ABSTRACT
Nonlinear mixed-effects (NLME) is an important tool for explaining inter- and intra-plot/stand variations, especially when modelling height–diameter relationships. The NLME technique has not been used to model the height–diameter relationship of Gmelina arborea trees in Nigeria. Therefore, in this study, NLME model was developed for G. arborea stands in southwest Nigeria. A total of 1263 height-diameter pairs measured from 63 sample plots of 0.04 ha size were used to develop the NLME model. The model was calibrated using random effect predicted from the measured height of one to eight sampled trees of an independent dataset. Empirical best unbiased predictor technique was used to make predictions for the random component. Model evaluation was based on different indices such as root mean square error (RMSE), adjusted coefficient of determination . The results showed that the NLME model improved the accuracy of predicting heights of G. arborea trees with RMSE and values of 2.378 and 0.828. Calibration response of the model involved the selection of five sample trees per plot nearest to the quadratic mean diameter. This is the ideal compromise between maintaining accuracy and reducing sampling cost.
Acknowledgements
Thanks to Sacramento Corral-Rivas for helping with some parts of the R script.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Supplementary material
Supplemental data for this article can be accessed here.