Diversity and assembly of weed communities: contrasting responses across cropping systems

Diversity and weed community composition of mid-season plant stands and autumn seedbanks were examined in spring barley–red clover cropping systems that varied according to crop rotation, tillage and weed management. Weed plant and seed density data collected over 4 years were used in the calculation of species richness (number of species), evenness (Shannon's E) and diversity (Shannon's H′), and in multivariate analysis (canonical discriminant analysis) of weed communities. Weed diversity indices were low (H′ < 2.0) but sensitive to management practices. Evenness had intermediate values (E = 0.4–0.8), suggesting little evidence of truly dominant species, particularly in the seedbanks. The difference in the number of species between treatments was never large (approximately two to four species). Overall, diversity indices were highest in the low disturbance treatments, particularly those with minimum weed management. Factors affecting ordination were somewhat different from those affecting diversity. Tillage had little effect on weed diversity indices but had a more major role in determining weed community composition. Seedbanks in no-till and monoculture-chisel plough treatments appeared to have more distinctive species composition compared with other treatments. Weed species assembly in seedbanks showed little discrimination across treatments and over time, confirming the ability of seedbanks to buffer disturbances across a variety of cropping systems. The use of diversity indices revealed part of the complexity of weed communities associated with disturbance in cropping systems, whereas ordination singled out species–cropping systems associations, which may be more meaningful to weed management.