Regenerating oak dominated stands descriptions, predictive models, and guidelines /

by 1974- Gould, Peter Jay

iii Regenerating oak-dominated forests is a major challenge faced by forest managers in the eastern United States. Even-aged silvicultural systems are considered the best systems for regenerating oaks, but in the central Appalachians (and elsewhere) oaks are nevertheless often replaced by other species after timber harvests. The present work addresses several aspects of the oak regeneration problem. The post-harvest development of 90 formerly oak-dominated stands ( > 50 percent oak before harvest) is described in terms of " developmental pathways. " The developmental pathways link stand conditions around the time of harvest to stand compositions and forest structures in the " third decade " (20 - 32 yrs) following harvest. Pathways were described by first identifying four outcome classes (based on third-decade compositions) using cluster analysis. Differences between outcome classes in stand compositions before harvest (overstory and advance regeneration compositions) and two years after harvest were then identified. Of the four pathways, only the OAK pathway resulted in the regeneration of oak-dominated stands. The loss of oak dominance in the other pathways corresponded with poor oak regeneration potential at the time of harvest. The RED MAPLE pathway featured a rapid post-harvest transition to dominance by red maple (Acer rubrum L.), a shade-tolerant species that established as advance regeneration in the pre-harvest stands. Stand compositions in the MIXED pathway changed markedly after harvest with the establishment of black birch (Betula lenta L.) and other pioneer species. Regeneration failure occurred in the UNSTOCKED pathway, which featured pre-harvest mortality and a post-harvest loss of regeneration. The pathways reflect several ecological models of forest succession. More importantly, they highlight the importance of stand conditions around the time of harvest to subsequent development. Models are presented to measure oak regeneration potential, expressed as expected third-decade oak stocking (i.e., the percentage of growing space occupied by oaks in the third decade after harvest). The models were created primarily using a longterm dataset developed from operational data collected in the course of timber sales, and a short-term research dataset collected as part of an on-going project examining iv regeneration development in Pennsylvania. A plot-level model was developed to measure the regeneration potential of advance oak regeneration sampled on milacre plots. The model uses plot aggregate height, the sum of oak seedling heights, to estimate the probability that an oak will occupy the plot in the third decade after harvest. Inverse modeling was employed to fit the plot-level model using both the long-term and shortterm datasets. A second model estimates the contribution of oak stump sprouts to thirddecade oak stocking, based on the species and diameter of parent trees. The two models are incorporated into comprehensive guidelines designed to aid forest managers in the regeneration of oaks. The guidelines recommend regeneration prescriptions based on management goals, expected third-decade oak stocking, and other stand conditions. Decision charts to assign prescriptions were developed using expert opinion and emerging information. Overstory removals are recommended when the current oak regeneration potential is adequate to meet management goals. When the current potential is lacking, treatments are to improve the seed-origin regeneration potential are recommended. The effects of competition on the development of individual oak seedlings following clearcutting are evaluated to test whether competition terms can significantly improve individual tree regeneration models. Current oak regeneration models do not explicitly consider competitive effects when predicting oak regeneration development. An unusual dataset from the Missouri Ozarks that tracks 405 naturally-established oak seedlings from before harvest to about 20 yrs after harvest was used. Models were fit to predict 20-yr success (reaching a dominant or codominant crown class) based on preharvest seedling height (size model) and pre-harvest height plus an index of local competition (competition model). The competition model produced predictions that better matched observations across a range of competitive conditions. The results suggest that including competition terms can significantly improve future oak regeneration models.