Comparisons of competitor selection approaches for spatially explicit competition indices of natural spruce-fir-broadleaf mixed forests

Determining the competitor selection methods for spatial explicit competition indices is important for individual tree growth modelling and forest management. There is a lack of systematic comparisons on these methods for natural mixed forest, however. In this paper, subplots with the area of 0.0625 ha were randomly sampled from 3 one-hectare remeasured and stem-mapped plots and repeated 200 times for natural spruce-fir-broadleaf mixed forests in northeast China with a bootstrapping method. Totally 600 subplots were used to examined the optimal competitor selection methods. Nine distance-dependent competition indices (CIs) and nine main competitor selection methods were tested by including them in the individual tree basal area growth models. After the analysis of the difference in the performance between partial model (only DBH was included) and full model (both DBH and CIs were included), we found that the difference among the statistically valid methods was weak (the increase in the adjusted coefficient of determination were 0.042 to 2.014%, the mean square errors were 0.060 to 2.851%). According to the magnitude and consistency of the contribution of the selection methods for a CI and the sensitivity to distinguish the competition effects on tree growth among the constituent tree species, we comprehensively concluded that the optimal competitor selection method was dependent on the CI, but the difference among these methods was weak. In addition, the CI from Alemdag (Alemdag IS (1978) Evaluation of some competition indexes for the prediction of diameter increment in planted white spruce. Information Report Forest Management Institute (Canada) No. FMR-X-108.) with the gradually expanding radius of circle method and 8 m radius of influence circle was the best combination which could efficiently detect the difference in competition effects on tree growth among tree species groups.