Regionalisation of climate variability used for modelling the dispersal of genetically modified oil seed rape in Northern Germany

摘要

The joint research project "Generic detection and extrapolation of genetically modified oilseed rape dispersal (GenEERA)" aimed at estimating the dispersal and persistence of genetically modified (GM) oilseed rape (Brassica napus) in a larger region by combining a small-scale individual-based model and an up-scaling approach, for which various data sources had to be evaluated to deal with local processes and spatial heterogeneities on the regional scale. The objective of the article at hand is to give a detailed account of the spatial variability of climate in Northern Germany (German Federal States of Brandenburg, Lower Saxony and Bremen, Mecklenburg-Western Pomerania, Schleswig-Holstein and Hamburg). Based on this, a method was developed that includes both, the determination of representative oilseed rape fields for modelling the dispersal of GM oilseed rape at field scale, and the subsequent generalisation of the results to the landscape level on a regional scale. Accordingly, land characteristics which could be important for the dispersal and persistence of oilseed rape were regionalised within a CIS environment: The beginning of flowering was modelled by variogram analysis and kriging estimation and was used to select satellite images for the detection of rape fields and to determine the period for the individual based modelling. The monthly means (1961-1990) of precipitation (P), air temperature (T), and sunshine duration (S) were regionalised by WARD cluster analysis. The PTS-clusters were combined to four climatic regions which together with WARD clusters on wind speed and direction as well as with land use clusters (crop rotation and management), which were provided by other project partners, enabled to finally define eight regions with a maximum of internal homogeneity. A meteorological station was selected to represent each of these regions. The data on wind speed and direction (hourly means), precipitation, sunshine and air temperature (daily) measured at that location were provided for modelling growth, dispersal and persistence of genetically modified oilseed rape on selected fields on the local level. Linking each of the modelled sites with a map on ecoregions which integrates the spatial patterns of soils, elevation, vegetation and climate, the model results were considered by analogy reasoning to be valid for all those ecoregions which are represented by the modelling sites and, thus, could be spatially generalised.