Predicting Regeneration in Appalachian Hardwood Stands Using the REGEN Expert System
Vickers, Lance Alan
MetadataShow full item record
A study was initiated to adapt the REGEN regeneration prediction model to the Appalachians of Virginia and West Virginia. REGEN generates predictions via expert created REGEN knowledge bases (RKBs) that contain competitive rankings and stochastic parameters for selected species and size classes of advance reproduction. We developed RKBs for four site productivity classes (xeric, subxeric, submesic, mesic), and tested two (subxeric and submesic) using field collected inventory data in this study. To test the model we collected data from 48 paired sites which contained a mature stand and an adjacent regenerating stand (clearcut) of similar site productivity harvested within the past 20 years. Across all 48 sites, model predictions were within 5% of measured values on average, and explained 32% (R2 = 0.32) of the variation in species composition in regenerating stands. The species compositions of 41 of the paired stands on the Appalachian Plateau in West Virginia were further analyzed to compare species composition. Species composition was compared between the mature and regenerating stands in the subxeric and submesic site classes. A comparison of the upper canopy (dominant and codominant) species composition in regenerating stands to that of all stems â ¥ 1.5 in dbh in the mature stands was conducted as well. Our results suggest that the future species composition of stands regenerating following clearcut harvests will likely differ from previous rotations with mesophytic, shade intolerant species being more numerous. Oaks will likely assume a smaller role as the clearcuts mature, particularly on the submesic sites.
- Masters Theses