Development of Diameter Distribution Yield Prediction Models for Simulation of Acacia Mangium Plantations

• Main Author: Abd Kudus, Kamziah
• Format: Thesis
• Language: English; English
• Published: 1998
• Subjects: Mangium - Simulation methods - Malaysia
• Online Access: http://psasir.upm.edu.my/id/eprint/9449/1/FSAS_1998_36_A.pdf

The purpose of this study is to develop diameter distribution yield prediction models for predicting probability density function parameters when age, spacing and number of trees per hectare planted are known. Five distributions, Weibull, Gamma, Johnson SB, Lognormal and Generalised Normal were compared in terms of their ability to model diameter data in uneven-aged and even-aged forest stands. The classical moments were applied as a measure of flexibility of the distribution in regard 'to their changes in shape. Diameter data were obtained from 16 uneven-aged stands of mixed timber species located at Bukit Lagong Forest Reserve, Kepong, Selangor and 14 even-aged stands of Acacia mangium located at Segaliud Lokan Project, Sandakan, Sabah. The stands were all plantations and the ages range from 2 to 22 years. The diameter data were fitted to the five distributions by the maximum likelihood estimation method. The Johnson SB distribution showed the best performance in terms of quality of fit to the diameter data based on relative ranking of the log likelihood criterion. The estimation of Johnson SB distribution was further investigated and the nonlinear regression method was proposed for the estimation of the SB parameters. This method was compared to five other estimation methods; namely the four percentile points method, Knoebel-Burkhart method, linear regression method, maximum likelihood method, and modified maximum likelihood method through simulation. The performance of the nonlinear regression was confirmed by using the real diameter data. Goodness-of-fit tests based on empirical distribution function (namely the Kolmogorov-Smimov statistic, Cramer-von Mises statistic and the Anderson-Darling statistic) were used in selecting the most superior parameter estimation method. The results suggested that the nonlinear regression method was superior for estimating parameters of the Johnson SB distribution for the diameter data. In order to simulate the stand characteristics, equations were developed for predicting average height, basal area per hectare, and number of trees per hectare surviving when age, spacing and number of trees per hectare planted were known. The predicted stand characteristics were then related to the estimated parameters of the Johnson SB and solving the resulting set of equations for the scale and shape parameters. This study revealed that the parameter prediction method yields reliable prediction equations of the stand characteristics, but the prediction equations of the scale and shape parameters suggested that further research is needed to improve the model.