Nonlinear height-diameter models for <i>Gmelina arborea</i> (Roxb.) in Achusa, Makurdi Local Government Area of Benue State, Nigeria

Abstract

Gmelina arborea belongs to the family Lamiaceae and is a rapid growing species that is often grown to produce timber, paper and pulp as well as fodder. The height of a tree is a very vital information to monitor biomass and carbon stocks. Due to the increasing demand and high consumption for wood and wood-based forest products has incessantly been on a spike, and it could multiply because of an increase in the growing population. There is therefore the need to develop height-diameter models for the study area. This study was carried out at Achusa, in Makurdi Local Government Area of Benue State. The simple random sampling technique was used for the establishment of ten (10) sample plot size of 20m x 20m in the woodland. Six (6) non-linear height-diameter relationship model forms were used and evaluated statistically. The results of the fitted height-diameter models showed that Chapman-Richards and Weibull model were the best function for estimating tree height of Gmelina arborea for the training dataset. They had the smallest RMSE 3.223, a Bias of -0.005, R² value of 0.734, AIC of 510.314 and 510.321 respectively. For the validation dataset the Logistics model was the best with the smallest RMSE 2.906 and Bias of 0.013. Chapman-Richards and Weibull model were the best models in predicting tree height based on diameter at breast height values within the study area. The research recommends that variables such as soil fertility, density, spacing, crown area, age and silvicultural practices can be incorporated in height-diameter models so as to improve the accuracy and reliability of the models.