Diameter Distribution Models for Tropical Natural Forest trees in Onigambari Forest Reserve

Abstract

The diameter distribution models of a natural stand in Onigambari forest reserve, South West Nigeria, were modeled with Beta distribution function. The data were obtained from 1.45ha temporary sample plot of size 25m x 25m. The data consisted of dbh measurement of all the trees with dbh ? 10cm in each plots. The data were sorted into species, families and dbh size classes. The models were developed using four – parameters Beta functions. Simple linear regression equation was used to fit the models for each of the parameters. To ascertain the accuracy and efficiency of the models, two-third of the data set was used for model validation which should not be significance to the observed values. The best model from each parameters were selected based on least values of mean residuals, standard deviation of residuals, sum of squares of residuals, coefficient of variation of residuals; significance and high coefficient of determination. The result revealed that sixty-two (62) tree species were present in the study area. Higher proportions of the trees were in the smaller diameter classes which is typical of natural forest. The maximum dbh observed was 98.4cm while the least was 10.0cm. The models fitted for each parameter are: (a+b) = 0.80lnDq – 200.79 (R² = 61.31%, SE = 20.13 and p = 0.0003),  b = 0.80lnDq – 210.78 (R² = 64.06%, SE = 24.78 and p = 0.0003), (p+q) = 6.78 – 0.59lnDq (R² = 34.37%, SE = 0.68 and p = 0.0216) and q = 5.70 – 0.55lnDq (R² = 24.76%, SE = 0.6343 and p = 0.0041). Quadratic mean dbh (Dq) appears to be the best predictor variable for the parameters estimated. The results of the models validation revealed that there were no significant differences between the observed and the predicted value of the parameters. Models developed in this study area are therefore recommended for application in the projection of diameter distribution for proper management.

Keywords: Diameter distribution models, tropical forest, management, Beta function