GenEERA I – Modelling for the dispersal behaviour of oilseed rape on a landscape scale – overall coordination

(2001 - 2004) University of Bremen, Center for Environmental Research and Technology (UFT), Dept. 10 Ecology, Faculty 2 (Biology/Chemistry)

Topic

The objective of the joint research project was to develop a methodological instrument that uses existing insights into the dispersal of oilseed rape on crop areas, bordering fringe biotopes and ruderal areas to derive conclusions for larger areas.

This was achieved by assessing outdoor findings, constructing models, remote sensing (analysis of satellite picture information) and the integration of mapped landscape correlations. Using the available information on the growth and dispersal behaviour of oilseed rape, a simulation model was developed that depicts areas of one square kilometre. This model was adapted in systematic scenario calculations by varying the environmental factors in such a way that all important factor combinations were depicted. The results can then be used to assess the dispersal dynamics over larger areas ("upscaling").

Summary

The models developed by the GenEERA project (GeneTraMP and RegioPol) are now available as a flexible instrument for depicting the spread of oilseed rape within the landscape framework of Northern Germany and for making preliminary forecasts. In so doing, it becomes clear that transgenic varieties used in agricultural practice would in the course of decades spread outside the cultivated area to an appreciable extent and would also cross into neighbouring fields. Not all the causes of variability have been included in the model yet. There is a need for further work at this level.

Important factors to be taken into account with transgenic oilseed rape:

Feral oilseed rape is more widespread in the agricultural area and in towns than was previously thought. On average, oilseed rape is found on one site per square kilometre.
Along roads and in towns herbicides are used. Herbicide -resistant plants would have a selective advantage here.
Because of the small field sizes in Northern Germany, an effective separation of conventional and GM crops would be complex.
After harvest, herbicide-resistant self-sown oilseed rape can lead to problems with volunteers in the following crop, since the herbicide resistance makes it harder to control the rape volunteers. This also applies to transgenes that have crossed into conventional populations.
Rape seeds can survive for more than ten years in the soil. This means that it is practically impossible to “recall” transgenes.

Experiment description

Since the dispersal behaviour of oilseed rape at the landscape level can no longer be understood by means of direct observation because of the size of the areas and the time involved for whole Länder, it is necessary to construct models and use geostatic extrapolation as new approaches for the assessment of environmental impacts. The approach developed in the project required intense cooperation between different fields.

The current project is divided into several part-projects and subordinate tasks:

The University of Bremen (UFT, Ecology Dept.) developed the modelling basis and obtained and processed satellite photographs showing the spatiotemporal distribution of oilseed rape cultivation in Northern Germany (Institute of Environmental Physics). Field data relating to the occurrence of oilseed rape and potential crossing partners outside crop areas were recorded in the Bremen area over more than 500 square kilometres and provided another modelling basis.
The information regarding the occurrence of oilseed rape in Bremen and the surrounding areas is processed by the Physiogeography Department of the UFT in a geographical information system in order to make it available for geostatic upscaling for the area of Northern Germany.
The botanical expertise for the dispersal ecology of oilseed rape and its relatives (Brassicaceae) is supplied by the Department of Systematic Botany of the University of Osnabrück and the Öko-Institut (Institute for Applied Ecology) in Freiburg as subcontractors.

Results

The project ran for three years and was completed at the end of 2004.

Dispersal analysis of oilseed rape on non-arable land

Oilseed rape and potential crossing partners were widespread in Northern Germany, even outside cultivated areas. On average one site of feral rape was found per square kilometre. Since the sites appear in clusters, the average distance of a few hundred metres fell within the possible pollen transfer range. Outside cultivated areas, oilseed rape occurred in the same habitats as its wild relatives. Ruderal rape and potential crossing partners showed a clear overlap in flowering times.

Diag.1: Modelling plant development. The model took account of the surroundings, development and cultivation measures.

Diag. 2: Model process diagram. The development of populations/individual plants is calculated based on the input data.

Diag. 3: Representation of the development of individual plants for three simulation years. A few plants reach seed maturity and the flowering periods of some plants overlap, despite high fluctuation, so that gene flow to related species is possible.

Diag. 4: Example of a graphic model readout. The plots represent an area of 6-13 hectares. Colour coding according to land use (WG = winter cereal) and GMO ratio in the oilseed rape crop. The model calculations depict cultivation on such areas for 10 years. It is possible to change the crop rotation, cultivation measures and climatic variation according to region.

Diag. 5: Pollen transfer from fields with GM rape to neighbouring conventional fields is more intense, the more fragmented the landscape. These are the simulation results for rape fields of different sizes from the scenario calculations.

Model construction and scenario development

Within the GenEERA joint project a method spectrum was developed that covers the description on a small scale of the processes involved in the spread and persistence of oilseed rape transgenes and enables upscaling of different local characteristics. The simulation models RegioPol and GeneTraMP were key elements. With the help of these models it was possible to simulate the development of individual rape plants and oilseed rape populations. The base information consisted of cultivation densities (analysis of satellite pictures, agricultural statistics), meteorological data (temperature, wind, precipitation, sunshine), maps (including classification of natural landscapes), the ecology of oilseed rape and related species and their dispersal data.

In order to be able to carry out the regional studies (GenEERA II to IV), the project team identified the climate conditions, prevailing regional crop rotations, etc. The models were calculated for the relevant conditions in a total of 1470 different combinations of input values and the results were made available in a database.

Simulation results

By way of example, Diagram 3 shows the simulation result for the survival of ruderal plants. A high dynamism is noticeable with populations appearing and spontaneously disappearing again. Only around ten to a maximum of twenty per cent of plants reach seed maturity, forming on average several hundred seeds per individual. The overlapping flowering phases of individual plants means that there is a possibility of gene flow to related species.

Diagram 4 illustrates the simulation of the dispersal process starting with a transgenic oilseed rape field using a sample map. Three oilseed rape plots are visible: the plot with GM rape has been coloured black (over 20 per cent GMO ). A dispersal zone with graduated grey tones (below 20 per cent GMO) stretches into the neighbouring conventional rape plot (1), showing the GMO content decreasing with distance. Another conventional oilseed rape plot (3) displays an even pattern of GMO. This plot was previously used for GMO, which led to volunteers in the current oilseed rape crop.

Simulation of model scenarios for Northern Germany

The scenarios were used to record the existing variability of the total area in small-scale simulations using a systematic combination of the input data. A total of 1470 different combinations were calculated.

The model can be used to investigate the interaction between the different effect paths. For instance, an assessment of the results by plot size clearly shows that because of the closer proximities in a fragmented landscape, higher GMO levels can be expected in conventional crops as a result of pollen transfer (Diagram 5). Further regional correlations are being investigated in studies by GenEERA projects II-IV.

gmoSafety: More on this topic

Further projects of the GenEERA research association (Generic analysis and extrapolation of oilseed rape dispersal):