794 
2.1 Temperature measurements 
Thermal camera used for 
temperature measurement 
Ground brightness temperature 
measurements were performed with 
a thermal camera (INFRAMETRICS 
760) equipped with 3 spectral bands 
(8-14 pm, 9.8-11.4 pm, 9.6-13 pm), 
(Figure 1) and calibrated with 17 
different blackbodies. Experimental 
tests performed in laboratory and in 
field conditions show that if it is 
flowered up more than about 75 
minutes (Figure 2) before the 
measurement, the measurement error 
with this camera is less than 0.3 °C. 
Test site: Crau (South-East of 
France) 
The experiment was conducted on 
the Crau test site situated in the 
south-east France, a very flat pebbly 
area. It has been widely used for 
absolute calibration of remote 
sensing visible and near-infrared 
sensors (Gu et al., 1991; Santer et 
al„ 1992). It has been also widely 
used for energy balance 
measurement (Seguin et al., 1982). 
The ground surface has a very high 
thermal emissivity (about 0.96) 
(Labed and Stoll., 1991) and the 
spectral difference of emissivities for 
the two AVHRR thermal channels 
can be neglected (Z.L. Li, 1993, 
personal communication). 
Sampling 
In order to obtain the brightness 
temperature of the site at the time of 
the satellite overpass, we have 
performed the measurements with an 
helicopter at 100 meters above the 
ground. For reducing the effect of 
emissivity and temporal change of 
the temperature of the test site on the 
calculation of the average brightness 
surface temperature, we have used 
only the measurements performed in 
the 9.8-11.4 pm band, and collected 
5 minutes before/after the satellite 
overpass. 
Figure 1. Spectral response of thermal camera INFRAMETRICS and 
spectral variation of emissivity of vegetation and of bare soil. 
o 
211 40 60 »0 I (HI UH 
Time after camera powered up (in minute) 
Figure 2. Temporal variation of radiometric response of the camera 
Emissivity