Do position and species identity of neighbours matter in 8-15-year-old post harvest mesic stands in the boreal mixedwood?

Neighbourhood competition indices (NCI), where position and species identity of neighbours are known, have been used to investigate growth and competitive interactions among adult trees. In this study, we used NCI in 8-15-year-old stands following clear-cutting in a boreal mixedwood forest of eastern Canada to improve our understanding of early successional forest dynamics. Trees of increasing diameter from the center (>= 1 cm) to the edge (>= 5 cm) were mapped in twenty-five circular 450 m(2) plots. Target trees (DBH >= 1 cm) were sampled in plot center to determine their annual radial stem growth. For each species, we compared a set of growth models using either a spatially explicit NCI or a non-spatial competition index. Both types of indices estimated a species-specific competition coefficient for each pair of competitor-target species. NCI were selected as the best competition model for all target species although differences in variance explained relative to the non-spatial index were small. This likely indicates that competition occurs at the local level but that the high density and the relative uniformity of these young stands creates similar neighbourhoods for most trees in a given stand. The effective neighbourhood radius for competitors varied among species and was smaller for shade tolerant species. Intraspecific neighbours were the strongest competitors for most species. Aspen (Populus tremuloides) was a weak competitor for all species as opposed to balsam fir (Abies balsamea) which was a strong competitor in all cases. These results are in contradiction with some widely used forest policies in North America (e.g. free-to-grow standards) that consider broadleaf species, such as aspen, as the strongest competitors. For these early successional forests, the decision regarding the use of spatial or non-spatial competition indices should rest on the intended use. For even-age management, spatial indices might not justify their use in high-density stands but they are needed for the simulation of novel harvest techniques creating complex stand structure. (C) 2010 Elsevier B.V. All rights reserved.