Environmental effects of genetically modified crops: differentiated risk assessment and management


  • D.E. Ervin
  • R. Welsh


A review of the literature shows that the environmental risks and benefits of genetically modified crops have varying degrees of certainty. For example, field studies have documented growing resistance to highly used pesticides. However, the risks of gene flow and deleterious effects on non-target organisms have not been evaluated at large field scales. Similarly, reduced pesticide use and toxicity have been estimated for some transgenic crops in some regions. Yet, the effects of herbicideresistant crops on erosion, carbon loss and supplemental water use generally have not been evaluated. Recent assessments have concluded that inadequate monitoring and evaluation of the ecological risks are being conducted. Among other limitations, the US regulatory system must rely on the small science base to assess the biophysical risks of transgenic crops. The system evaluates the occurrence of a suite of hazards for all such crops and applies the standard science protocol of minimizing type-I error (i.e., rejecting the null hypothesis of no environmental risk, when in fact the null is true). However, genetically modified crops vary widely in their potential for environmental risks, some with minor and others with major possible ecological disruptions. We illustrate a differentiated risk-assessment process based on the ‘novelty’ of the genetically modified organism, as measured by the genetic distance from its source of variation. As ‘novelty’ increases, information about hazards and their probabilities generally diminishes and more precautionary risk-assessment standards would be invoked. Three different models are illustrated: (1) the current US approach that controls type-I error for crops that are close to conventionally bred crops; (2) a model for transgenic and similar crops that minimizes type-II error (i.e., accepting the null hypothesis when the alternative of significant ecological effect is true) at a moderate power of test standard, and (3) a model for the most novel and complex genetically modified crops that imposes a very high power of test standard. The discussion then develops parallel risk-management approaches that include economic costs for the first and second models. The paper concludes with a discussion of how a biosafety regulatory system that effectively distinguishes the relative risks of genetically modified organisms can stimulate public and private research into a new generation of biotechnology crops that reduce unwanted environmental risks and perhaps provide ecological benefits