Analysis of landscape pattern: towards a ‘top down’ indicator for evaluation of landuse

Landscapes as highly complex systems are subject to many different assessment procedures despite the fact that their development is not really predictable. Every synthetic model of landscape functioning will probably fail in demonstrating the landscape behaviour due to the decreasing precision and relevance of its output with increasing complexity. Two approaches are presented here which use deductive methods to describe landscape behaviour as indicated by its spatial structure. The one is a correlative concept which is characterized by all the advantages and disadvantages of statistical regression methods. It is strongly dependent on data quality. In the second study an expert knowledge system is developed. Fuzzy set theory is applied to transfer rules of landscape ecological experience to image parameters derived from satellite data. Both approaches are compared regarding the problems faced in the work during implication. Theoretical restrictions and the applicability in landscape assessment is discussed.