Determination of tree stem volume : A case study of Cinnamomum
Modelling of trees has attracted scientific research in various fields and disciplines since trees and forests play very important roles in the global system. It helps in the proper decision makings and implementation of policies. Hence, this research is designed such that the idea of determining th...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English English |
| Published: |
2013
|
| Subjects: | |
| Online Access: | https://eprints.ums.edu.my/id/eprint/41818/1/24%20PAGES.pdf https://eprints.ums.edu.my/id/eprint/41818/2/FULLTEXT.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Modelling of trees has attracted scientific research in various fields and disciplines since trees and forests play very important roles in the global system. It helps in the proper decision makings and implementation of policies. Hence, this research is designed such that the idea of determining the best models and solving their parameters that give the best estimates are conceptualized. The significant factors and their relationships are identified through a modelling approach. A modeling approach is developed which focuses on the phases in the model-building procedures, effects of interactions variables on the model, minimizing the effects of multicollinearity on the variables and recommending remedial techniques to overcome them, identification of the significant variables by removing insignificant variables, selecting the best model using the eight selection criteria (8SCs), and finally using the residual analysis to validate the chosen best model. Illustrations and algorithms are incorporated into the procedures. Non-normal and nonlinear data variables are addressed, hence data characterization is presented. A data transformation approach is introduced, based on the different data characteristics using the maximum coefficient of determination (R2) and maximum p-value approaches. Transformations are numerically optimized for linearity and normality of models. The three stem biomass equations adopted are namely, the Newton, Huber and Smalian’s formulae, based on the multiple regression (MR) and polynomial regression (PR) techniques. Relevant mathematical models are identified from the 684 models obtained in estimating the volume and the biomass equations used. The best MR model is model M52.5.5 Newton, however, the best PR model (P57.14.6 Newton) is found to give an improved estimation. Comparisons between the MR and PR models of the case studies are analyzed based on the eight selection criteria (8SCs). Factors contributing to the stem volume estimation are identified as tree height (T) and diameter at base (Db) as main contributors, while diameter at the middle (Dm), breast height (Dbh) and top (Dt) are significant contributors. Simulations of the best models are done using the Maple software. |
|---|