A site index model for lodgepole pine in the sub-boreal spruce biogeoclimatic zone of British Columbia.

An interior lodgepole pine stand level height growth and site index model is developed using stochastic differential equations and state-space approaches. Two data types, stem analysis and permanent sample plots were used. Four formulations of the Bertalanffy/Richards growth equation were evaluated using the EasySDE program. Data was partitioned based on data type, stand origin and combinations thereof. The best model was identified using maximum likelihood, and differences among data types and stand origins were evaluated for using likelihood ratio tests. A Power formulation was found to be statistically superior for all data types and origins. Combining the Stem Analysis and Permanent Sample Plot data in such a way that their respective variance structures was accommodated in parameter estimation was significantly better than the individual and origin specific runs, even though this increased the number of parameters. The resultant model was compared to the existing provincial model and differences were observed. --P. ii.The original print copy of this thesis may be available here: http://wizard.unbc.ca/record=b173786