Hofmann Forest Site Quality Modeling and Estimation

Show full item record

Title: Hofmann Forest Site Quality Modeling and Estimation
Author: Dodrill, John D.
Advisors: Robert C. Abt, Chair
Raymond B. Palmquist, Co-Chair
Frederick W. Cubbage, Member
Barry K. Goodwin, Member
Jacqueline Hughes-Oliver, Minor Rep, Member
Abstract: An evaluation of site quality was conducted on the Hofmann Forest located in Jones and Onslow Counties of North Carolina. Site quality is a key element in the plan for management to reach its objectives. Accurately measuring site quality can determine the optimal locations for intensively managed plantations and the areas that may be better managed for other purposes. The identification of these areas will allow management to maximize the returns from the forest regardless of the objectives. There are a number of methods available for measuring site quality that includes both direct and indirect methods. The standard method among foresters in the United States is the direct method of the site index approach. The most common indirect method is the soil-site study. The objectives of this research are to examine these two procedures for their reliability and accuracy in determining the site quality on the Hofmann Forest.The foundation that the site index approach is built upon is the theory that tree height growth in relation to age is very sensitive to site quality but is independent of various factors; i.e. stocking density, species compositions, etc (Avery and Burkhart, 1994). Therefore if height growth is not independent of these factors then the site index approach will provide erroneous predictions. The current measure of site quality on the Hofmann Forest was provided by a soil-site study performed by T. S. Coile in 1966. A recent timber inventory of the forest provides a means of testing the accuracy of these measures. An intensive database was constructed from the timber inventory collected by F & W Services, Inc., the soil mapping performed by T. S. Coile, and previous climatic and environmental data for the forest. To achieve the objectives of this research the data was used to develop a non-linear model containing three categories or groups of variables. These categories include variables representing management decisions, soil characteristics, and environmental factors. The model produced an R-squared value of 0.62 with a p-value of <0.0001. A correlation analysis of the height predictions and the actual heights collected in the timber inventory produced a correlation coefficient of 0.79. Nested structural tests found management decisions very significant in influencing height growth. Soil characteristics and precipitation were not significant individually but were when used in conjunction with each other. A correlation analysis was performed on the model predictions and the previous predictions found in Coile's soil-site study, which produced a correlation coefficient of 0.06. Also, the correlation of Coile's predictions and the actual tree heights was 0.22. A forest wide average height at a base age of 25 years was predicted at 10.94 feet lower than that found by Coile.
Date: 2001-07-25
Degree: MS
Discipline: Forestry
URI: http://www.lib.ncsu.edu/resolver/1840.16/1648


Files in this item

Files Size Format View
etd.pdf 854.9Kb PDF View/Open

This item appears in the following Collection(s)

Show full item record