Herbicide resistance evolution can be tamed by diversity in irrigated Australian cotton: a multi-species,...

Because herbicide resistance evolves in very large populations over periods of many years, modelling is an important tool for investigating the dynamics of the problem. The Diversity model tracks the simultaneous evolution of resistance to multiple herbicides, using multiple genetic pathways, in several weed species at once. Tracking multiple species and simultaneous resistances is an important development in resistance modelling. We used the Diversity model to test weed management strategies for new cropping cotton varieties with multiple herbicide tolerances ('triple-stacked' varieties), in an Australian context.

The diversity required for long-term control of three key weeds in Australian cotton goes beyond using three herbicides, especially where there is already a substantial background of existing resistance to one or more of these herbicides. Assuming some glyphosate resistance is already present, simulations showed that glyphosate-resistant summer grass populations reach 20 000 seeds m^-2 within 12 years using the triple-stack herbicides (glyphosate, glufosinate and dicamba) and a minimum of other tactics. Adding three pre-emergent modes of action plus cultivation to the system effectively controls glyphosate-resistant grasses for over 30 years. In conditions where resistance genes are as frequent as 1 in 100, however, highly fecund weeds such as Conyza bonariensis are hard to control beyond 15 years even with very highly diverse management.

Stacked herbicide tolerances in new crop varieties offers potential for increased herbicide diversity, but existing glyphosate-resistant weed populations need substantial extra management beyond what a glyphosate/glufosinate/dicamba resistance stack provides. More diverse systems can provide robust management over 30 years in the absence of very high levels of background resistance to other herbicides. © 2018 Society of Chemical Industry.