Landscape research, landscape policy and Earth observation

'Landscape' refers explicitly to spatial structures of outdoor environments and the role of these structures in nature and society. Since repetitive land coverages of Earth observation systems such as SPOT (Systeme Probatoire pour l'Observation de la Terre) and Landsat provide a wealth of landscape images from local to continental scale, one would expect a more systematic use of these data by landscape researchers and planners than is currently the case. This paper states that the divergence between the central goals, vocabularies and methods in the distinct fields of landscape research, landscape policy and remote sensing may be prohibiting a systematic implementation. Examples are given in which the methodological steps necessary to bridge the gaps between these fields are explained. These are based on experiences in remote sensing data analysis fbr landscape ecological and landscape planning projects. The coverage over extensive areas, the grid format and the multi-temporal facility remain major assets of remote sensing. At continental scales, the coverage of remote sensing images in Europe is less complete than expected given the length of operation of both SPOT and Landsat Thematic Mapper (TM). Integration with spatial datasets of different origin (soil maps, road maps, etc.) is necessary to cover basic information needs. For landscape planners and researchers in general, specific qualities of spatial data are more relevant than their remote sensing origin. This is an important fact in stimulating the use of Earth observation data in landscape research. The mere presence of satellite images and their derived products (such as land cover datasets) is a 'push' factor both for landscape research and policy and the implementation of scientific knowledge and models into planning and policy. The task of 'intermediators' should be recognized against classical, if not mythical,'data providers' and 'end users' commonly identified in Earth observation.