1202 
• Research teams developing methodologies for an optimal use of satellite data or studying biosphere 
processes and needing remotely sensed data in combination with ground data for local studies. Their need 
can be global data but generally is local data and for periods of time not exceeding 1 or 2 years. 
• Projects based on multisource data, which are operational and need data for long periods of time (more 
than 10 or 15 years) for the purpose of research or operational applications. Among them are the 
international programmes on Global Change or programmes on agriculture monitoring or for a better 
management of ressources. With respect to Global Change studies, many of them are undertaken in the 
framework of the International Geosphere -Biosphere Programme. The relevant projects are Global 
Climate Terrestrial Ecosystems (GCTE), with for instance the SALT experiment studying dry african 
savannas, Biospheric Aspects of Hydrological Cycle (BAHC) with studies of desertification mechanisms, 
Land Use land Cover Change (LUCC) aimed at studying the impact of mankind actions on climate 
(deforestation,etc.).The HAPEX and BOREAS experiments enter in the ffamew'ork of GEWEX. Besides, 
European projects like TREES for the monitoring of tropical forests (TREES, 1992) or MARS (Monitoring 
Agriculture with Remote Sensing) (Meyer Roux et al, 1992) have demonstrated the credibility of 
operational application programmes based on satellite data. Some of these projects are now able to precise 
mission and products requirements for a space mission definition. 
These applications are based on existing systems which were generally designed for another use. It thus appears 
a need for operational satellites specially designed for vegetation monitoring. 
2 - RATIONALE FOR VEGETATION MONITORING SYSTEMS 
Many applications described are based on a long period monitoring. They consequently need a continuity in the 
end-products for more than 10 or 15 years which in nuns implies a continuity in the basic space data routinely 
used in these applications, and consistently to propose systems providing data quite similar to those presently 
available.To some extent, a continuity in the inversion procedures must also be maintained. 
Global studies can only be based on data sets resulting from the piecemeal collation of diverse data sets, which 
leads to major problems of spatial and categorical consistency. A first achievement in this respect is the 
AVHRR 1km global maps now provided by USGS as a result of a cooperative effort from 
NOAA/NASA/ESA/USGS at the request of IGBP DIS. Nevertheless the time needed for making a global 
composite remains prohibitive for a quick use. There is thus a need for a system able to provide global data, 
with a single processing, and with a time delivery acceptable for near real time applications. 
Obviously, the accuracy presently achieved for the various products may be increased, thanks to a better data 
quality (improved calibration, atmospheric correction, etc.). A higher accuracy can be expected as well from 
new observing techniques (other wavelength domains or other measurements : polarization, interferometry, 
etc.) which can be explored with experimental instruments. 
Finally the development of new methods is essential in the improvement of the accuracy of some applications. 
Systems can be characterized in terms of spatial resolution, revisit frequency, wavelength domain. Furthermore, 
time alloted for product delivery, one of the main requirement of some applications may be a driver of the 
system design. 
2.1 Spatial resolution 
The decametric resolution of SPOT or Landsat is required for the estimate of vegetated surfaces and yields at a 
field scale. Nevertheless, the swath of such sensors doesn't allow a wall to wall cover required in some 
applications. Furthermore, the revisit frequency remains insufficient to resolve the seasonal variations of some 
vegetation characteristics.The need for sensors with coarser resolution, like AVHRR, is justified for a frequent 
global coverage. The combined use of both kinds of data may be envisaged thanks to scale integration methods. 
All weather measurements at larger scale provided by microwave radiometers (SSMI) and geostationary 
satellites imagery with their higher repeat frequency will also benefit this integration procedure. 
2.2 Revisit frequency 
It is admitted that the adequate time sampling to monitor vegetation growth and maturation is around 10 days. 
This is obviously not accessible to high resolution radiometers, but to systems like NOAA/AVHRR. For tropical