Journal of Applied and Natural Science 6 (2): 738-743 (2014)
Comparative study of some non-linear models for predicting the yield of Gmelina
F. O. Oboite1* and V. D. Ade-Oni2
1Department of Forestry and Wildlife, University of Benin, P.M.B. 1154, Benin City, NIGERIA
2Department of Forest Production and Products, University of Agriculture, P.M.B.2373, Makurdi, NIGERIA
*Corresponding author. E-mail: Felixoboite@gmail.com
Abstract : Yield models are important for effective forest management and as such were developed for the
University of Benin Gmelina arborea plantation, Nigeria. The objectives of the study were to develop, evaluate and
compare predictions from some non-linear models for timber volume estimation. A total of nine non-linear models
comprising of three models each for weibull, logistic and log-normal models were developed using the three
independent variables combinations (Basal area and merchantable height, diameter at base and merchantable
height, diameter at middle and merchantable height). The assessment criteria (correlation coefficient (R), coefficient
of determination (R2), standard error of estimate (SE)) with the validation results (using percentage bias and
probability plots of residuals) showed that all categories of weibull and logistic models generated in this study
discovered to be very adequate for tree volume estimation. The highest R2 (93.80), lowest SE (0.25) and lowest
bias% (1.29) in the study were achieved from Weibull model 1a. The log-normal models were the least adequate for
tree volume estimation with the highest bias%. The one way analysis of variance revealed that there were no
significant differences in the performance of the non-linear models when varying predictor variables were used. The
weibull, logistic models were therefore recommended for further use in this ecosystem and in any other forest
ecosystem with similar site condition.
Keywords: Gmelina arborea, Plantation,.
Keywords : Gmelina arborea, Plantation, Tectona grandis, Yield models.