Comparing performance of Douglas fir growth and yield models in the South Island of New Zealand.

Type of content
Theses / Dissertations
Publisher's DOI/URI
Thesis discipline
Forestry
Degree name
Bachelor of Forestry Science
Publisher
Journal Title
Journal ISSN
Volume Title
Language
English
Date
2023
Authors
Walker, Liam
Abstract

Douglas fir (Pseudotsuga menziesii) is New Zealand's second most important plantation tree species. Of the total plantation area (100,105 ha), approximately 75% is planted in the South Island regions of Canterbury, Otago, and Southland. There are three common growth and yield models for Douglas fir in the South Island: the 500 Index model (500 Index), South Island Douglas fir model (SIDFIR) and the Douglas fir National model (DFNAT). Although frequently used, it is unknown how the models perform on datasets outside those used for initial validation.

Predictions of mean top height (MTH), basal area/ha and stocking by the three models were compared to 315 growth measurements across 8,376 ha of Douglas fir forests throughout Canterbury, Otago, and Southland. The ability of each model to match actual historical growth measurements in an independent dataset formed the basis for comparison. The effects of region and several site characteristics were also tested for their impact on residual errors of model predictions. Site characteristics shown to affect residual errors significantly were used to adjust model predictions to increase precision and reduce bias.

Substantial imprecision, systematic bias, and regional variations were found in predictions of MTH, basal area/ha and stocking by the 500 Index, SIDFIR and DFNAT growth and yield models. Regional variations and significant effects of site characteristics were also shown to exist. The SIDFIR model performed the best with the most precision and least bias of the three models; however, predictions still displayed considerable imprecision and bias.

Thus, developing a new growth and yield model for Douglas fir in the South Island is recommended. This model should utilise a hybrid modelling approach to account for climatic variations between sites and provide increased precision and reduced regional variation. A new model would allow forest managers to make effective decisions to ensure the productivity, profitability, and sustainability of Douglas fir forests in the South Island of New Zealand.

Description
Citation
Keywords
Ngā upoko tukutuku/Māori subject headings
ANZSRC fields of research
Rights
All Right Reserved