Date of Award


Level of Access Assigned by Author

Campus-Only Dissertation

Degree Name

Doctor of Philosophy (PhD)


Forest Resources


Robert Seymour

Second Committee Member

Laura Kenefic

Third Committee Member

John Brissette


In New England, eastern white pine (Pinus strobus L.) and northern red oak (Quercus rubra L.) commonly grow together in mixed-species, multi-cohort stands, accompanied by eastern hemlock (Tsuga canadensis (L.) Carr.), red maple (Acer rubrum L.), and other hardwoods. Within-stand heterogeneity, with great variability in the spacing and juxtaposition of trees of different species, age, and size classes within small-scale neighborhoods of adjacent trees, makes such stands difficult to model and manage. Spatially-explicit data was collected on 121 plots representing a range of structural and compositional characteristics at the U.S.D.A. Forest Service Northern Research Station's Massabesic Experimental Forest. Two-year average annual growth rates for individual and per-hectare volume and biomass were modeled, using linear mixed effects, as functions of neighborhood-scale composition, and vertical and horizontal structure. The relative magnitude and direction of potential influences on growth were evaluated through comparisons among models, and examination of model coefficients. Exact neighbor positions relative to subject trees did not appear to be relevant to individual growth rates; plot-level averaged information (ie: stem density and quadratic mean diameter) performed equally well in prediction. Size of neighbors relative to subject trees was a critical variable in determining individual growth rates, determining both relative-size thresholds for one-sided competition, and degree of partial competitive asymmetry within the range of two-sided competition. Relative size metrics based on height or crown characteristics were superior to those based on diameter. Neighbor tree species composition influenced individual tree growth rate beyond relative size alone, and the reaction to neighbor species varied with subject species. Fit comparisons between models of varying levels of detail and complexity suggested vertical structure, after crown surface density, as the primary variable controlling per-hectare production rates in study stands. The contribution of residual-cohort white pine to productivity was not fully explained by its effects on structure alone, and hemlock was found capable of increasing total density and canopy layering when in the understory. Growth models for stands similar to this study should incorporate species composition and vertical structure information, but need not contain explicit spatial data. Managers are encouraged to consider stand structures exploiting the capacities of white pine and hemlock to create layered canopies above and below hardwoods.