Comparison of scale and scaling issues in integrated land-use models for policy support

• Published in: Agriculture, ecosystems & environment ecosystems & environment, 2011-07, Vol.142 (1), p.18-28
• Main Authors: van Delden, H., van Vliet, J., Rutledge, D.T., Kirkby, M.J.
• Format: Article
• Language: English
• Subjects: Decision Support System (DSS); Dynamic tests; End users; issues and policy; Land use; Land-use modelling; Mathematical models; Model integration; Policies; Policy support; Representations; Scale; Scaling; Temporal resolution; Tradeoffs
• ISSN: 0167-8809; 1873-2305
• DOI: 10.1016/j.agee.2011.03.005

► Integrated modelling for policy support is a balancing act between user, science and data. ► Users often require information at a scale that is not necessarily the appropriate scale to model the underlying processes. In such cases underlying processes need to be considered at finer or coarser scales than what is requested. Final results subsequently need to be upscaled or downscaled to meet user requests. ► In general, large-scale applications use less detail in the representation of the processes and coarser temporal and spatial resolutions. However, both the thematic resolution and the range of information provided by the model seemed to depend more on the requirements from the user than on the scale of the application. ► The spatial levels included in integrated models are generally closely linked to user requests. Furthermore we found that intermediate spatial levels are often necessary to downscale information from high spatial levels to local level. ► Integration between model components is much more than an upscaling or downscaling of model results. It is a challenge technically, but more important conceptually, as processes included operate at different characteristic scales and the models representing them are build on different underlying assumptions that can be contradictory. Recently an increasing number of integrated land-use models have become available that support policy making. Inevitably, their model components represent processes that act on different scales and that use different levels of detail to represent those processes. Therefore, it is a challenge to integrate them properly. In this paper we analyse and compare scaling issues from four integrated models that are explicitly spatial and dynamic. All have a strong agricultural component and are developed to support policy making. From these examples we find that scaling issues in model integration typically involve trade-offs among four factors: (1) the scale at which end users or policy makers require information; (2) the scale at which processes take place and the representation of those processes in a single model; (3) the way to integrate model components representing processes occurring at different scales; and (4) the limitations posed by practical restrictions such as data limitations and computation speed. Furthermore we conclude that the complexity of the model components and the spatial and temporal resolutions applied in the models are generally related to the size of t