Evaluation of a weed relative leaf area model for predicting yield loss in wheat

Abstract

The ability of a simple "two parameter" model based on the relative leaf area of weeds to describe grain yield losses in wheat (Triticum aestivumL.) was assessed. Wheat was sown at 100 and 300 plants/m 2 and oversown with six densities of mustard (Brassica nigra L.) to generate different levels of "weed" leaf area. Results showed that the model could be simplified to a one parameter model in this particular experiment. The single parameter q, defining the rate at which crop yield declines with increasing relative weed leaf area, did not vary between times of leaf area determination or between wheat sowing densities. However, other workers have shown that q varies substantially between experiments, limiting the predictive value of the model.

Evaluation of a mechanistic model of wheat and weed growth

“…The development of such a strategy requires (1) prediction of the long-term population trajectory of the weed and how this is influenced by alternative weed control methods (Cousens and Mortimer 1995), and (2) prediction of the economic loss caused by the weed (Auld et al 1987). Simple empirical models relating crop yield loss to relative weed leaf area are adequate models for the experimental system from which they are derived (Bourdôt et al 1997;Kenezevic et al 1995;Kropff et al 1995). However, their parameter values are too unstable among years and environments to enable reliable predictions of crop loss from the weed in other environments (Lotz et al 1996).…”