Effect of the structural variables of landscapes on the risks of spatial dissemination between GM and non-GM maize
Résumé
Crops grown from non-genetically modified seed should contain no more than 0.9% GM material. The very diverse nature of EU landscapes implies that coexistence measures must be adapted to regional conditions. Rigid rules, such as the requirement of large isolation distances everywhere, may raise the cost of coexistence measures, particularly in densely planted areas, without necessarily meeting the objectives of the European Commission. One of the objectives of the SIGMEA programme (Sustainable Introduction of GM Crops into European Agriculture is a Specific Targeted Research Project in the EC), within which this work was conducted, was to provide tools for the management of coexistence at the landscape level. We compared four contrasting maize production zones (in Alsace and Aragon) using MAPOD (R), a spatially-explicit model of gene flow, and randomly allocating GM crops to maize plots. We calculated two indicators of potential interest to different stakeholders (not only farmers and regulators, but also collecting firms and industry): (1) the proportion of the non-GM maize area not complying with a given threshold and (2) the mean cross pollination rate obtained if the harvests of all the conventional plots were to be mixed in a silo.Our results suggest that it is possible to classify the regional landscapes in terms of the level of risk (for different thresholds depending on market standards-e.g. 0.1%; 0.4%; 0.9%) with few structural characteristics easily calculated from GIS and agricultural statistics data (Mean field area; proportion of the Agricultural Used Area (AUA) devoted to maize, GM maize proportion, maize flowering dynamics and wind characteristics). The mean cross pollination rate increases with GM maize proportion in all the landscapes but the effect of this factor is modulated by structural variables of the landscapes.