Abstract
Competition among individual trees is a fundamental structuring feature of uneven-aged mixed-species stands. We developed a Bayesian model for neighborhood competition in uneven-aged mixed-species stands, in which the interspecic variability in competitive impact is explicitly represented. We used data from 16 permanent plots in the University of Tokyo Hokkaido Forest, where the diameter at breast height (DBH) of all trees has been repeatedly measured and mapped. To analyze the competitive effect of neighboring trees, we used the “Neighboring Competition Index”, an index in which the competitive effect of neighboring trees is predicted as a function of its DBH, distance from the target tree, and its species. Among three alternative models, a model that represented the interspecic variability to have a common mean value using hierarchical parameter structure exerted the lowest DIC (a criterion for model selection), followed by a model that represented it to be independent. The estimated interspecic variability was smaller than previously reported, in that we found only three out of 39 species were considered to have a signicant difference from other species. Our results showed that although there is clear evidence for interspecic variability, the species-specific effects are not independent of one another, but rather have some similarity. Our results provide a novel approach for efficiently predicting the complex competition in uneven-aged mixed-species stands.
