ABSTRACT
Historically, the terrestrial surface has been studied from a disciplinary perspective driven by each discipline’s own speci‡cations and well-de‡ned objectives. Vegetation mapping by ecologists can be traced back to a century-long tradition, describing the surface in terms of presence and abundance of speci‡c plant species. Early work in terrain classi‡cation systems focused on the physiographic description of the land forms and of the plant physiognomic types. As for geographers, they were more concerned with land-use information gathered manually through ‡eld and socioeconomic observations. Similarly, public administrations and agencies relied on their own data speci‡- cations and data collection and categorization methodologies for de‡ning and recording land-based features of interest. However, such a disciplinary perspective for terrestrial surface characterization is no longer affordable. In addition, the scienti‡c agenda of these disciplines has shifted from land inventory to process understanding and numerical modeling. Meanwhile, speci‡c disciplines dealing with georeferenced information to study the terrestrial surface have emerged, supported by technological development in earth observation (EO), geographical information systems, and image processing.
