Abstract:

Taking the establishment of single-tree biomass equations for larch (Larix) of the northeast and Masson Pine (Pinus massoniana) of the south in China as the example, the goodness evaluation and precision analysis were studied in this paper. Firstly, based on the comprehensive analysis of various statistical indices for model evaluation, six essential evaluation indices were presented which proved to be feasible and effective through evaluating the tree biomass models of two species above. Secondly, according to basic assumptions of linear regression estimation and related formulas of confidence intervals for prediction, the approach to estimate confidence intervals of conditional mean and single predicted value of tree biomass models was provided, and the confidence intervals of tree biomass models of two species were determined. Finally, the Monte Carlo simulation was used for random resampling test, and the test results showed that examination of applicability based on validation samples was inadvisable, and the adequate procedure was to utilize all sample trees for constructing the model, not splitting into two groups for model construction and model validation, which could make full use of the information of entire sample and decreased the prediction error to the highest degree.

Key words: single-tree biomass, allometric equation, goodness-of-fit, model evaluation, Monte Carlo simulation, prediction precision, confidence interval