54 
linear regression fit shown in Figure 2 reveals a strong linear correspondence (R2=0.71) 
between the classified and the EUROSTAT data and indicates that the classification under 
estimates the EUROSTAT cropland surface area statistics (AVHRR% = 0.89 EUROSTAT% 
+ 2.56). These results must be regarded with some caution however as surface area estimates 
made by pixel counting over large regions are biased when there are mixed pixels and when 
the classification accuracy is not high (Czaplewski, 1992). Further work will be performed to 
assess the classification results in a more comprehensive manner. 
CONCLUSION 
This paper has illustrated the complementary usage of remote sensing data and GIS 
information in the production of digital European forest and land cover maps. 
Régionalisation, pixel class labelling, and preliminary classification validation procedures 
were performed at a variety of spatial scales using ancillary remotely sensed preclassified data 
and spatial and aspatial information held in a GIS. The study indicates that summary and 
agglomerated statistics of high spatial resolution remote sensing images may be used to assign 
and validate AVHRR clustering results. The processing procedures described in this study 
could not be implemented in an efficient, reliable or timely manner without the use of GIS 
techniques.