Weed resistance risk management in glyphosate-resistant cotton

Abstract

The introduction of glyphosate resistance into Australian cotton systems will have an effect on conventional weed management practices, the weed species present and the risk of glyphosate resistance evolving in weed species. Therefore, it is important that the effects of these management practices, particularly a potential reduction in Integrated Weed Management (IWM) practices, be examined to determine their impact on weed population dynamics and resistance selection. The study began in 2003 with a survey of 40 growers in four major cotton growing regions in Australia to gain an understanding of how adoption of glyphosate resistance had influenced the weed spectrum, weed management practices and herbicide use after three years of glyphosate-resistant cotton being available. The 10 most common weeds reported on cotton fields were the same in glyphosate-resistant and conventional fields. In this survey, herbicide use patterns were altered by the adoption of glyphosate-resistant cotton with up to six times more glyphosate being applied and with 21% fewer growers applying pre-emergence herbicides in glyphosate-resistant cotton fields. Other weed control practices, such as the use of post-emergence herbicides, inter-row cultivation and hand hoeing, were only reduced marginally. A systems experiment was conducted to determine differences in the population dynamics of Echinochloa crus-galli (barnyardgrass) and Urochloa panicoides (liverseed grass) under a range of weed management regimes in a glyphosate-resistant cotton system. These treatments ranged from a full IWM system to a system based soley on the use of glyphosate. The experiment investigated the effect of the treatments on the soil seed bank, weed germination patterns and weed numbers in the field. All applied treatments resulted in commercially acceptable control of the two grass weeds. However, the treatments containing soil-applied residual herbicides proved to be more effective over the period of the experiment. The treatment with a reduced residual herbicide program supplemented with glyphosate had a level of control similar to the full IWM treatments with less input, providing a more economical option. The effectiveness of these treatments in the long-term was examined in a simulation model to determine the likelihood of glyphosate resistance evolving using barnyardgrass and liverseed grass as model weeds. Seed production and above-ground biomass of barnyardgrass and liverseed grass in competition with cotton were measured. In all experiments, seed production and biomass plant⁻¹ decreased as weed density increased while seed production and biomass m⁻¹ tended to increase. Seed production m⁻¹ reached 40,000 and 60,000 for barnyardgrass and liverseed grass, respectively. In 2004-05, weeds were also planted 6 weeks and 12 weeks after the cotton was planted. Biomass and seed production of the two weeds planted 6 weeks after cotton were significantly reduced with seed production declining to 12,000 and 2,500 seeds m⁻¹ row for barnyardgrass and liverseed grass, respectively. Weeds planted 12 weeks after cotton planting failed to emerge. This experiment highlighted the importance of early season weed control and effective management of weeds that are able to produce high seed numbers. A glyphosate dose-mortality experiment was conducted in the field to determine levels of control of barnyardgrass and liverseed grass. Glyphosate provided effective control of both species with over 85% control when the rate applied was greater than 690 g ae ha⁻¹. Dose-mortality curves for both species were obtained for use in the glyphosate resistance model. Data from the experimental work were combined to develop a glyphosate resistance model. Outputs from this model suggest that if glyphosate were used as the only form of weed control, resistance in weeds is likely to eventuate after 12 to 17 years, depending on the characteristics of the weed species, initial resistance gene frequencies and any associated fitness penalties. If glyphosate was used in conjunction with one other weed control method, resistance was delayed but not prevented. The simulations suggested that when a combination of weed control options was employed in addition to glyphosate, resistance would not evolve over the 30-year period of the simulation. These simulations underline the importance of an integrated strategy in weed management to prevent glyphosate resistance evolving from the use of glyphosate-resistant cotton. Current management conditions of growing glyphosate-resistant (Roundup Ready &reg) cotton should therefore prevent glyphosate resistance evolution.