Simple to complex: Modelling crop pollen-mediated gene flow

Coexistence among genetically modified (GM) and non-GM cropping systems and identity preservation at the field level are increasingly important issues in many countries. Different types of pollen-mediated gene flow (cross-fertilization) models have been released during the past decade, primarily as a decision-support tool to achieve the European Union (EU) 0.9% GM adventitious presence (AP) labelling threshold for food and feed. We review key empirical or mechanistic models for four diverse crop types—canola or oilseed rape (OSR) (Brassica napus L.), maize (Zea mays L.), wheat (Triticum aestivum L.), and creeping bentgrass (Agrostis stolonifera L.). Their strengths, weaknesses, relevance, and utility in simulating pollen-mediated gene flow are examined. Many empirical models simulate gene flow well, although their utility is often restricted by datasets with limited environmental variability or spatial scale. Few mechanistic models have been developed, reflecting the challenge in accurately simulating pollen-mediated gene flow by wind or insects; such models have not been validated for commercial field scenarios. Many models tend to provide upper-end or worst-case outcrossing predictions and management recommendations, either because of experimental design underlying datasets, biological and environmental stochasticity, or chosen statistical analysis. Both experimental results and modelling predictions of outcrossing in OSR, maize, and wheat reveal that isolation distance or a pollen barrier (buffer zone) generally is only recommended between small grain maize fields (ca. <5 ha) to maintain field-average AP due to pollen-mediated gene flow below the EU threshold. Recent advances in modelling pollen-mediated gene flow in commercial fields are encouraging, but simulating gene flow in heterogeneous landscapes remains an elusive goal. Moreover, practical, user-friendly decision-support tools are needed to inform and guide farmers in implementing coexistence measures.