Methods for assessing regeneration establishment and height growth in uneven-aged mixed species stands

Abstract

This study shows how information from routine inventory data can be used to predict juvenile tree height growth for trees ≤1.3 m using simple competition measures derived from such data. Actual 5-year height increment is predicted by using: (1) the competition of the remaining overstorey, (2) the intra- and interspecies competition among the regeneration itself and (3) a modifier for the edge effect on incident light. For situations where no regeneration information is available, predictions were made of the occurrence of regeneration during the last 5 years using artificial neural networks. The results indicate that the probability of juvenile trees and the probability of a tree species per unit area are predictable with neural networks. The juvenile tree height growth analyses demonstrate that overstorey and inter- or intraspecific competition among the regeneration as well as the modifier for the edge effect on incident light were important to ensure accurate model predictions.