Simulation of control strategies for decision-making regarding Digitaria sanguinalis in glyphosate-resistant soybeans

Abstract

A bioeconomic model was developed for decision-making regarding large crabgrass (Digitaria sanguinalis) control in glyphosate-resistant soybeans in the Rolling Pampas of Argentina. The model was used to evaluate the economic returns of four different glyphosate-based strategies for weed control. In the absence of herbicide application (T1), the soil seed bank increases to an equilibrium density of 12,079 seeds m-2 in three years. A single herbicide application during the early stages of the crop (T2), which was intended to be highly effective in the control of an early weed cohort, allows a late, unaffected cohort to produce sufficient seeds to maintain population densities in the soil seed bank. A single, delayed herbicide application (T3), which was intended to control both early and late cohorts, results in a soil seed bank increase up to an equilibrium density similar to that achieved without treatment. Two sequential herbicide applications per year (T4), targeting the two cohorts, leads to a soil seed bank density after 10 years of 107 seeds m-2. Model predictions indicate that in the absence of control measures, a 93% reduction in soybean yield was predicted due to weed interference. The lowest reduction in crop yield (27%) was predicted using strategy T4, which is the most common control measure used by local farmers. This strategy clearly outperforms the other options tested, leading to lower D. sanguinalis seed bank densities and higher soybean yields and economic returns compared to those obtained using the alternative strategies.