A Height Prediction Model with Random Stand and Tree Parameters: An Alternative to Traditional Site Index Methods

Publisher Website
Google Scholar
Abstract
A new method for predicting the height development of dominant trees is presented and illustrated with slash pine plantation data. The prediction procedure is based on a statistical model that explicitly describes the major random components in the variation of height curves. The average height of dominant and codominant trees in the population is expressed as a function of age. Then a model is developed for the variance-covariance structure of deviations from this average height curve due to stands and trees within stands. The model takes into account the fact that, within a given stand, heights of trees at a specific age are correlated, and tree heights and average stand height are correlated over time. The ability to specify and estimate this special covariance structure is obtained by including random stand and tree parameters in the height growth model. With the estimates of variances and covariances, and of the parameters of the function for average height, the height development of a given stand or single tree can be predicted using statistical prediction methods that utilize all available height measurements. With an example that includes two trees on one plot, one measured at one age only and the other measured at two ages, calculation of the variances of prediction errors is illustrated, and confidence limits on predicted values are calculated. For. Sci. 34(4):907-927.

This publication has 0 references indexed in Scilit: