Weed aboveground and seedbank community responses to agricultural management systems

The development of integrated weed management programs requires a clear und erstanding of the factors and mechanisms conditioning weed community dynami cs in agroecosystems. This study evaluated the effect of different agricult ural management Systems on the aboveground and seedbank weed communities in annual row crops at the Long Term Ecological Research project in agricultu ral ecology at the W. K. Kellogg Biological Station, Michigan, USA. Weed bi omass and species composition were sampled for six years over two corn-soyb ean-wheat sequence cycles in four agricultural management systems: (1) conv entional (high external chemical input, moldboard plowed); (2) no-till (hig h external chemical input, no tillage); (3) low-input (low external chemica l input, moldboard plowed), and (4) organic (no external chemical input, mo ldboard plowed), A greenhouse germination study assessed variation in the a bundance and composition of the weed seedbank across the studied systems in the first and sixth year of this study. Aboveground weed biomass, species density, and diversity were lowest in the conventional system, intermediate in the no-till system, and highest in th e low-input and organic systems, but there were significant year-by-system interactions. Monocot and dicot species were equally common in the conventi onal system, whereas annual grasses, such as Digitaria sanguinalis (large c rabgrass) and Panicum dichotomiflorum (fall panicum), dominated the no-till system. Two perennial weed species (Trifolium pratense [red clover] and El ytrigia repens [quackgrass]) and one annual dicot (Chenopodium album [commo n lambsquarters]) dominated the low-input and organic systems. A multivaria te ordination of all four systems revealed close associations between the c onventional and no-till systems and between the low-input and organic syste ms. Separate ordinations of the four management systems revealed a crop eff ect in the low-input and organic systems, but no differentiation in the con ventional and no-till ones. The seedbank study revealed a significant increase in the number of weed se eds and species, mainly of annual grasses such as D. sanguinalis and P. dic hotomiflorum, in the conventional and no-till systems over the six years of study. During the same period, the number of weed seeds declined in low-in put and organic systems. Three annual dicots (Stellaria media [common chick weed], Veronica peregrina [purslane speedwell], and C. album) dominated the seedbank of the low-input and organic systems. Weed aboveground and seedba nk community composition were more constant over time in the low-input and organic systems than in the conventional and no-till systems over the study period. These results demonstrate that agricultural management systems can have both immediate and long-term effects on weed species density, abundan ce, and diversity. The differences observed among management systems in wee d biomass, species composition, diversity, and community constancy indicate challenges that exist for the development of ecologically based weed manag ement systems in row crop agriculture.