Relative influence of crop rotation, tillage, and weed management on weed associations in spring barley cropping systems

Abstract

Generalizations concerning the effects of management practices on weed community dynamics often lack robustness, most likely because of the concomitant effects of agronomic and environmental factors. However, such generalizations, when valid, provide useful grounds for predictions and are thus desirable. This study attempted to evaluate the relative importance of crop rotation, tillage, and weed management as factors affecting weed communities and tested the hypothesis of an association between management practices and weeds from certain life cycle groups. Principal component analysis (PCA) of weed density data from a 4-yr field study conducted on a Kamouraska clay and a Saint-André gravelly sandy loam at La Pocatière QC, Canada, identified groups of weed species, while an analysis of variance (ANOVA) of PCA scores associated these groups with management factors. A multivariate analysis of variance (MANOVA) of regression coefficients describing time courses of density for each species confirmed treatment effects. Species segregated roughly according to life cycles. Interactions among weed management intensity, tillage, and crop rotation mostly explained species dominance in the various cropping systems. A first group of species, mostly annual dicots, largely dominated in minimum weed management treatments; their relative importance in each rotation varied with their level of susceptibility to postemergence herbicides. A second group included annuals and perennials, whose commonality seemed to be their tolerance to herbicides; these species also had a particular affinity for chisel and no-till treatments. A third group was formed by perennial species, each with a different response to tillage. The tenuous correspondence between commonly used classification schemes and management factors suggests that other aspects of weed biology (e.g., seed size, dispersal, production, germination requirements, and seedbank longevity) should be considered when trying to explain and predict the presence and dominance of certain weed species with regard to management practices.