Old-growth temperate forests: How their functional structure and composition changed over 30 years?
Major changes in biodiversity have recently been recorded in old-growth forests of North America. How the structure and composition of plant functional groups respond to these changes? Which environmental and spatial processes are responsible for these changes? To address these questions, we compared historical (1980) and recent (2011) surveys of shrub and herb communities conducted in an old-growth temperate forest located in the urban setting of Montréal (Canada). The species sampled were first classified into 26 emergent groups (EGs) based on four functional traits (life cycle, plant height, lateral extension, seed type). Temporal changes in community composition of EGs were assessed using a space-time interaction test (STI). For each survey, biotic, abiotic and disturbance variables collected within and in the surrounding of the forest were used to evaluate the EG-environment relationships, while spatial variables constructed using the Moran’s eigenvector approach served to test and visualize the presence of a spatial structure in community composition of EGs. Finally, variation partitioning analyses of environment and spatial variables were used to isolate the roles of deterministic (environmental) and stochastic (dispersal limitation) processes in shaping the spatial structure of communities.
The spatial distribution of 35% of EGs significantly shifted between 1980 and 2011, which resulted in substantial changes in community composition (STI: p = 0.016). In both surveys, the abundance of EGs in communities were explained by the soil drainage, the openness of vegetation cover and beaver damages, while some variables of distance to the edges (walking trails, forest margin) were significant only in 2011. This edge effect had a strong impact on the distribution of EGs with tall plants dispersed through vegetative propagation and/or fleshy fruits. A spatial structure affecting the community composition was detected for each survey. Overall, the abundance of EGs forming a clumped dispersion pattern had increased in 2011 compared to 1980. Variation partitioning revealed that clumped pattern found in 2011 was not only caused by the edge distances (deterministic process), but also by a stochastic process, suggesting that the environmental conditions and the intrinsic dispersal abilities of plants were concomitant causes of changes in community composition of EGs. Open areas linearly arranged (e.g., trails) and traits associated with competition for space (e.g., vegetative propagation) provide an opportunity for some functional groups to move between communities, and therefore to change the whole structure of biodiversity.