5 Conclusion

Despite the large number of available land-use change models, approaches for regional to global scale land systems which include the key elements of both the human and the environment sub-system, are scarce in literature. One reason is the levels of complexity and interactions that need to be represented by this type of models (which is also reflected in the large number of person months required for their development). The reviewed models all have their individual strengths, but are still limited in their ability to simulate all the complexities involved in the functioning of land systems. Based on the discussion in the last section, we have identified five major areas where future research activities should focus on.

• The first area is the improvement of model components for land-use decision making in order to facilitate more differentiated scenario analysis (i.e. in the main field of application of the reviewed models). Open research questions on for large scale include the simulation of changes of decision making mechanisms over time, the spread of innovation in terms of agricultural management practices and the influence of national or international institutions on land-use decision making. Here, future research can build upon experience and techniques from the field of agent-based modelling.
• There is also need to develop a deeper understanding of feedback effects on the model behaviour. In this context, a central question for further research will be the development of methods for testing and validating these highly integrated models.
• On the spatial scales which are subject of this review most models focus on the spatially dominant land-use and land cover types such as agriculture. Future studies should pay more attention to processes related to urbanisation which is recognized as an important driver for environmental change (Grimm et al., 2008).
• Currently, almost exclusively the effect of agricultural management on the environment is taken into account by the reviewed models. It will be a major step forward to incorporate more detailed representations of forest and water management. This will open up many new fields of model applications.
• Finally, on a more technical level we propose to investigate the application of open software frameworks such as OpenMI (Gregersen et al., 2007) to support integrated model building and to facilitate the exchange of model components between research groups.