Plant life span and persistence of soil seedbanks predict the emergence of herbicide resistance in noxious weeds

Abstract

AbstractPlant control methods have been developed to reduce weed species that are often problematic in agricultural systems. However, these methods can create new challenges, such as herbicide resistance. Determining which plant traits are associated with herbicide resistance can assist managers in identifying species with the potential to develop herbicide resistance and to better understand factors contributing to the evolution of herbicide resistance. We used random forest models to model herbicide resistance of noxious weeds as a function of 10 biological and ecological plant characteristics. Three noxious weed characteristics—plant life span, seedbank persistence, and occurrence in riparian or wetland microsites—predicted herbicide resistance with 87% accuracy. Species with persistent seedbanks and with short life spans (i.e., annuals) that occurred outside riparian or wetland areas were most likely to develop herbicide resistance. Short life spans indicate short generation times enabling faster evolution for herbicide resistance. Persistent seedbanks may increase the survival of resistant genotypes within a population or may be co-selected as an alternate form of escape from control methods. Species occurring in riparian or wetland microsites may be a case of “avoidance” rather than resistance, as managers typically avoid applying herbicide in these areas. Currently, 47 of the noxious weed species analyzed in this study are herbicide resistant, and our models identified an additional 63 species with traits that are highly associated with herbicide resistance, potentially indicating species that are at risk of developing resistance under conducive conditions. Further data-driven analyses with more plant traits and species from around the world could help refine current risk assessment of herbicide-resistance development.