Current weed demography models were reviewed to evaluate how the effects of cultural practices on weed dynamics were integrated into the models and to suggest possible ways to improve the simulation of cropping system effects. Several models were chosen to illustrate the interactions between cropping systems and weed dynamics. The first one described the structure of the weed life cycle. The second model integrated the effects of a wide set of cultural practices; the comparison of this example with other models suggested how the integration of cropping system effects could be improved. The last two models introduced the interactions of cultural practices with intraplot weed variability, either spatial variability of weed densities or genetic and phenotypic variability within weed populations. This review indicates some ways to make weed population models more comprehensive, robust, and accurate in order to improve their contribution to the evaluation and management of cropping systems.

It is now essential to reduce the negative impacts of weed management and especially herbicide use. Weed-suppressive crop species/varieties hold promise for integrated and sustainable weed regulation. Competition for resources and allelopathy are the two main underlying mechanisms. Unlike competition, which is well studied and established, allelopathy by living crops remains a contentious mechanism. A major difficulty to demonstrate the effects of allelopathy in the field is to dissociate them from those of competition. Here, we systematically and quantitatively review the literature, searching for field-based evidence of the role of allelopathy (by root exudation of living crops) in weed regulation, independently of competition, focusing on studies comparing different varieties of a given crop species. Our critical literature analysis also aims to identify weaknesses and strengths in methodology, providing insights on optimal experimental designs and avenues for future research. Our main conclusions are: (1) in most articles, the role of crop competition is disregarded or not exhaustively studied. Consequently, contrary to authors’ conclusions, it cannot be determined whether weed regulation is due to allelopathy and/or to competition. (2) Few articles provided convincing evidence of the presence/absence of allelopathy in the field. (3) To further investigate allelopathy in the field we recommend to (i) finely characterize crop competition by measuring traits in the field, (ii) assess crop allelopathic potential with complementary experiments in controlled conditions or by quantifying allelochemicals in the field, and (iii) quantify the contribution of each studied trait/mechanism in explaining weed regulation in the field with multiple regression models. In conclusion, the consistent use of the suggested guidelines, as well as alternative approaches (e.g., creation of varieties with deactivated allelopathic functions, development of process-based simulation models), may provide a basis for quantifying the role of allelopathy in the field and, subsequently, for designing weed management strategies promoting weed biological regulation.

Gene flow in rapeseed is a process taking place both in space and over the years and cannot be studied exclusively by field trials. Consequently, the GENESYS model was developed to quantify the effects of cropping systems on transgene escape from rapeseed crops to rapeseed volunteers in neighbour plots and in the subsequent crops. In the present work, this model was used to evaluate the risk of rape harvest contamination by extraneous genes in various farming systems in case of co-existing GM, conventional and organic crops. When 50 % of the rape varieties in the region were transgenic, the rate of GM seeds in non-GM crop harvests on farms with large fields was lower than the 0.9 % purity threshold proposed by the EC for rape crop production (food and feed) harvests, but on farms with smaller fields, the threshold was exceeded. Harvest impurity increased in organic farms, mainly because of their small field size. The model was then used to evaluate the consequences of changes in farming practices and to identify those changes reducing harvest contamination. The effects of these changes depended on the field pattern and farming system. The most efficient practices in limiting harvest impurity comprised improved set-aside management by sowing a cover crop in spring on all set-aside fields in the region, permanently banning rape crops and set-aside around seed production fields and (for non-GM farmers) clustering farm fields to reduce gene inflow from neighbour fields.