391 
USE OF ERS-1 WINDSCATTEROMETER FOR SURFACE 
CHARACTERISTICS STUDY ON SEMI-ARID AREAS. 
R. D. MAGAGI and Y. H. KERR 
LERTS, 18 AV. EDOUARD BEUN 
31055 TOULOUSE 
ABSTRACT 
This study was performed within the Hapex-Sahel (Hydrologic Atmospheric Pilot Experiment in the Sahel) 
framework. One of Hapex Sahel objectives is to improve the parameterisation of surface hydrology over semi- 
arid areas. To assess surface parameters (soil moisture, surface roughness, biomass) we have studied ERS-1 
windscatterometer (WSC) and NOAA/AVHRR data. The ERS-1 WSC (5.4Ghz) operates in VV polarization. It 
has 3 antennas (fore, mid, aft) which observe the target with incidence angles ranging from 18 to 59°. Its spatial 
resolution is of 50Km. It was primarily designed to assess wind characteristics (speed and direction) over ocean 
through its antennas which measure the backscatiering coefficient (a 0 ) with different viewing angles. In this 
study we will present the opportunity to use the multiangular observations of WSC to quantify soil and 
vegetation characteristics. In order to separate the different contributions of the signal and to reduce its angular 
dependence, it is useful to optimize the WSC parameters (incidence angles). We will adapt for each acquisition, 
the linear regression method between 3 measured a 0 and 3 correspondent incidence angles. The slope and the 
intercept of this function will be analysed in a first step to assess roughness and biomass. The ERS-1 WSC data 
will be coupled with NOAA/AVHRR (visible and near infrared) data. A first result of this synergism will be 
presented. Temporal evolution of angular signatures of WSC data was correlated to surface features. But surface 
roughness effects seemed difficult to retrieve with such method consequently, we invert an empirical model to 
extract soil moisture and surface roughness whose effects will be studied. 
KEYWORDS: Semi-arid, backscattering coefficient, multiangular, roughness, soil moisture, biomass, 
inversion. 
1. INTRODUCTION 
The arid and semi-arid areas have been studied in thermal infrared [1], visible and near infrared [2], microwaves 
[3]. These efforts took into account of the particularity of vegetation repartition on these areas. In this study we 
will be interested in the Sahelian zone of West Africa (Niger) where the Hapex Sahel experiment [4 to 6] took 
place in 1992. One objective of this experiment is to quantify water balance in semi-arid areas.This requires the 
knowledge of surface characteristics which variations influence the radiative balance. Many studies showed the 
ability of aircraft scatterometer or ground scatterometer to identify soil features: soil moisture, roughness, 
biomass. These parameters have been correlated to measured backscattering coefficient [7 to 11]. 
Considering WSC data, we will work at large scale (the spatial resolution of WSC is of 50Km), with a 
frequency of 5.4 Ghz [12]. This gives us a large spatial and temporal (atmospheric effects are negligible at this 
frequency) coverage, and thus an interest in temporal evolution of surface parameters. 
The goal of this study is to verify if the relationships between a 0 and surface parameters were valid for 
satellite data, and to show how wind scatterometer data could provide useful informations on arid and semi-arid 
areas. Over the study site (13°3r08“N-02°39’37”E) the vegetation is strongly dependent on soil moisture 
which is linked to rainfall. The latter alters the soil roughness. We will present consequently a method to 
uncouple these factors. At the end, inversion method to retrieve soil moisture and roughness is presented.The 
results will be discussed.