47 
IN FIELD CHARACTERIZATION OF SAHARIAN SITES REFLECTANCE 
PROPERTIES FOR THE CALIBRATION OF OPTICAL SATELLITE SENSORS 
H. Cosnefroy(l), X. Briottet(l), M. Leroy(2), P. Lecomte(3), and R. Santer(4) 
(1) ONERA CERT/DERO, 2, av. E. Belin. 31055 Toulouse, France 
(2) LERTS, UMR CNES/CNRS, 18, av. E. Belin, 31055 Toulouse, France 
(3) LOA, Université de Lille, 59655 Villeneuve d'Ascq, France 
(4) LOA, Université du Littoral, 62123 Wimereux, France 
ABSTRACT 
During a joint in-field campaign conducted in February- March 1993, we have characterized 4 Saharian 
desertic sites in Algeria identified from satellite imagery to be suitable for calibration purposes. Bidirectional 
properties of the surfaces in reflectance and polarization were collected in different plans as well as 
measurements to describe the spatial and temporal variabilities of each site in 4 spectral bands in the visible, 
near infrared and shortwave infra red. The Saharian in-field experiment is part of a more global effort, which 
aims at the characterization of some selected test sites in North Africa and Saudi Arabia, with ground truth, 
airborne experiments and satellite data, in order to prepare metrological standards for the calibration of present 
(AVHRR, Meteosat, SPOT, etc) and future (POLDER, Vegetation, MER1S, etc) satellite sensors. 
1 - INTRODUCTION 
Stable desertic sites are promising candidate terrestrial targets for the evaluation of the calibration of 
optical satellite sensors, not only for multitemporal calibration purposes (Holben et al., 1990; Kaufman et al., 
1990; Staylor, 1990; Henry et al., 1993), but also for the multiangular calibration (Delphin et al., 1991; 
Martinuzzi et al., 1992; Cosnefroy et al., 1993) of wide field of view sensors equipped with CCD detectors, 
such as POLDER/ADEOS (Deschamps et al., 1994), VEGETATION/SPOT4 (Amaud and Leroy, 1991), 
MISR/EOS (Diner et al., 1989), or MERIS/ENV1SAT (Rast et al., 1991). A third objective, not yet explored, 
could be the assessment of multiband calibration of optical sensors. 
To investigate the potential of desert sites for the calibration of space instruments, we have selected 20 
desertic zones in Saharian North Africa and Saoudi Arabia, using a criterion of spatial uniformity in a 
multitemporal serie of cloud-free METEOSAT-4 visible images, and we have verified that the temporal 
stability of these sites at hourly and seasonal time scales has a satisfactory level (Cosnefroy et al., 1993). A key 
next step is the metrological characterization of these calibration sources, that is, the evaluation of the spatially 
averaged bidirectional spectral reflectance seen at the top of the atmosphere of each of the selected zones. This 
reflectance may be denoted by p 7 . OA (A.,9 i ,0 v ,<t>), that is a function of wavelength X, Sun and view zenith 
angles 0 5 and 0 V , and relative azimuth 4> between Sun and view directions. Satellite measurements, necessarily 
limited in spectral and directional space, can not be the only source of metrological measurements; they must 
be supplemented by in field and airborne campaigns, for which the spatial, spectral and directional 
measurement coverage can be performed with more flexibility. 
This paper reports some results obtained in a field campaign in Saharian south Algeria, from February 
20 to March 21, 1993, in which measurements of bidirectional reflectance have been collected in 4 zones 
preselected by Cosnefroy et al. (1993). The objectives of the campaign have been, for each zone: (i) to 
characterize their surface spectral and directional reflectance, in order to serve as an input for the evaluation of 
the TO A (Top Of the Atmosphere) spatially averaged reflectance mentioned above; (ii) to characterize the 
spatial and temporal variabilities of these reflectances. Investigated length scales of variability range from tens 
of meters to a few kilometers, so that local measurements of reflectance can be compared with satellite 
measurements such as those of METEOSAT or AVHRR. Investigated time scales are from a few hours to a few 
days and have been limited essentially by logistic considerations; (iii) to apprehend, zone by zone, the major 
causes of spatial and temporal variabilities (aerosols, clouds, surface variability, or humidity); (iv) to 
qualitatively document the surface of each zone (type of ground, vegetation, structure and orientation of dunar 
patterns).