scope:

This VDI Standard is to be used in the monitoring of the environmental effects of genetically modified organisms pursuant to EU Directive 2001/18/EC and in compliance with the  provisions of the Council Decision 2002/811/EC for surveying wild bee fauna. The primary aim is the longterm observation of potential significant changes in the inventory or abundance of wild bee populations. This standard lays down the methodological standards relating to the production of data for a species inventory and on the population density for adult wild bees. Whether monitoring of wild bee fauna is necessary must be decided based on the GMO that is to be investigated. Conservation issues must always also be assessed to ensure that any contravention of the prohibition to kill or the prohibition on the destruction of the habitat of the highly protected wild bees is avoided BNatSchG Art. 44 Paragraph 1. The standardization of data collection methods guarantees a high degree of reproducibility and comparability of the data. The methods must be feasible and efficient, it must be possible to carry them out with an acceptable workload and they must generate data suitable for statistical analysis. Insofar as is possible, sampling and the selection of sampling plots must also provide the option of collecting causal evidence for a GMO effect. If this is not possible, then these methods must allow the derivation of hypotheses on the effects of the GMO on wild bee fauna that must then be specifically tested subsequently. The study design described here is essentially transferable to other research questions and groups of organisms.

The standard treats wild bees in their entirety. It describes the required periods for data collection and the field data collection methods and considers the specific requirements for the different data collection methods on bee imagos. Recommendations are given on plot selection and design as well as on the determination of statistical power. The use of the standard aims to ensure that data are collected on regional and supraregional effects in a variety of relevant habitats through a representative study.

Cultivation practices and the potential distribution of GMO and non-GMO plots pose some essential problems. Within the scope of monitoring studies, it is not possible to make a selection based solely on the comparability of the variants, as would be the case in a planned scientific field experiment, but the GMO cultivation plots specified by the farmers must be used as the starting point. Framework conditions similar to those in the laboratory will therefore hardly be possible to achieve on this basis

The handling of this fundamental problem constitutes an important basic element in this standard. When using the standard, minimization of (environmental) variation and improvements to the validity of the data are achieved through the following provisions, under consideration of the workload involved:

• Paired comparisons are carried out between similar GMO and non-GMO plots to minimize the confounding effects of variable environmental factors.

• Different complementary data collection methods are used (transect method and bowl traps) to ensure species documentation is as complete as possible.

• Combined data collection is carried out on species communities and their population densities in the form of a method with a detailed design.

• The validity of the number of possible repetitions (number of paired comparisons) will be estimated in advance using a power analysis, such that clear information is available on the dependability of the work involved.

• A variety of statistical methods are proposed to improve the validity, such that the optimum level of information is extracted from the available data, depending on data structure.

• The systematic collection of the data in a regional database (e.g. "Entomon" [54]) will allow or facilitate a variety of statistical analyses.

• Documentation and quantification of the actual state (habitat structure) of the study areas at landscape and field-level during the individual study years forms an important basis for the analysis of the study results, e.g. when using multivariate analysis.

• Indirect effects on wild bee populations that arise from the application of total herbicides to herbicide-resistant crops can be documented through a comparison of GMO and control plots.

Overall, this will result in the currently bestpossible treatment of the topic, with optimization of the validity of the results and a manageable workload.