45th Congress of the European Regional Science Association: "Land Use and Water Management in a Sustainable Network Society", 23-27 August 2005, Amsterdam, The Netherlands
Heterogeneity of agricultural landscapes is supposed to be of significant importance for species diversity in agroecosystems (Weibull et al. 2003). Thus it is necessary to account for structural aspects of landscapes in land management decision processes. Spatial optimization models of land use can serve as tools for decision support. These models can aim at various landscape functions like nutrient leaching and economical aspects (Seppelt and Voinov 2002), water quality (Randhir et al. 2000) or habitat suitability (Nevo and Garcia 1996). However neighbourhood effects stay unconsidered in these approaches. In this paper we present an optimization model concept that aims at maximizing habitat suitability of selected species by identifying optimum spatial configurations of agricultural land use patterns. Bird species with diverging habitat requirements were chosen as target species. Habitat suitability models for these species are used to set up the performance criterion. Landscape structure is quantified by landscape metrics (McGarigal et al. 2002) estimated within the species home range. Statistical significance of these metrics for species presence was proven by a logistic regression model (Fielding and Haworth 1995). The landscape is represented by a grid based data set. Based on a genetic algorithm the optimization task is to identify an optimum configuration of model units. These model units are defined by contiguous cells of identical land use. Within this concept we can study how optimum but possibly artificial landscapes vary in structure depending on the selected species for which habitat suitability is maximized.