Field evaluation of regression equations to estimate crop yield losses due to weeds

Abstract

Various regression equations based on weed density alone, or relative time of weed and crop emergence or crop density in addition to weed density have been developed in western Canada to estimate the effects of wild oat (Avena fatua L.) and volunteer cereals on yield loss of field crops, and to advise farmers on the economics of weed control with herbicides. In 1997, 1998, and 1999, several of these equations were evaluated in 9 barley (Hordeum vulgare L.), 9 wheat (Triticum aestivum L.) and 11 canola (Brassica napus L.) fields in Alberta. Wild oat was the dominant weed in the barley and wheat fields, and wild oat or volunteer cereals in the canola fields. In barley and wheat, more complex equations based on both weed density and either crop density or relative time of weed and crop emergence were more reliable in estimating yield losses due to wild oat than those based on weed density alone. In canola, an equation based on volunteer barley and canola density provided the most reliable estimates. Under the assumed crop prices and herbicide costs, these equations also resulted in the best estimates of whether or not a herbicide application resulted in a net profit or loss. Herbicide application was rarely economical in barley, but usually economical in wheat and canola reflecting the different market value of the crops. The implementation of the weed economic threshold concept is likely to be more feasible in low-value crops such as feed barley than in higher-value crops such as canola. Key words: Barley (Hordeum vulgare L.), canola (Brassica napus L.), volunteer cereals, weed economic threshold, wheat (Triticum aestivum L.), wild oat (Avena fatua L.)