Date of Award

12-2012

Level of Access Assigned by Author

Campus-Only Thesis

Degree Name

Master of Science (MS)

Department

Forest Resources

Advisor

Robert S. Seymour

Second Committee Member

Aaron R. Weiskittel

Third Committee Member

Brian E. Roth

Abstract

Pure stands of eastern white pine are common across New England as a result of agricultural abandonment; however a consensus on the optimal management regime has not been reached after nearly 40 years of debate (Leak, 2003; Seymour, 2007). To assist with developing thinning schedules, a density management diagram (DMD) was developed for eastern white pine using permanent and temporary plot data from self-thinning stands over 200 years of stand development. A maximum size-density relationship was fit to data from 72 plots using least squares regression which was independent of site quality. Using this biological maximum, relative density (RD) can be calculated for any stand as a proportion of this maximum. Data used to fit the self-thinning lines were well above A-line on the white pine stocking guide, suggesting these guides underestimate maximum density, and overpredict selfthinning. RD can be used to achieve a variety of objectives from maximizing stand-level growth to low density crop tree management and can be used to assess the relationship between stand density and growth. Leaf area is highly correlated with volume growth (Gilmore and Seymour, 1996; Innes et al., 2005; Jose and Gillespie, 1997) and is commonly used to quantify canopy responses to silvicultural treatments. Relationships between leaf area index (LAI) and relative density (RD) were examined in even-aged stands of Pinus strobus (L.) using long-term litterfall measurements over a wide range of stand development in both natural and silviculturally treated stands. Projected LAI ranged from 0.7 to 7.2 over RDs of 0.14-1.0. RD alone was a poor predictor of LAI, but when combined with TOPHT (mean height of tallest 100 trees per hectare) and site index (SI), stand development patterns of LAI were effectively modeled using a linear mixed effects model. LAI reached a peak at a TOPHT of approximately 12 meters followed by a gradual decline. Data suggests a nonlinear increase in LAI with increasing RD. At a given RD, LAI was strongly and positively correlated with site index (SI). Stand development patterns of LAI have implications for establishing rotation lengths and implementing thinning schedules for white pine and can be used to estimate LAI for even-aged stands. Leaf area is an important metric of growing space occupancy and is arguably the most important determinant of stemwood formation in trees and stands; however individual tree leaf area is difficult to measure directly. As a result, allometric equations using sapwood area have become the preferred predictor due to the strong physiological relationship between the conducting xylem and the amount of foliage a tree can support. Archived data from 61 destructively sampled trees was used to develop individual tree projected leaf area (PLATree) prediction models using both sapwood and crown-based parameters. PLATree models were tested at the stand-level by comparing allometric LAI (by summing up trees within the plots) to LAI obtained from litterfall. The Valentine model using a combination of basal area and modified live crown ratio performed the best at both the tree- and stand-level with the added benefit of not requiring destructive coring, however all models provided poor estimates of LAI when extrapolated to the stand-level. Tree-level models generally underpredict LAI in thinned stands and overpredict LAI in nonthinned plots likely due to a limited sample size of destructively sampled trees as well as a lack of larger trees in the dataset. To address the error when summing up PLATree equations to the stand -level a variety of stand-level models were developed by modifying tree-level equations and tested by estimating litterfall LAI for each plot. The best model predicted LAI as a function of sapwood area per hectare and crown length per hectare and reduced the bias from scaling up the PLATree equations. These models provided more precise estimates of stand-level LAI across a wide range of stand age and density, but are less practical as they require destructive coring of all trees.

Share