Korean pine (Pinus koraiensis Sieb. et Zucc.) is one of the highly commercial woody species in Northeast China. In this study, six nonlinear equations and artificial neural network (ANN) models were employed to model and validate height-diameter (H-DBH) relationship in three different stand densities of one Korean pine plantation. Data were collected in 12 plots in a 43-year-old even-aged stand of P. koraiensis in Mengjiagang Forest Farm, China. The data were randomly split into two datasets for model development (9 plots) and for model validation (3 plots). All candidate models showed a good perfomance in explaining H-DBH relationship with error estimation of tree height ranging from 0.61 to 1.52 m. Especially, ANN models could reduce the root mean square error (RMSE) by the highest 40%, compared with Power function for the density level of 600 trees. In general, our results showed that ANN models were superior to other six nonlinear models. The H-DBH relationship appeared to differ between stand density levels, thus it is necessary to establish H-DBH models for specific stand densities to provide more accurate estimation of tree height.
The two dipterocarp species, Dipterocarpus alatus and Hopea odorata, have been widely planted in degraded forest land in Southern Vietnam in the last decades. However, study on growth characteristics of these species and their associated factors is still limited. Therefore, the present study aimed to examine growth performance in different stand densities, and classify tree quality in 28-year old D. alatus and H. odorata plantations. Our results of analysis of covariance (ANCOVA) indicated that in pure stand aged 28 years, D. alatus outperformed H. odorata in tree growth, biomass and volume. In addition, except for four growth variables including tree height at the first branch (Hb), crown length (CL), crown ratio (CR) and linear crown index (LCI), the remaining variables were negatively affected by stand density. We observed that medium quality trees occupied the greatest proportion in both D. alatus (47%; n = 425) and H. odorata (50%; n = 400). Except for CR and LCI, the class of good quality trees had the greatest values in the remaining examined growth variables. In linear discriminant analysis (LDA) model, the classification accuracy of testing set was relatively high in both D. alatus (85%) and H. odorata (91%). The most important variables for tree quality classification in D. alatus were crown diameter (CD) and diameter at breast height (DBH). Meanwhile, the most important variables in H. odorata were CD and tree height (H). These obtained results suggest that controlling crown size is important for shaping individual tree quality. Our data evidenced that D. alatus and H. odorata planted in Southern Vietnam with density equal or less than 500 trees per ha could yield high proportion of good and medium quality trees.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.