Most of the site index models recommended for use in British Columbia are base-age specific. These models are being updated to take advantage of newer fitting techniques, such as the Generalized Algebraic Difference Approach (GADA) and the grounded-Generalized Algebraic Difference Approach (g-GADA), which create base-age invariant models that have superior properties to those of base-age specific models. The main objective of this research is to develop a g-GADA model for lodgepole pine (Pinus contorta var. latifolia Engelm.). The g-GADA based on the Chapman-Richards function was applied to a set of lodgepole pine stem analysis data that were collected in British Columbia. For comparison, a GADA model and a mixed-effects model were also developed with the same data and base function. An independent set of data was used to test the models. Based on the model fitting and testing, the g-GADA model was generally better than the other models in terms of bias and standard error of the height prediction errors. The calibration of a model based on the g-GADA will likely require a numerical technique; pseudo-code for such a calibration is provided. The g-GADA, while similar to the GADA, has some advantages over the GADA.
Download Research Report 31 PDF file (882 KB)
To view this document you need the current version of
Adobe Acrobat Reader, available free from the Adobe Web Site.
Updated November 21, 2017