Simulating evolution of glyphosate resistance in Lolium rigidum I: population biology of a rare resistance trait

Summary Despite frequent use for the past 25 years, resistance to glyphosate has evolved in few weed biotypes. The propensity for evolution of resistance is not the same for all herbicides, and glyphosate has a relatively low resistance risk. The reasons for these differences are not entirely understood. A previously published two-herbicide resistance model has been modified to explore biological and management factors that account for observed rates of evolution of glyphosate resistance. Resistance to a post-emergence herbicide was predicted to evolve more rapidly than it did to glyphosate, even when both were applied every year and had the same control efficacy. Glyphosate is applied earlier in the growing season when fewer weeds have emerged and hence exerts less selection pressure on populations. The evolution of glyphosate resistance was predicted to arise more rapidly when glyphosate applications were later in the growing season. In simulations that assumed resistance to the post-emergence herbicide did not evolve, the evolution of glyphosate resistance was less rapid, because post-emergence herbicides were effectively controlling rare glyphosate-resistant individuals. On their own, these management-related factors could not entirely account for rates of evolution of resistance to glyphosate observed in the field. In subsequent analyses, population genetic parameter values (initial allele frequency, dominance and fitness) were selected on the basis of empirical data from a glyphosate-resistant Lolium rigidum population. Predicted rates of evolution of resistance were similar to those observed in the field. Together, the timing of glyphosate applications, the rarity of glyphosate-resistant mutants, the incomplete dominance of glyphosate-resistant alleles and pleiotropic fitness costs associated with glyphosate resistance, all contribute to its relatively slow evolution in the field.