Modelling interactive effects of herbicide application timing and dose for the control of Bromus japonicus in wheat in an arid environment

摘要

Timely herbicide application is a key factor in herbicide efficiency that may provide some possibility of reducing herbicide dose. Field experiments were conducted to study the interactive effect of dose of the herbicide mesosulfuron-methyl and application timing on the control of Bromus japonicas in wheat. Herbicide application timings were based on Zadoks growth stages of wheat. An empirical model was developed for deciding herbicide dose. Relationships between herbicide dose and weed biomass were described with the standard dose-response model at each application time. Parameter estimates suggested a linear increase in parameters W-0 (weed biomass with no herbicide treatment) and ED50 (herbicide dose required for 50% reduction in W-0) with thermal time. A combined dose-response (CDR) model, with W-0 and ED50 being replaced by their linear relationships with thermal time, was used for weed biomass prediction. CDR was incorporated into the rectangular hyperbolic relationship to predict wheat yield with different herbicide doses and application timings simultaneously. The combined model was rearranged to give the minimum dose required to restrict weed competition and achieve an acceptable wheat yield. For example, to obtain a wheat grain yield of 4.4kgh(-1) (90% of the potential yield), the model suggests 75% of the recommended dose, if herbicide is applied prior to tillering. With delayed herbicide application to wheat jointing, the respective amount of wheat yield is achieved with herbicide rates higher than the recommended dose. The model developed in this study can be used to decide herbicide dose for controlling B.japonicus based on application timing.