Partitioning of the variance of predictions of a conceptual forest growth model

• Main Authors: Gertner G, Guan B, Parysow P
• Format: Conference Proceeding
• Language: English
• Published: 1996
• Subjects: analisis estadistico; analyse statistique; encuestas forestales; enquete forestiere; etapas de desarrollo de la planta; evaluacion; evaluation; forecasting; forest stands; forest surveys; modele; modelos; peuplement forestier; pinus resinosa; plant developmental stages; rodales; stade de developpement vegetal; statistical analysis; technique de prevision; tecnicas de prediccion
• ISBN: 9783905620528; 3905620529

Prediction quality assessment was performed on a carbon-balance model based on pipe model theory and self-thinning. The model was adapted to red pine (Pinus resinosa Ait.). As with many conceptual models, many direct and indirect sources were used to obtain the parameters for the conceptual models. Low order response surface models that were approximately orthogonal were developed to predict the variance and bias of projections made with the conceptual model. These response surface models were used to approximate error budgets. Error budgets show the effects of individual errors and groups of errors on the quality of predictions made with the overall model.