Oct 31, 2005

Research Projects

Send

The importance of landscape structures for the development of post-market monitoring

(2001 – 2004) Bureau of Landscape Ecology and Environmental Research, (BLaU-Umweltstudien), Göttingen

Topic

Genetically modified crop plants with market authorisation must, in accordance with European Directive 2001/18/EU, be accompanied by post-market monitoring for unexpected environmental effects. Depending on the plant species and cultivation method, they might be tested for e.g. their specific multiplication, persistence and dispersal possibilities in farming areas. Farming areas, which are usually found in regionally diverse agricultural landscapes, are characterised by a mosaic of different habitats, often in close proximity and forming structures. Some of the habitats not used for agriculture have field-like properties and can present alternative habitats for crop plants.

Depending on the genes introduced, genetically modified varieties can, under certain circumstances, have impacts on the environment, particularly if the new trait gives the varieties better feral and persistence characteristics. According to observations to date, environmental effects in local farming areas are possible mainly with flowering and fruiting species like oilseed rape and cereals. Examples of new traits that might be habit-changing include:

• Disease/pest resistance,
• Changes to metabolic performance,
• Morphological markers with impacts on the phenotype,
• Improved winter hardiness.

If the new traits lead to changes in the behaviour of crop plant species, the following fields of action can be derived for post-market monitoring with regard to protecting ecosystems and conserving biodiversity:

• Monitoring the dispersal behaviour,
• Investigating changes in life communities,
• Monitoring impacts on species diversity.

The agriculture aspect includes the following fields of action:

• Novel cultivation methods,
• Impact on pesticide balances,
• Modified plant characteristics.

In the interests of efficient site selection and to obtain ecologically meaningful results, when implementing a monitoring network to register the aforementioned influence factors on agricultural landscapes, the possibility of testing both aspects simultaneously should be established.

The project investigated the following questions:

• How are landscape structures and habitats that may be suitable for the dispersal and establishment of oilseed rape and cereal species structured?
• Where are they found?
• Which habitats in a cultivated landscape offer favourable conditions for cross-pollination between oilseed rape and wild relatives?
• Where are the regional cultivation centres for the crop species in question?

These questions were to be used to derive a network of suitable landscape segments (agricultural landscape monitoring network) to form the geographical basis of a post-market monitoring system. Criteria for the selection of suitable landscape segments were developed with the help of empirical methods. In addition, best practice approaches were followed in collaboration with the Centre for Agricultural Landscape Research (ZALF e.V., Müncheberg).

Summary

Using 22 large agricultural landscape segments, each measuring nine square kilometres, criteria were established for setting up a network of sites suitable for comparative post-market monitoring of cultivated GM plants. The idea is that the monitoring network sites should serve as an indication of potential environmental impacts on agricultural and non-agricultural land (e.g. to detect dispersal tendencies).

A monitoring network focused on agricultural and environmental concerns needs to incorporate adequately sized landscape segments. The proportion of disturbed areas is an important selection criterion. The greater the proportion, the higher the probability of flowering and fruiting crop species becoming established on non-agricultural land.

Experiment description

Landscape analyses

Twenty-two landscape segments, each measuring nine square kilometres, were selected for characterisation of regional cultivation and habitat conditions in six Brandenburg landscapes (Fläming, Nuthetal, Barnimer Platte, Oderbruch, Uckermark and Prignitz). The distribution, sizes and types of the agricultural and non-agricultural ecosystems were mapped and described. Unused habitats are either linear (e.g. edges of paths, roads, ditches and fields) or extensive areas (fallow areas, uncultivated corners of fields, storage areas, etc.). Vegetation photographs formed the backbone of the mapping process, which was repeated over three years. Using digital aerial photographs and special CAD software, the recorded data were statistically prepared and typical landscape features were derived.

Long-term observations

Long-term observation areas were set up, primarily on farmland. Vegetation mapping was used to indicate changes at the level of plant communities. The idea was that the comparison between cultivation systems with and without genetically modified crop plants in the same region would form the basis for statements about changes. However, because of the steep decline in release projects since 2000, only a very limited comparison was possible. Most of the investigations were carried out on conventionally cultivated areas (status quo survey). Phenological surveys and similarity assessments also served as indicators for change.

Cost evaluation and optimisation

Further aims of the project were to evaluate the human, technical and financial resources and to investigate possibilities for cutting costs, e.g. by optimising monitoring networks. This was carried out in collaboration with the Institute for Plant Virology, Microbiology and Biosafety of the Federal Biological Research Centre for Agriculture and Forestry (BBA) and the BioMath company.

Results

Landschaftsanalysen

Je nach Landschaft zeigen sich oft regionaltypische Landnutzungen. Gleiches gilt auch für die Ausstattung mit nicht-landwirtschaftlich genutzten Biotopen (Abb. 1). Unter letzteren sind für ein umweltorientiertes anbaubegleitendes Monitoring vor allem störungsgeprägte Lebensräume von Bedeutung, z.B. Lagerplätze, Böschungen, Weg-, Straßen-, Grabenränder, Ackerränder, selten benutzte Wege und Anbauflächen von bestimmten Sommerkulturen. Sie besitzen stellenweise offene Bodenoberflächen, da sie durch menschlichen Einfluss regelmäßig gestört werden. Herrschen derartige Bedingungen an einem Standort, können sich auch konkurrenzschwache einjährige Nutzpflanzen außerhalb eines Ackers behaupten.

Classifying disturbed habitats

Disturbed habitats display a gradation in the incidence of annual species. The more field-like the conditions, the greater the probability that crop plants – whose seeds usually end up there as a result of handling losses – and some wild relatives will be found on these sites.

The mapping results suggest establishing a ranking according to the suitability of non-agricultural habitats for crop plants. Dividing the habitats into four categories reflects the probability of crop species becoming established outside agricultural areas:

• no occurrence (category 0),
• rare occurrence (category 1),
• occasional occurrence (category 2) and
• frequent occurrence (category 3)

For post-market monitoring, research into disturbed biotopes in categories 2 and 3 is to be recommended.

Significance of disturbed habitats in agricultural landscapes

The results of the inventory of disturbed biotopes in the 22 landscape segments are shown in Diagram 1. This shows that the original classification according to strict geographical landscapes does not always make sense. Where landscape segments are characterised by certain site factors, e.g. soil wetness or soil drought, these factors dominate cultivation conditions to a greater extent than the geographical location.

Based on the 22 landscape segments, the relative proportion of disturbed biotopes is on average between four and five per cent, although with considerable variations. The proportion of disturbed areas in the northern Uckermark region is on average 1.9 per cent; in Rädekow (moraine flatlands in central and eastern Brandenburg) it is 12.2 per cent. Low values are typical of particularly unstructured landscapes, while high values are typical of richly structured landscapes.

In order to take account of the variations in agricultural use and in habitats in a monitoring network focused on agricultural and environmental concerns, the network needs to consist of specially selected and adequately sized monitoring sites.

Long-term observations

It was possible to set up long-term observation areas on and near release sites in only two cases; all other observation sites were on conventionally cultivated areas. Plant stock levels were recorded using phytosociological methods during the three years of the experiment on at least three dates per year. Since the experiments with GM plants involved investigations over several years with phosphinotricin-tolerant (HR) oilseed rape and/or (HR) maize, the problem of oilseed rape volunteers appeared in subsequent years. Compared to conventional herbicides, however, the use of phosphinotricin does not necessarily result in a reduction of plant species diversity.

Selection of landscape segments for a monitoring network

When selecting agricultural landscape segments for a monitoring network, there are several aspects that need to be taken into account. The bigger the area planted with a crop species in a landscape segment, the greater the probability of finding crop plants in disturbed habitats outside fields. A high or low relative area of a (genetically modified) crop species in a region therefore provides a first indication of the importance of an area as a monitoring network site. An additional criterion to be taken into account is the proportion of disturbed areas in a landscape segment (see above). If high values of both criteria are used to make the selection, the resulting network will represent a ‘worst case scenario’ of landscape segments. This type of monitoring network is most likely to indicate dispersal tendencies among crop species.

The procedure for selecting monitoring network sites for oilseed rape monitoring is shown in Diagram 2. The diagram displays the proportion of land planted with oilseed rape and the proportion of disturbed land in a two-dimensional coordinate system for each landscape segment. A horizontal and a vertical (boundary) line help with the selection process. The vertical line shows the average proportion of farmland used for oilseed rape cultivation for all landscape segments and research years. It is located at about eleven per cent of farmland. The horizontal line represents the average proportion of farmland occupied by disturbed areas over the 22 landscape segments and research years (about 3.4 per cent of farmland). In the case shown, the selection would fall on the three landscape segments above the two red lines. In principle, additional decision criteria are conceivable. However, the two factors presented here take priority in the context of the project’s aim of developing a site selection method based on ecological criteria.