• Berven, Kate
  The Graduate School, University of Maine
  

This study investigates U.S. Forest Service silvicultural research in the northern mixed-conifer forest of the Northeast. Chapter I is an overview of three closed experimental forests and is a cautionary tale regarding lost research opportunities. The Paul Smith, Finch-Pryun, and Gale River Experimental Forests (EFs) were established in the early to mid 1900s. Changing societal needs and research priorities led to redirected staffing and funding; all three were disestablished. Initial investments were lost and outcomes of the experiments not fully realized.

The Penobscot Experimental Forest is an exemplary illustration of replicated, long-term research. I used data from the silvicultural experiment in an evaluation of the Northeast Variant of the Forest Service's Forest Vegetation Simulator, a growth and yield model (Appendix B). In addition, chapters 2 and 3 examine seedlings and saplings in seven treatments on the Penobscot EF.

In Chapter 2, I examine the factors (treatment, density of regeneration, species, treatment interval, and site-specific factors) influencing recruitment into the sapling class. I found that shade-tolerant softwoods had a lower probability of recruiting into the sapling class than other species groups. Tolerant softwoods recruited at greatest densities in three-stage shelterwoods; other treatments had a slower, continuous rise in in growth. Species-specific trends should be investigated further.

Chapter 3 focuses on seedling herbivory. My analysis indicated that a seedling's probability of being browsed was a function of its height class and species. Tolerant hardwoods had the highest probability of being browsed, and tolerant softwoods the lowest. Among species of interest, red spruce and northern white-cedar had the highest species-specific probabilities of being browsed. Eastern hemlock and eastern white pine had relatively low browsing probability and severity, in contrast to findings from other regions. Because larger height classes had lower probabilities of browsing, release treatments may be beneficial.

Long-term data provide information regarding forest dynamics through time. Data from the Penobscot EF provide the opportunity to investigate trends over long temporal periods and across multiple silvicultural regimes. This knowledge may be useful when developing long-term goals and defining desired outcomes for silvicultural treatments.