A novel approach for assessing the neighborhood competition in two different aged forests

Competition plays an important role in stand dynamics, but the study of competition conducted in natural forests with complex structure and composition is still limited. The traditional methods, which quantify the competition between individual trees, commonly ignore the information of species identity and spatial distribution. This study introduced a new competition index named Structure-based Competition Index (designated as SCI), which was calculated according to the characteristics of neighborhood spatial structure. The effectiveness of the new index was tested by: (1) analyzing its correlation with the traditional competition indices (Hegyi and BAL); (2) comparing its performance in predicting tree growth with the traditional indices, and (3) analyzing the competition effect on tree mortality. The data were collected in two different aged natural forests located in the Jilin and Gansu provinces, China. Results showed that SCI was significantly correlated with the traditional competition indices. Basal area increment (BAI) was positively related to initial tree size, and negatively related to competition. Overall, the combination of tree size and competition explained 60.0-70.6% of the total variance in BAI of trees in both sites. The BAI prediction models incorporating SCI as one explanatory variable showed better statistical performance than those including the traditional indices. The performance of traditional competition indices showed mixed results in differently aged forests: the Hegyi index performed better than the BAL index in old-growth forest, while the contrary result was found in secondary forest. In conclusion, SCI could be a useful indicator for assessing the competitive status of individual trees in natural forests. It was essential to consider the relative importance of tree size, competitive status and the species identity when analyzing tree growth and mortality in natural forests characterized by complex structure and composition.