Covariance of cropping systems and foxtail density as predictors of weed interference

Abstract

Regression models of the effect of weed density on crop yield can form the basis of weed management programs by helping growers decide whether weed control is economically justified. However, few studies have examined whether one regression model can be used across a wide range of tillage systems and crop rotations. We used a nonlinear analysis of covariance to examine experiments conducted in 1990 and 1991 on the interaction of weed interference with conventional, fall chisel, and no-till systems, and rotations of corn, soybean, and wheat on a clay loam soil. Corn and soybean suffered heavy losses due to interference by green foxtail (a mixed population of robust purple and robust white varieties). Both tillage system and crop rotation altered the relationship between weed density and yield for corn in 1990 and 1991, but tillage was not a factor for soybean in 1991. Companion experiments on a sandy loam soil found no relationship between weed density and dryland corn yield in the drought year 1990, but weed density greatly decreased yield in irrigated corn. In 1991, the same model fit both dryland and irrigated corn grown in sandy loam soil. Foxtail density did not affect average weight per foxtail plant in any of our experiments, which indicates a lack of intraspecific competition. Competitiveness of corn better explained variation in dry weight per foxtail than did weather. Economic thresholds for foxtail interference are not constant but vary with weather, cropping system, and soil type.