Species-habitat association affects demographic variation across different life stages in an old-growth temperate forest

The sustainability of tree populations in old-growth temperate forests has motivated many ecologists to investigate the mechanisms by which species coexistence is maintained in a local-scale habitat. Species-habitat association could indicate assembly rules that support demographic processes in a particular habitat due to the effect of spatial distribution. In this study, we hypothesized that plant performance with a positive association is better than that for species with a negative association or with no association. We used demographic census data from 2010 and 2015 from the 9-ha Liangshui forest dynamic plot in northeastern China. A one-way ANOVA was used to examine the variation in the relative growth rate (RGR) and mortality rate (MR) in three life stages (saplings, juveniles, and adults) in different habitat types and to analyse the variations in the recruitment rate (RR) in different habitat types. The habitat types were determined by multivariate regression trees (MRT). The three habitat types are Hab I (very high soil organic carbon with a moderate slope), Hab II (low soil organic carbon with a gentle slope), and Hab III (low soil organic carbon with a strong slope). A t-test comparison was performed to compare the differences in RGR and MR between the positive associations group and the non-positive associations (negative and no association) group. The species habitat association statuses were determined by a torus translation test. We used a generalized linear mixed-effects model to analyse the effect of habitat types, diameter and habitat association status on the demographic rate in three different life stages. The RGR of the saplings was significantly higher than that of the juveniles and adults in the three habitat types (P < 0.001). The variation in the MR in the saplings and juveniles in habitats with high soil organic carbon was significantly higher than that in the adults (P < 0.001). However, the variation in the RR showed no significant difference among the three habitat types. The RGR in the saplings and juveniles with positive habitat associations was significantly lower than in those with negative and no habitat associations (P < 0.001). Moreover, the MR in the saplings with positive habitat associations was significantly higher than in those with negative and no habitat associations (P < 0.05). The species-habitat association status significantly affected tree survival, growth, and recruitment. These results demonstrate both positive and negative effects on forest dynamic processes across different life stages. Our study indicated that habitat association status has a significant influence on variation in demographic rates. We suggest that including different life stages in different habitat types is the key to obtaining ecological insight about forest dynamic processes, mechanisms of species coexistence, and maintenance of sustainable forest ecosystems.