Gradient analysis and numerical classification of apple orchard weed vegetation

Importance values based on cover and frequency were determined for 36 weed species in understory samples from 47 apple orchards located in Henderson County, North Carolina. Orchard sub-groups identified by cluster analysis of the weed composition data included two major groupings dominated by grasses and woody perennials, respectively. Ordination by detrended correspondence analysis (DECORANA) was used to extract axes of compositional variation in orchard weed vegetation. Orchard sub-groups identified by cluster analysis could be recognized in the DECORANA ordinations. All grasses, most perennial forbs and all summer-annual forbs decreased in importance along a first ordination axis based upon data from 43 orchards. Two perennial forbs and all woody perennials increased in importance along the same axis, while species-richness of weeds declined. Results of multiple regression indicated that the annual frequency of paraquat application from 1976 to 1979 was the single factor accounting for the greatest compositional variation in the first two axes of the 43-orchard ordination. Tree canopy width and in-row spacing, total extractable soil cations, and percentage soil organic matter also accounted for significant amounts of variation in weed composition. Similar results were obtained from models developed for an ordination of the 35 grass-dominated orchards, suggesting the importance of management, structural and soil factors in determining weed composition in apple orchards.