145 
GVM - GAC PROCESSING CHAIN : 
A REMOTE SENSING TOOL FOR ENVIRONMENTAL APPLICATIONS AT 
CONTINENTAL LEVEL 
C. Carmona-Moreno*, M. Turchini*, M. F. Cruz Martinez**, J. Baron**, J. Lopez Raya José**, A. 
Moreno Ruiz** 
*JRC, SAI-GVM unit 
e-mail : cesar.carmona-moreno@irc.it ; www: //www.mtv.sai.jrc.it 
**UAL, GIASAAI 
ISPRS Commission VI Working Group 3 
Abstract 
The importance for Europe of maintaining independent space-based systems for environmental monitoring in 
the post-Kyoto context is currently the subject of debate by the Commission Services, European space 
agencies and industry. This is in line with the concept of the Pioneer Customer outlined in the Commission’s 
Communication on Space to Council and Parliament. Space Application Institute (SAI) is part of the 
partnership testing this concept and our work in the 5 th Framework Program will provide important 
information on the definition, design and implementation of any such system. 
In this context, “Global Vegetation Monitoring” (GVM) unit in SAI is in charge of “... the specification and 
implementation of dedicated observation systems (space and ground segments) that will meet the requirements 
for a reliable and accurate, space-based, global environmental information system for use by the European 
Commission...” This information system will be dedicated to monitor compliance with international treaties 
and conventions, in particular the Kyoto Protocol. 
GVM unit is currently implementing Remote Sensing processing tools in order to compliance with our 
objectives defined below. 
GVM - GAC processing chain was developed in this context in order to study the behavior of vegetation from 
long temporal data series. This is able to process AVHRR - GAC data set (15 years of daily data) for 
obtaining land-cover, fire and burned surface maps. 
This paper deals with the general presentation of GVM - GAC processing chain where the 
radiometric/geometric pre-processing algorithms implemented, the characteristics of the images from the 
processing, notably special vegetation indices: GEMI3, GEMI, ... will be considered. A comparison with the 
NASA - GODDARD pre - processed data will be presented in the framework of different thematic 
applications (e.g. burned surface assessment). We will especially see the absolute location accuracy (better 
than 1 pixel) that allows analyzing long temporal data series for reliable changes monitoring.