Projection variance partitioning of a conceptual forest growth model with orthogonal polynomials

An uncertainty analysis was performed for a conceptual forest growth model for red pine (Pinus resinosa Ait.). Low order orthogonal response surface models were developed to predict the variance of projections made with the conceptual model. Based on these orthogonal response surface models, it was possible to partition the variance of the projections. Approximate error budgets were developed, and the power of hypotheses tests based on model predictions were assessed.