DISTURBANCE EFFECTS ON HERBACEOUS LAYER VEGETATION AND SOIL NUTRIENTSIN POPULUS FORESTS OF NORTHERN LOWER MICHIGAN

Recent disturbance models have identified changes in resource availabi lity as factors that control plant community response. Soil nutrient r esources typically are assumed to change following forest disturbance, usually with nutrient availablity increasing initially and subsequent ly decreasing through later stages of succession. We examined the effe cts of disturbance (clearcut harvesting with a brief recovery period) on soil organic matter, pH and extractable soil nutrients in successio nal aspen forests of northern lower Michigan to determine relationship s of these variables to changes in herbaceous layer vegetation. Two si te types were identified: dry-mesic (glacial outwash sands, low in org anic matter) and mesic (calcareous clay till, high in organic matter). Extractable nutrient concentrations were 1.5 to 3 times higher in the A(1) horizon of mesic sites than those of dry-mesic sites. Soil pH an d cations increased after disturbance on mesic sites, but not on dry-m esic sites. Patterns of change with disturbance were less pronounced i n lower horizons on both site types. Herb-layer species diversity incr eased after disturbance on mesic sites, but with decreases in the impo rtance of shade-tolerant tree species and Maianthemum canadense. Speci es characteristic of open habitats (e.g. Pteridium aquilinum, Rubus sp p., Fragaria virginiana, and Diervilla lonicera), increased in importa nce. Soil factors, species composition and diversity on dry-mesic site s changed little after disturbance, with Pteridium aquilinum and erica ceous species remaining dominant in both mature (55 - 82 yr) and distu rbed (less than or equal to 15 yr) stands. These results suggest that soil nutrient resources do not always change through secondary success ion and that patterns of change can be distinctly site-dependent. Dist urbance response patterns in the herbaceous layer of these aspen fores ts are also site-dependent.