Progress on Predicting Crop Yield Loss from Weeds

Increasing concerns about the environmental impacts of herbicides, the development of herbicide-resistant weeds, and the need to carefully weigh costs and benefits have led to greater interest in developing rational, economical, and sustainable approaches to weed management (O'Donovan 1996). In addition, greater emphasis on post-emergence applied herbicides such as glyphosate has allowed for more flexibility in herbicide application timing, but more information is often required to effectively time applications to maximize efficacy (Boerboom, 2002; Kropff and Walter, 2000). A continuum of approaches exists for making weed management decisions, ranging from informed but subjective assessments made by farmers to recommendations produced by computerized decision-support tools, such as WeedSOFT (Mortensen et al., 1999). An optimal weed management approach will not simply maximize weed control but will also maximize economic gains while minimizing environmental risks. Since several outcomes must be considered simultaneously, and because interactions among numerous variables contribute to each outcome, the task of deciding upon an optimal weed management approach is complex and thus the role of comprehensive decision support tools has become more important. Furthermore, previous research has indicated that recommendations generated by decision support tools are often superior to subjective assessments in reducing weed populations and maximizing gross income (Wilkerson et al., 1991).