Adjusting a process-based growth model for varying site conditions through parameter estimation.
1993
Sievanen R. | Burk T.E.
The problem of estimating the parameters of a process-based growth model using typical stand growth measurements (of tree dimensions) is studied. The data consist of measurements of diameter, height. number of trees, and live crown ratio obtained from plots representing different site qualities. An analysis of the identifiability of model parameters is made which shows (i) that the structure of the model makes certain parameter combinations unidentifiable and (ii) that the data at hand do not support all the parameters. It is not possible to reduce the number of parameters in the model without losing its biological significance. Therefore, as a remedy for identifiability problems, the model has been modified slightly on the basis of identifiability analysis and initial estimations and the set of parameters to be estimated has been restricted. A division of the parameters into two groups is sought: those estimated with all plots combined and those that are allowed to vary from plot to plot. The estimated parameters have biologically reasonable values, and the variation in accordance with plot site quality is logical. Data-based analysis shows that apart from some unidentifiable parameter combinations, the parameters of the present model are estimated rather consistently. Analysis of the loss function components indicates that measurements of diameter, height, density, and live crown ratio are needed for reliable fitting of the model.
显示更多 [+] 显示较少 [-]