Evaluating environmental risks of genetically modified crops: ecological harm criteria for regulatory decision-making

► Consensus on criteria that would allow a commonly accepted evaluation of environmental harm is presently lacking. ► Establishing criteria that define harm is not a scientific process, but a process of analysing and implementing policy requirements. ► Protection goals as specified by legislation are the exclusive starting point for risk managers for a definition of harm. ► We concentrate on providing guidance to risk managers on how they can specifically define what should be protected. European risk managers currently face substantial difficulty in evaluating the risks of genetically modified (GM) crops for biodiversity. This difficulty is not primarily due to a lack of scientific data (the data are abundant) but rather to a lack of clear criteria for determining what represents environmental harm. Establishing criteria that define harm is not a scientific process but a process of analysing and implementing policy requirements, and policy-makers and regulatory authorities need to define what is to be regarded harmful based on existing legislation. This process is a necessary pre-condition for the environmental risk assessment of GM crops. The present paper proposes a systematic approach on how harm can be explicitly and operationally defined for decision-making. Most legal frameworks require the protection of the environment or more specifically of biodiversity from harm. It follows that the first step in defining harm should be the characterisation of protection goals; protection goals are those valued environmental resources that should not be harmed by GM crop cultivation or by any other agricultural practice. In a second step, one must derive scientifically measurable entities (so-called assessment endpoints) on the basis of the protection goals. Such endpoints are required for regulatory decision-making because they specify what deserves protection. They furthermore allow quantifiable predictions of adverse changes during environmental risk assessment. Definitions of harm also require decisions on which environmental changes should be regarded as relevant and thus represent unacceptable harm. Using a case study comparing different effects of various pest management practices, the current paper proposes an approach that differentiates between intended effects that are acceptable and harmful unintended effects. By making explicit the assumptions underlying policy choices, the ecological criteria proposed here may result in a better and more transparen