Date of Award


Level of Access Assigned by Author

Open-Access Thesis

Degree Name

Master of Science (MS)


Forest Resources


Robert S. Seymour

Second Committee Member

Michael Day

Third Committee Member

Laura Kenefic


Managing uneven-aged, mixed-species stands requires balancing the need for high leaf area allocation in the overstory where it is most efficient versus the need to allow for sufficient growth of younger cohorts in the understory. To help forest managers make informed decisions to maintain this balance, the understory growth dynamics of northern conifer species in stands managed under uneven-aged silvicultural systems were studied. Sapling height growth of Picea rubens Sarg., Abies balsamea (L.) Mill, and Tsuga canadensis (L.) Carr. were modeled as a function of overstory canopy openness (gap fraction) using regression analysis. Research was conducted in four uneven-aged northern conifer stands on the Penobscot Experimental Forest in eastern Maine; two replicates each of selection cutting on five- and ten-year cycles. Gap fraction estimates were obtained directly above 167 sample trees between 0.5-6.0 m in height, using a LI- COR LAI-2000 plant canopy analyzer. These estimates were tested in several model forms along with initial tree height to predict sapling height growth. The effect of different vertical distributions of foliage on sapling height growth was also explored using analysis of covariance. Using cluster analysis, plots were grouped into one of three categories based on similar vertical leaf area structure. Species-specific height growth was then compared between groups of similar vertical structure using initial tree height as a covariate. An innovative method employing vertical point sampling was used to obtain leaf area estimates to quantify plot-level vertical leaf area structure. To validate the use of vertical point sampling, plot-level leaf area index (LAI) and basal area (BA) estimates based on vertical point sampling were compared with conventional horizontal point sampling using a 2 m2/ha basal area factor (BAF) prism. Tree-level LA1 estimates were replaced with specieslspecific constants based on projected leaf area (PLA)-height squared and PLA-DBH' linear regression coefficients in an effort to develop a quick and accurate method to estimate LA1 in the field using both vertical point sampling and prism sampling. Leaf area index measurements, BA, and tree tallies from vertical point sampling were also related to gap fraction measurements to determine if an efficient method for in-the-field gap fraction estimation could also be developed. Regression modeling demonstrated that sapling height growth of all three species followed a monotonically increasing pattern with respect to decreasing canopy closure. Abies balsamea appeared to be the most aggressive competitor demonstrating the greatest response to changes in gap fraction while Tsuga canadensis appeared to be the least responsive to changes in gap fraction. Although total plot-level LA1 was not significant in predicting height growth in these complex stands, the vertical distribution of leaf area was. While height growth of Abies balsamea and Tsuga canadensis were not significantly different between vertical leaf area structures, height growth of Picea rubens was significantly higher in plots with well-developed understories with high LAI, regardless of overstory LAI. Vertical point sampling showed strong promise in providing LA1 estimates, and in particular facilitating in-the-field LA1 estimation with the use of species-specific tree- level LA1 constants that remove the need for individual tree measurements. More field- testing of this technique needs to be done. Simple vertical point sample measures were not successful in accurately predicting gap fraction.