Spatially explicit bioeconomic model for weed management in cereals: validation and evaluation of management strategies

摘要

Spatially explicit weed population models are very useful to explore long-term weed management scenarios. The economic implications of the management practices have been rarely considered in those models. Their inclusion could enhance weed management decisions. A bioeconomic spatially explicit model was developed and validated using field data. The model integrated a weed population submodel, a weed-crop competition submodel and an economic analysis. It was employed to evaluate long-term management strategies to control Lolium rigidum Gaudin, a troublesome weed infesting cereals in Mediterranean climates. A total of eleven individual and integrated management strategies, including cultural and chemical control, were evaluated. The predicted and observed distributions of population growth rates were similar for the four experimental plots used in the validation. One plot was fully validated with most proposed validation methods, while in the remaining plots, the predicted growth rates, in general, underestimated the observed growth rates. Our results showed that most individual management programmes were worse at controlling L.rigidum populations than integrated management ones. The programmes integrating cultural control strategies and herbicide at its full rate achieved the best long-term control. However, other strategies that were not as efficient at decreasing weed densities were more profitable economically, such as herbicide application at full rate. Economic returns were negative in all the strategies simulated. A sensitivity analysis showed that the seed bank and the economic outcomes were especially sensitive to fecundity and to the control exerted by management.Synthesis and applications. Our results suggest that the use of herbicide and low-input cultural control methods or full-dose herbicide applications are the most economically efficient strategies controlling L.rigidum populations. However, cultural control methods are not recommended as the only management strategies. Herbicide applications at full doses are preferable to applications at low doses, producing more successful weed control and, generally, less economic losses. The spatially explicit bioeconomic model developed has proved to be useful evaluating long-term weed management strategies in dry land cereal agrosystems. This kind of model would enhance weed management decisions for farmers based on an economic as well as an agronomic point of view.