Agron. Sustain. Dev. 25 (2005) 355-368
DOI: 10.1051/agro:2005035

Spatial aspects of gene flow between rapeseed varieties and volunteers

Nathalie Colbach^a, Nicolas Molinari^b, Jean-Marc Meynard^c and Antoine Messéan<sup>d

a</sup>  UMR Biologie et Gestion des Adventices, INRA, 17 rue Sully, BP 86510, 21065 Dijon Cedex, France
^b  Laboratoire de Biostatistique, IURC, 641 avenue Gaston Giraud, 34093 Montpellier, France
^c  UMR d'Agronomie, INRA-INAPG, BP 01, 78850 Thiverval-Grignon, France
^d  INRA Éco-Innov & CETIOM, BP 01, 78850 Thiverval-Grignon, France

(Accepted 25 February 2005)

Abstract - GeneSys models the effects of regional cropping systems on gene flow between cropped and volunteer rape plants. Its sensitivity to field patterns and cropping system input variables was analysed using Monte Carlo simulations of which the output was analysed by regression trees and linear regressions. The studied simulated output variable was the harvest contamination of conventional crops by genetically modified seeds. The study showed that field characteristics (areas, shapes, distances and orientation) have little effect compared with cropping systems. Gene flow increased with the area of the pollen- and seed-producing field and decreased with the area of the gene-receiving field; it increased for elongated vs. square fields and decreased with the distance between fields. The sensitivity analysis identified the distance and time during which cropping systems influence harvest contamination in various field plants. The harvest purity of a cluster of fields was influenced by the cropping history of the fields located less than 200 to 400 m around the cluster, and of three past years in addition to the ongoing analysed year for major variables such as crop succession or sowing date. The analysis of these major variables pointed to strategies for reducing gene flow, such as increasing the time between successive rape varieties; avoiding high-risk crops such as genetically modified (GM) rape and set-aside in the neighbouring fields for the year preceding the cultivation of a non-GM crop as well as the year during which the non-GM rape is grown; delaying the sowing of non-GM rape.

Key words: sensitivity analysis / model / cropping system / gene flow / oilseed rape / volunteers / Monte Carlo / input variables / survey plan / GENESYS

Corresponding author: Nathalie Colbach colbach@dijon.inra.fr

© INRA, EDP Sciences 2005