Assessing climate change impacts on tree growth and yield with process based and statistical models in the province of Newfoundland and Labrador
Loading...
Files
Date
2015
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
University of New Brunswick
Abstract
Forest Growth and Yield (G&Y) modelling is essential for timber supply analysis and
sustainable forest management. Empirical G& Y models are typically developed based
on historical data obtained from permanent sample plots (PSP) with an assumption that
forest will grow back at the same rate as in the past. With the predicted trend of global
warming, growth and yield projections with traditional G& Y models would become
invalid because the projected climate change will likely affect forest growth in the
future. Therefore, new and innovative approaches are required to account for the
potential impacts of climate change on tree growth and yield predictions.
In this study, process based and statistical models were used to assess climate change on
forest growth and yield in the province of Newfoundland and Labrador, Canada. A
process based model (JABOWAIII) was used to assess the impact of climate change on
forest G& Y. The model was calibrated with seventy PSPs that cover the entire
spectrum of soil and weather conditions. The model results were used to estimate
species specific climate change modifiers, which can be used to adjust the existing yield
curves to account for the effects of climate change.
A multiple regression and an Artificial Neural Network (ANN) models that take into
account biophysical conditions and historical climate factors were developed to predict
the G& Y of individual trees of forest stands. Thirteen independent variables were
included in the statistical based G&Y model. Stand and tree-level independent
variables included species distribution, stand age, tree height, tree diameter, stand basal
area. Biophysical and climate variables including growing degree days, potential solar
radiation, and annual precipitation were also included in the new model as input variables. First- and second-level auto-correlations of growth in individual trees were
also considered in the models. Results from the ANN-model were compared with those
produced with linear-regression models.
Both JABOWAHIII model and statistical model predicted that forest G& Y will be
negatively affected by warmer temperature. Uncertainties related to both process-based
models and statistical models are also discussed in the thesis.