Analysis of pattern-process interactions based on landscape models-Overview, general concepts, and methodological issues

Pattern-process analysis is one of the main threads in landscape ecological research. It aims at understanding the complex relationships between ecological processes and landscape patterns, identifying the underlying mechanisms and deriving valid predictions for scenarios of landscape change and its consequences. Today, various studies cope with these tasks through so called "landscape modelling" approaches. They integrate different aspects of heterogeneous and dynamic landscapes and model different driving forces, often using both statistical and process-oriented techniques. We identify two main approaches to deal with the analysis of pattern-process interactions: the first starts with pattern detection, pattern description and pattern analysis, the second with process description, simulation and pattern generation. Focussing on the interplay between these two approaches, landscape analysis and landscape modelling will improve our understanding of pattern-process interactions. The comparison of simulated and observed pattern is a prerequisite for both approaches. Therefore, we identify a set of quantitative, robust, and reproducible methods for the analysis of spatiotemporal patterns that is a starting point for a standard toolbox for ecologists as major future challenge and suggest necessary further methodological developments.