Improving the management of coexistence between GM and non-GM maize with a spatially explicit model of cross-pollination

Abstract

The European Commission have established the concept of coexistence, according to which, farmers should be able to grow whatever type of agricultural crops they wish (genetically modified (GM), conventional or organic), provided that they comply with the legal obligations for labeling and/or purity standards. In the case of maize, the main factor conditioning the feasibility of coexistence is gene flow from GM fields to other types of production. The distance between fields has been identified as a key factor governing this gene flow. As a consequence, the existing regulations mostly concern the maintenance of a fixed isolation distance between GM fields and the closest non-GM field. However, other factors, such as temporal dynamics of pollen shedding, wind, relative field sizes and shapes and the spatial distribution of the different types of fields, may greatly modulate the effect of distance. Moreover, uniform distance-based rules create a “domino effect”, in which it is difficult for GM crops and non-GM crops to co-exist at the landscape scale. In this study, we hypothesized that the use of a spatially explicit gene-flow model, MAPOD®, would result in a significant gain in proportionality and freedom of choice for the farmer over uniform distance-based rules. To test this hypothesis, we performed a global sensitivity analysis on this process-based model but, instead of exploring a random set of situations, the sensitivity analysis was carried out on a subset of realistic scenarios based on farmers’ strategies. To select those scenarios, we constructed a multicriteria decision-making model describing the decision process used by farmers when deciding whether or not to grow GM maize, and used this model to generate realistic allocation scenarios for GM, non-GM conventional and organic maize cultivation. We showed that the coexistence method based on the MAPOD® model allowed the presence of a higher percentage of GM maize in the landscape than the distance-based method. This made it possible to follow the farmer’s field intended allocations more closely, whilst complying with the legal threshold requirements. This gain in proportionality was greater at high maize densities, for which the distance-based method allowed almost no cultivation of GM crops. However, in case of high proportions of organic fields, our study indicated that coexistence between GM maize and organic maize at the landscape level is difficult, if not impossible in case of farm-saved seeds, without a spatial aggregation of fields, leading de facto to separate non-GM and GM zones. Finally, the use of MAPOD® resulted in better discrimination between acceptable and risky situations, and greater flexibility, which is crucial for the implementation of an efficient coexistence strategy.