Crop sequence, crop protection and fertility management effects on weed cover in an organic/conventional farm management trial

Abstract

A survey of 128 plots, in 2008, of a trial where the effects of crop protection can be separated from those of fertility management, generated weed cover data within six crops (winter wheat, winter barley, spring barley, potatoes, cabbages and a grass/clover ley). The effects of the 2008 crop types, of the two preceding crops and of organic and conventional crop protection and fertility management, were assessed using mixed-effects models and constrained ordination. Cover data for 22 weed species and for monocotyledon, dicotyledon, annual, perennial and total weed cover were used. Cover of 15 weed species, and of the five weed groups, was significantly affected by 2008 crops, with cover highest in spring beans and cabbage. Nine and four weed species 2008 cover were significantly related to crops grown in 2007 and 2006 respectively, as were dicotyledon, annual and total weed cover, but not monocotyledon or perennial cover. Cover of 15 species, and the five groups, was significantly higher in plots with organic crop protection, but only eight species and annuals were significantly affected by fertility management. Crop:crop protection produced the most significant interactions with most cover in organically managed plots. Five species, perennials and total weed cover produced significant three-factor models. The greatest weed cover was in organic crop protected but conventionally fertilised spring barley and the least in totally conventional winter barley. Other factors such as crop density and mechanical weeding also affected 2008 weed cover. The ordination indicated that most of the 22 species were strongly associated with crops from all three years. The sequence of crops in the rotation had a profound effect on weed cover. Where three spring-sown, difficult to weed, crops were grown in sequence (spring beans, potatoes and vegetables, spring barley) weed cover increased. However, cover was limited in grass/clover and some cereal plots with different preceding crops. Models predicting weed cover may need to take into account crop sequences within crop rotations, as well as the more usual management inputs.