659 
Characterization of Different Canopies from POLDER Polarization 
Measurements during La Crau 1991 Campaign. 
J.L. Deuzé*, R. Santer, P. Lecomte* 
Laboratoire d'Optique Atmosphérique (L.O.A) 
* Université des Sciences et Techniques de Lille 
59655 Villeneuve d'Ascq, France 
Université du Littoral, Station Marine de Wimereux 
62910 Wimereux, France 
be made apparent from space observations. In order to emphasize this contribution corresponding to a 
Fresnel reflection, we will work (i) around the specular direction, (ii) at 650 nm over vegetation where it 
is believed that a significant contribution of polarized light can be reflected by the leave surfaces. To 
observe this phenomena we used an aircraft version of the POLDER instrument, for details on this 
instrument the reader may reefer to Deschamps et al (1990) or to companion papers published in these 
proceedings and using the POLDER instrument. One objective of POLDER is to use the polarized 
reflectance measured over land for aerosol remote sensing and this ability is described in Deuze et al 
(1993) during an experiment held in La Crau (France) in June 1990. Two favorable circumstances helped 
to emphasize the aerosol contribution: first the day was quite turbid with an aerosol optical thickness of 
0.4 at 550 nm and second La Crau is a flat plain of white pebbles which is quite lambertian and then 
generated a little polarization. During an other campaign in June 1991, POLDER flew over La Crau and 
an agricultural site under clear circumstances. At the same time, we had ground-based measurements of 
the outgoing reflectances and degrees of polarization over different types of canopies as well as 
atmospheric measurements to describe the aerosols and the water vapor content. Two objectives behind 
these measurements to link them to the POLDER images: atmospheric correction (and we will define a 
simple approximation of the signal to solve this problem), a validation of this correction by comparing 
corrected POLDER data and ground based measurements. This last point also raises the problem of the 
spatial scale between the 20 meter pixel size of POLDER and the 50 cm spot on the ground for the 
photopolarimeter. We will include this study by comparing different methods of agricultural inventories 
provided by the experiment: NDVI, polarized reflectance and BRDF. 
For this mission, POLDER was equipped at three wavelengths (550 nm, 650 nm and 850 nm) 
with a set of 3 polarizors to give the total radiance, the polarized reflectance and the direction of 
polarization. The objectives of this campaign were multiple with absolute calibration activities on the 
calibration test site of La Crau involving SPOT-2, AVIRIS and AVHRR and on the other hand with 
agricultural inventories based on spectral measurements (AVIRIS) and on bi-directional reflectances and 
degrees of polarization. POLDER images were acquired over La Crau for north-south and east-west legs 
on June 18th and 19th. The agricultural test site was over flown on June 21th during three legs. 
Figure 1 is a POLDER image of the site at 850 nm on which we labeled sites we identified from 
the ground. Along track , a same target is present on 12 images; each of them separated by 10 degrees in 
view angle. The same fields were investigated by Leroy et al (this issue) focusing on the use of the 
BRDF derived from POLDER and INRA did on the site biological and physical measurements. 
1-INTRODUCTION 
In the solar spectrum, the light reflected by the Earth-atmosphere system is partially polarized. 
The polarization originates both from the atmosphere (molecules and aerosols) and from the ground. The 
general purpose of this paper is to investigate if a polarized component, originated from the surface, can 
2- PRESENTATION OF THE MEASUREMENTS 
2.1 The POLDER measurements