Creating a synthetic landscape: Spatial allocation of non-spatial single farm data

Abstract

Spatially explicit farm data are often not available for farm level studies and models, limiting the accuracy of simulating environmental impacts and posing limitations to agent-based models that simulate spatial decisions under explicit consideration of neighborhood effects (e.g. as on a land market). For large regions, such as the whole EU, different approaches have been developed to address this problem, e.g. by downscaling economic model results to lower-level administrative regions or by spatially allocating farm types to homogenous spatial entities. For smaller regions and if interactions between farms have to be considered, these approaches are too coarse.Some studies make use of a synthetic landscape approach for this scale and type of analysis. Farms are placed on a grid of the area under study by ensuring that the localization of the farms fulfills the area claims of the farms. Available approaches use relatively simple techniques by considering only the grassland share in the total area of farms or farm types and use a relatively coarse resolution (1ha*1ha). In this article, we present a method for improving the synthetic landscape approach, by considering landscape parameters in the allocation of farms, using a finer resolution (25m*25m) and introducing allocation quality indicators that allow for an assessment of the overall allocation result. For the German case study region Ostprignitz-Ruppin (NUT3 administrative level), the approach delivered a relatively high overall allocation quality. We conclude that the farm maps generated by our method can be used by mathematical optimization models or agent-based models to improve their capability of simulating spatial decisions and effects. The transferability of the method to other regions is possible, subject to data availability and the type of analysis targeted.