To reduce reliance on herbicides and maintain crop productivity, integrated weed management (IWM) seeks to optimize synergies between diverse sets of weed management practices combined at the cropping system scale. Nevertheless, data on weed community response to the long-term implementation of IWM practices remain scare. Here, we assessed the effects of four IWM systems with contrasting objectives and practices (S2: transition from superficial tillage to conservation agriculture; S3: no-mechanical weeding; S4: mixed mechanical and chemical weeding; S5: herbicide-free; all with 6 year rotations) compared to a conventional reference (S1: herbicide-based with systematic plowing and a 3 year rotation) on taxonomic and functional weed community composition and structure after 17 years of continuous implementation. We examined the legacy effects of these systems with a uniformity trial consisting of winter wheat managed uniformly across the systems as well as with a novel in situ weed seedbank approach involving tilled strips. We found that resulting weed communities in IWM systems were more species rich (species richness from 1.1 to 2.6 times greater) and more abundant (total density from 3.3 to 25 times greater) than those observed in the reference system, and differed in term of taxonomic and functional composition. In addition, we found that, when systems shared the same weed species, germination patterns of two thirds of the species differed between systems, highlighting the selection pressures some IWM practices exert on weeds. We showed that analyzing the superficial germinable seedbank in situ with tilled strips could provide a comprehensive view of resulting weed communities and be helpful in developing cropping systems that foster agroecological weed management.
