Niche based plant community assembly emphasizes species sorting along environmental gradients. Determining whether large-scale anthropogenic disturbances alter species sorting with abiotic factors is necessary for understanding community recovery. Does large-scale timber harvesting, which has altered the landscape of millions of acres in eastern North America, affect whether species sort along environmental gradients? The understory herbaceous layer of rich cove forests in the Southern Appalachians provides fine-scale heterogeneity in the environment along with high species diversity within a forest landscape that has undergone large-scale logging. Three rich cove forests that were logged a century ago where selected to compare to three old growth forests that have never been logged. Within each forest, the diversity of the herb layer was assessed in early and late spring within 60-0.25 m² plots at each site. Similarity of species composition was analyzed with non-metric multidimensional scaling (NMS). Environmental parameters of canopy cover, tree basal area, tree species richness and 19 soil variables were measured at each site. Soil variables were reduced to two principle component axes (PCA). Multiple regressions were used to test whether environmental parameters explain species composition.

Results/Conclusions

Soil variables demonstrate a significant age effect for organic matter, estimated nitrogen released, phosphorus, zinc, iron and nitrate along the second PCA. In early spring, NMS results in a two dimensional solution with the second axis grouping old growth and hundred year forests separately, explaining 34% of variation in species composition. Species composition along the first NMS axis correlates with environmental parameters within both old growth and hundred year old forests. The second axis shows similar composition in each forest age with no environmental correlations. Early spring provides a more homogeneous light environment than late spring, with soil components explaining most of the variation in species composition along the first axis without deference to forest age. In late spring, multiple regressions between species composition axes and environmental parameters demonstrate that species composition does not correlate with environmental parameters within hundred year forests. In old growth forests, however, species composition in late spring is significantly explained by environmental parameters suggesting niche based processes in this heterogeneous environment. This demonstrates that long term recovery of late spring herbaceous layer species from large-scale logging disturbance results in decoupling of species composition from the abiotic environment.