Derivation of regional crop sequences as an indicator for potential GMO dispersal on large spatial scales

A methodological approach is presented which aims to visualise the constraints for crop sequence planning in agriculture in a regional, large-scale context. In particular, the relationship between the scope of oilseed rape cultivation and the overall regional cropping structure, the share of particular farm types and the interactions between single crops have been analysed. The identified constraints have been applied to specify current and regionally typical crop sequences as input data for large-scale ex ante assessments, here exemplary for the genome dispersal risk in the case of GM oilseed rape cultivation.The regional and spatio-temporal variation of crop sequences for oilseed rape was analysed and generalised through a combination of analytical, classification and up-scaling techniques. In order to anticipate and assess the dispersal risks of transgenic oilseed rape, the methodology was tuned on crop sequences, which strongly influence the temporal dispersal of genetically modified oilseed rape. The regional cropping patterns for oilseed rape were analysed for the four northernmost German federal states: Schleswig-Holstein, Mecklenburg-Western Pomerania, Lower Saxony and Brandenburg. For typical regional crop clusters, specific crop sequences were derived, taking into account the constraints between crops and the weights for the particular crops as related to farm type. Real land-use data obtained at particular research sites were used to precisely determine the frequency of the single crops, as well as to discover sub-dominant crop combinations, which may have a high impact on dispersal processes. The introduced methodology stresses the following aspects: (i) reflection of the current situation due to links to periodically updated statistical data, (ii) implementation of the relationships and constraints between the different crops through statistical analyses, (iii) transfer of extensive, spatially limited agricultural data and expert knowledge to a large-scale context and (iv) integration of sub-dominant measures that are highly sensitive for particular processes.