399 
STUDY OF TERRESTRIAL ECOSYSTEMS AT A GLOBAL SCALE 
WITH ERS-1 WIND SCATTEROMETER DATA 
Mougin £., Frison P.L., Lopes A, Proisy C M Lo Seen D. 
Centre d'Etude Spatiale des Rayonnements / CNRS / UPS 
9 avenue du Colonel Roche 31029 Toulouse Cedex / France 
fax : (33) 61.55.67.01. email : mougin@cnesta.span.cnes.fr 
ABSTRACT 
This paper presents some observations made with the ERS-1 wind scatterometer data over land surfaces. Global 
backscattering coefficient images of the whole African continent are shown, indicating a strong correlation 
between the radar echo and the distribution of the main vegetation types. In the northern hemisphere, arid and 
semi-arid vegetation formations (savannas, woodland,...) are clearly visible. Furthermore, Guinean forests and 
the tropical dense forests of the Congo basin are well identified. Multitemporal variations of the backscattering 
coefficient a° are also illustrated for different vegetation types along a North-South latitudinal gradient. 
Backscattering coefficient data are then compared with NOAA/AVHRR Global Vegetation Index (GVI) data. 
Results indicate the strong capability of the ERS-1 scatterometer to discriminate and monitor vegetation 
evolution at a global scale. 
Keywords : ERS-1, wind-scatterometer, terrestrial ecosystems, vegetation, global scale, GVI. 
INTRODUCTION 
Spacebome scatterometers are primarily designed for estimating wind speed and wind direction over sea 
surfaces. The ERS-1 scatterometer provides a measure of the radar backscattering coefficient o° along different 
azimuthal directions, from which wind parameters can be retrieved. In addition, the ERS-1 wind-scatterometer 
is characterized by a low spatial resolution (about 47 km), a high temporal repetitivity (the same point can be 
seen every 4 days), and above all, a frequency (5.3 GHz, W polarisation) sensitive to land surface parameters 
such as vegetation cover and soil surface moisture content (Ulaby et al, 1986). The radar signal is delivered and 
analysed by three sideways looking antennae; one pointing normal to the satellite flight path (midbeam 
antenna) and the two others pointing 45° forward (forebeam antenna) and 45° backward (afibeam antenna) 
respectively. The illumination swath is equal to 500 km and across this swath, local incidence angles are 
ranging from about 20° to 55° depending on the antenna considered. Consequently, a point on the earth can be 
seen, at different times, by various viewing conditions both in azimuth and incidence angles. 
As already pointed out by a few previous studies using Seasat scatterometer data (Birrer et al, 1982; Kennett 
and Li, 1989a; 1989b) acquired at a higher frequency (Ku-band), ERS-1 wind-scatterometer characteristics 
make this instrument "a priori" suitable for land applications. However, due to the inherent low spatial 
resolution, vegetation studies must necessarily be limited to regional and global applications. At present, two 
years of fully calibrated data (Lecomte and Attema, 1992) acquired by the ERS-1 scatterometer over the whole 
land surfaces offer a unique opportunity to assess the contribution of low resolution active microwave systems 
to global monitoring of the terrestrial surfaces. 
The objective of this paper is to present some observations obtained with the ERS-1 scatterometer data over 
land surfaces. Emphasis is put on the African continent which presents numerous vegetation formations along a 
North-South latitudinal gradient. In the first part, we describe the method to elaborate global backscattering 
coefficient images of the whole continent. In the second part, examples of spatial and temporal ct° signatures of 
different vegetation types are given. Finally, backscattering coefficient images are compared with global maps 
derived from the NOAA/AVHRR Global Vegetation Index (GVI) series. Their differences and complementarity 
are pointed out.