Assessing unintended effects of GM plants on biological species

Abstract

A model for assessing the unintended environmental risks of GM plants to other biological species is outlined. The model starts by allowing the local values and concerns in a country to frame the identification of endpoint entities and endpoints. It classifies biological diversity into ecological functional groups that are directly associated with the values and concerns. Multi-criteria decision analysis is used to identify the highest priority endpoint entities for further evaluation. For each of these entities, risk hypotheses are identified and prioritized to specify assessment endpoints. This is illustrated with Bt cotton in Vietnam, with particular reference to potential harms if Bt cotton adversely affected a natural enemy species. The possible harm can be evaluated using an equivalence test. A method for establishing equivalence limits for generalist natural enemies is described using pest management theory. These limits are likely to be large, which implies that large changes in natural enemy populations are necessary to cause significant harm, and that sample sizes may be quite manageable to evaluate equivalence statistically. The model emphasizes clear, expert-driven, ecologically based decision-making and provides formal methods for conducting an environmental risk assessment.