This research compared effects of the weed control practice, soil cultivation, and the conventional practice, glyphosate application on weed seedbank, in a vineyard system. The experiment was conducted in a commercial wine-grape vineyard in the Napa Valley of northern California from 2003 to 2005. The annual treatments were "winter-spring glyphosate," "spring cultivation," "fall-spring cultivation," and "fall cultivation-spring glyphosate," and were applied "in-row," under the vine. Composition of the weed seedbank collected in 2002 before treatment establishment did not differ among treatments. After 3 yr of weed treatments, detrended correspondence analysis indicated that the composition of spring cultivation and winter-spring glyphosate tended to differ from each other, but the remaining two treatments showed little differentiation. As determined by linear discriminant analysis, the specific weed species were associated with seedbanks of certain treatments. These were Carolina geranium, annual bluegrass, brome grasses, California burclover, and scarlet pimpernel, which do not pose problems with regard to physical aspects of grape production. Although 'Zorro' rattail fescue was ubiquitous among treatments, its distribution between depths in the cultivated treatments indicated that tillage provided some homogenization of seedbank along the vertical soil profile. The seedlings from the seedbank study were not congruent with those measured aboveground in the field, suggesting that both treatment and microclimatic effects in the field may have influenced germination, and thus, aboveground composition.
