Which model species for weed seedbank and emergence studies? A review

Models predicting the effects of cropping systems on weed demography are important tools for testing new rules for integrated weed management that may reduce the use of herbicides and preserve the biodiversity of agro-ecosystems. Such models already exist for a few species and should now be extended to a larger flora, in order to predict and understand the effects of agricultural practices on the evolution of weed communities. This review analysed the literature from 1973 to 2006, focusing on 45 species, to identify past reasons for choosing particular species when modelling the effects of cropping systems on the processes leading to seedling emergence. The frequency or harmfulness of the species were the main reason for studying them. It appears that the studied species were mainly autumn-emerging in north-western Europe cropping systems and summer-emerging in North America; the effects of deep soil tillage were studied mainly in Europe, as simplified sowing techniques are more often practised in North America. A voluminous literature exists on seed persistence in the soil, dormancy, germination and emergence, but rarely with the attempt of establishing generic relationships between species characteristics and model parameters. Until now, such an approach has been mostly developed in ecological studies. Taxa, as well as ecological preferences, seed size and the relationships of these characteristics with weed emergence model parameters should be considered when selecting a range of species for multi-specific modelling purposes.