Competitive interaction among forest trees is a central process driving long-term forest tree demographic rates. Understanding tree competition is critical to both forest ecosystem dynamics and sustainable resource management. A likelihood-based spatially explicit model was developed to predict individual tree growth on a 1-ha boreal forest plot in western Canada. A series of competition indices that based on different assumptions of competitive interactions are evaluated for each dominant tree species on the plot. A competition index with the best performance in effectively predicting tree growth was selected and the effective neighborhood radius and effective neighborhood size were estimated for each species separately. The inter- and intra-specific competition effect were compared. All the alternative models are fitted by maximum likelihood approach and compared by likelihood ratio test.
Results/Conclusions
I found a distance-dependent competition index based on horizontal angle sums having an overall best performance for aspen and birch, NCI is relatively better for predicting growth of white spruce. Effective neighborhood radius and size varies markedly from species. Effective neighborhood size reflected asymmetric competition existed between large and small trees. There is a strong interspecific competition effect between shade-tolerant species and shade-intolerant species.