Tree survival is important for forests as the community structure and dynamics are determined by the demographic rates of individual trees. Once establishment has occurred, tree survival could be the most crucial driver of the maintenance of diversity and consequently ecosystem functioning. The often observed positive biodiversity–ecosystem functioning (BEF) relationship can be due to either increased growth or increased survival in more diverse forests. However, most of the survival analysis with standard methods do not consider the effects of covariates that vary with time and tree age, which might be why in the past we failed to explain the survival mechanisms. Detecting such complex controls on individual-level survival rates per year can largely improve our understanding of the mechanisms underlying BEF relationships and provide guidance for forest restoration. Using a dataset from a large forest BEF experiment in subtropical China (“BEF-China”), we analyzed yearly survival rates over 10 years for 40 species with different functional traits in monocultures and mixtures of 2, 4, 8 and 16 species.
Results/Conclusions
We found a large number of complex interactions among the covariates: 1) Survival strongly differed among species. Species with high survival tended to have thinner leaves and higher leaf N:P ratio. Species richness did increase the survival of species with lower leaf N:P ratio and affected species of different orders differentially. 2) Over time, survival increased more strongly in species with thick leaves and survival rate of species responded differentially to species richness. 3) Survival rates showed significant variations from year to year with better survival in wetter and warmer years, while species varied in their response to climate. 4) Coexistence may be even further increased by various richness levels with lowest survival for species with low leaf N:P ratios at high diversity in dry years and highest survival for species with high leaf N:P ratios in monoculture in wet years. These interactions between species, traits, years and climatic effects on survival demonstrate the balance between species across the forest development helping to maintain diversity. We conclude that only an integrated statistical analysis including time-dependent covariates and their interactions allowed us to elucidate the full complexity of tree survival patterns in developing forest plots of different species compositions.