COMPARISON OF DIAMETER DISTRIBUTION MODELS (Beta, Gamma, and Weibull) FOR TECTONA GRANDIS L.f. PLANTATION IN UNIVERSITY OF BENIN, BENIN CITY, EDO STATE

ABSTRACT

This study aimed to determine the most appropriate statistical distribution model for characterizing the diameter of a teak plantation in the University of Benin. Total enumeration was carried out in 34 temporary plots measuring 0.04ha (20m x 20m) easch, and a total of 234 trees with DBH ≥ 10.0cm were measured. The three distribution models considered were Beta, Gamma, and Weibull. Goodness-of-fit statistics such as bias, mean absolute error (MAE), mean square error (MSE), and Kolmogorov-Smirnov (KS) test were used to evaluate the performance of each model. The results showed that the Gamma distribution which had the smallest MSE and MAE of 39.9706 and 15.26254 respectively provided the best fit for the teak plantation data in terms of minimizing absolute and squared errors between the observed and predicted values. The Beta distribution which showed the lowest Bias of -39.9706 was recommended for consistent and unbiased predictions, while the Weibull distribution was recommended for preserving the shape of the distribution having showed a mean KS value of 0.10562. Additionally, Model 1 was recommended for volume prediction in the plantation due to the statistical criteria used in comparing three different models. It was shown in the result that model 1 had the lowest RMSE, AIC and BIC of 25.22834, 3.227968 and 8.696028157 respectively. However, Model 2 was ranked second even though it had the highest R² value (0.7494). The study concludes that the choice of the most appropriate distribution model depends on specific requirements of analysis, and further research should be done on larger data sets.