NOAA/AVHRR Vegetation Indices : Suitability for Monitoring Fractional Vegetation 
Cover of the Terrestrial Biosphere 
C. Leprieur (1), M. M. Verstraete(2), B. Pinty(3) & A. Chehbouni(4) 
(1) LERTS, UMC0001O/CNES/CNRS, 18 av. Edouard Belin, F-31055 Toulouse Cedex, FRANCE. 
(2) IRSA, CEC Joint Research Center, 1-21020 Ispra (VA) ITALY 
(3) LAMP, URA 267/CNRS Université Biaise Pascal, F-63177 Aubière, FRANCE. 
(4) ORSTOM Niamey, (NIGER), présent affiliation : JPL,Mail Stop 300-233,4800 Oak Grave Drive, 
Pasadena, CA 91109, USA 
ABSTRACT: 
Various economic activities are dependent on the growth and development of vegetation, particularly 
agriculture and forestry. The state and evolution of the vegetation cover therefore present a great interest and 
must be repeatedly monitored over large areas but the quantitative interpretation of the satellite observations 
may be difficult A variety of vegetation indices have been used to assess the state and monitor the evolution 
of the terrestrial biosphere. Early indices were easy to compute but very sensitive to soil and atmospheric 
effects. Modified indices with a reduced sensitivity to soil brightness changes were proposed such as the 
Weighted Distance Vegetation Index (WDVI), Soil Adjusted Vegetation Index (SAVI) and the Modified Soil 
Adjusted Vegettion Index (MSAVI). More recently, new indices have been designed to be less affected by 
both atmospheric and soil conditions such as the Global Environment Monitoring Index (GEMI) or the 
Atmospherically Resistant Vegetation Index (ARVI). This paper evaluates a representative sample of these 
indices with respect to their capability to retrieve the fractional vegetation cover from AVHRR optical data. 
Flexible criterions to assess the performance of indices are proposed. These criterions are provided by the 
ratio of signal to noise at the surface level and at the top of the atmosphere. This approach has been applied to 
simulated data for various ranges of fractional cover, soil types and atmospheric conditions. These criterions 
will help quantify how the various indices transport the surface information at the top of the atmosphere 
MOTS CLES: fractional vegetation cover, terrestrial biosphere, vegetation indices 
1- INTRODUCTION 
The state of the terrestrial vegetation cover is directly affected by climatic events, as well as by human 
activities. The biosphere, however, also influences the climate of this planet through its partial control of the 
exchange of water, energy, carbon and other biogeochemicals between the surface and the atmosphere. 
Furthermore, various economic activities are dependant on the growth and development of vegetation, 
particularly agriculture and forestry. The state and evolution of the vegetation cover therefore present a great 
interest and must be repeatedly monitored over large areas. The international scientific community has 
organized a number of large scale coordinated efforts to address the issue of climate change and environmental 
degradation. These activities known as the International Geosphere Biosphere Program (IGBP) or the Global 
Change Research Program (GCRP) have been described in the littérature (e.g., NRC, 1986, 1988; IGBP, 1988, 
1990; Houghton et al, 1990; Jäger and Ferguson, 1991). 
Artificial satellites provide the only convenient way to repetitively perform the monitoring activity over 
extended areas required by these programmes, but specific tools must be developed to convert the radiation 
measurements made in space into useful information relative to the biophysical parameters of the canopy. 
Ideally, the analysis of satellite data should be made with the help of detailed physically-based models capable of 
representing the interactions of radiation with the surface and the atmosphere. Such models are under 
development but are currently lacking. In the mean time, empirical approaches have been developed to address 
some of these issues and provide preliminary answers. This paper focus on comparing a representative sample of 
of vegetation indices that are commonly used to exploit AVHRR data,for the specific purpose of estimating the 
fractional vegetation cover. 
1103