATSR is not great probably because 
the surface studied has some vegetative cover. However, using high resolution TIMS data the 
emissivity angular effects are very evident. 
A good estimate of the nadir emissivity is vital for obtaining accurate LSTs. Because ol 
lack of ancillary data, constant values have been assumed for some of the important param