Modelling and Simulation of Agricultural Landscapes

An agricultural landscape is a section of a region shaped by its natural landscape features primarily involving agricultural land use and land management. Intensive anthropogenic activities have left a permanent mark on agricultural landscapes, which have been developed over hundreds of years. Agricultural landscapes constitute a spatiotemporal structure. Hence they represent a complex system in which a large number of processes occur continuously that, due to their temporal dynamics, lead to constant changes in the state of the system. Traditional experiments are inappropriate for the impact assessment of anthropogenic and naturally occurring changes in agricultural landscapes. The only option here is to conduct virtual landscape experiments at the computer level. To this end, a relevant set of spatial, quantifiable landscape indicators is defined that can be used to map the landscape on the computer. Building on the extensive expertise in agricultural landscape research, indicators can be mapped using validated and robust models and their dynamics described involving temporal aspects. Various model types can be used in this process. Special simulation environments involving the use of spatial data and accounting for possible land use and global changes enable forward-looking scenario simulations to provide answers to the question of the sustainability of agricultural land use systems. Decision support systems (DSS) that exploit the latest possibilities offered by information technology, statistics and artificial intelligence provide the framework for integrating models, spatial data concerning the state of the landscape and scenario data, simulation techniques as well as tools for interpreting and visualising results. Such DSS are also the basis for quantifying the complex impact of site conditions, changes in land use or management, and of potential climate change on individual landscape parameters or landscape indicators. A number of examples show how indicator-based models of different types can be used to assess the impact and sustainability of land use systems on a landscape scale. As a prerequisite for the development and validation of integrated dynamic landscape models, more long-term ecological studies and monitoring systems are required. This also means that more resources are necessary to support these activities. The use of models and virtual simulation experiments within a DSS framework at the computer is a very promising way of finding suitable site-specific complex measures for the adaptation of agriculture to climate change.