A comparison of fitting techniques for ponderosa pine height-age models in British Columbia

Abstract

The ponderosa pine height models currently in use in British Columbia, Canada, were calibrated for southwest Oregon, USA. Height growth patterns in British Columbia may be different from those in Oregon. Furthermore, they may be different between biogeoclimatic zones within British Columbia. To check this, 80 stem analysis plots were established to develop a new ponderosa pine height model. One tree in each 0.01 ha plot was intensively sampled to obtain annual heights from the pith nodes. A conditioned log-logistic function was used as the base height model. Various model fitting procedures were employed to meet assumptions about the data and the regressions. These procedures included using an autoregressive model to account for serial correlation, and using nonlinear mixed modelling so that site index could be treated as having a random component. The final version of the model tested for differences in height growth patterns across the four biogeoclimatic zones where ponderosa pine most often grows. Although growth differences between the zones were detected, the results may be uncertain due to small differences in height growth trajectories and small sample sizes for some zones. A new height model for ponderosa pine is now available for British Columbia. This model gives only slightly different height estimates from the current models, so the use of the previous model in the past has not led to poor forest management decisions.