Ensuring the purity of conventional crops grown in the vicinity of genetically modified (GM) crops depends on understanding both short and long distance pollen flows. New research shows that current guidelines on the safe isolation distances for GM maize may not adequately prevent cross pollination of conventional crops.

Contamination of conventional crops can occur where GM pollen cross-fertilises non-GM maize. The proportion of cross-contaminated seeds in the conventional field is the 'impurity rate' for that crop. Under European Union rules1, if the accidental proportion of GM to non-GM seeds exceeds 0.9 per cent then the crop must be reclassified and labelled as GM. Existing safe distances were largely established using 'paired field' comparisons, where contamination from a GM field is measured in a specific nearby field. The distances between the two 'paired' fields can then be adjusted to determine a 'safe' distance between fields. However, on a landscape level, other GM or non-GM maize crops in the vicinity may have an effect on pollen flow.

French researchers modelled the spread of pollen in a landscape containing a patchwork of GM and non-GM maize fields, as well as other non-maize fields. By taking into account the pattern of both short and long distance dispersal of GM pollen, the study explored the additional impact of more distant GM maize fields (i.e. not the closest GM field) on the impurity rate of the non-GM maize. For comparison, the impurity rates in a conventional field were also calculated using only the distance to the closest GM field.

Overall, the study showed that pollen from GM fields closest to conventional fields and the size of the conventionally planted fields have the greatest impact on the degree of contamination. However, as the proportion of GM maize to non-GM maize increases within the landscape, the impurity rate of conventional fields also increases. This increase was caused by long distance pollination from GM fields further from the conventional fields and suggests that if GM maize becomes more widely adopted by farmers, then existing models will underestimate the 'safe' distance between GM and non-GM crops. Importantly, the level of underestimation increased as more GM maize was included in the modelled landscape and when the isolation distance between GM and non-GM fields increased.

The researchers therefore suggest that, as long-distance dispersal of GM pollen can contaminate fields of non-GM crops and potentially raise the impurity rate above 0.9 per cent, pollen from all GM fields in the landscape needs be considered when setting isolation distances between fields of GM and non-GM crops. Further research is required to determine how to model these effects at the landscape level.