A framework for uncertainty assessment of mechanistic forest growth models: A neural network example

Biing T. Guan, George Z. Gertner, Pablo Parysow

Research output: Contribution to journalArticlepeer-review

Abstract

A method for assessing the prediction quality of mechanistic forest growth models was presented. The method consists of four steps: assuming distributions for parameter values, parameter screening, outlining model behavior through sampling, and approximating model behavior based on the sampled points. The proposed method was then applied to a carbon balance stand level forest growth model. The Monte Carlo method was employed to perform the sampling, whereas the approximation was carried out with a 'patterned' artificial neural network. A massively parallel computer was used for the sampling and approximating the model behavior with the neural network. It was found that the initial parameter variance did not cause significant bias in predictions, and the variances of the predictions were mainly contributed by only a few parameters. Such information allows us to analyze the contribution of different model components and provides a basis to improve the model.

Original languageEnglish (US)
Pages (from-to)47-58
Number of pages12
JournalEcological Modelling
Volume98
Issue number1
DOIs
StatePublished - May 16 1997

Keywords

  • Artificial neural network
  • Conceptual model
  • Ecological modeling
  • Pipe model

ASJC Scopus subject areas

  • Ecological Modeling

Fingerprint Dive into the research topics of 'A framework for uncertainty assessment of mechanistic forest growth models: A neural network example'. Together they form a unique fingerprint.

Cite this