COS 63-3 - Diversity stabilizes but does not increase sapling survival in a tree diversity experiment

Background/Question/Methods

Forest restoration is a popular conservation option given the potential to restore landscape diversity and sequester carbon. However, planting failure and early mortality pose significant barriers to restoration success. Biodiversity-Ecosystem Function theory suggests that diverse tree plantings should increase the likelihood of successful restorations through either the portfolio or facilitation effect on sapling survival, yet there remain few tests of this for trees. Here, we use a large-scale tree-diversity experiment (BiodiversiTREE) to test whether (1) tree species diversity reduces the risk of mortality, and (2) tree diversity stabilizes the risk of planting failure (i.e., a portfolio effect of reduced variation in overall survival at higher diversities). In 2013, almost 20,000 trees were planted into 70 plots at three levels of species richness: 1 (n = 32 plots), 4 (n = 19), or 12 (n = 19) species. Nearly 7000 individuals have been monitored yearly for survival. We analyzed survival data using both standard logistic regression and a discrete time survival analysis. To test for the portfolio effect, we calculated the coefficient of variation for proportion survival across all plots with bootstrapping to estimate 95% CIs.

Results/Conclusions

Survival varied significantly by species (χ2 = 684.8, p < 0.0001), with the lowest survival among the Carya spp. (51.6% ± 3.7%) and highest survival for Platanus occidentalis (97.5% ± 1.1%). In contrast, species richness did not affect survival (χ2 = 1.62, p = 0.44). However, higher tree species richness strongly reduced variation in survival, with more than 2-fold higher coefficients of variation in monocultures (24.4%, 22.5 - 26.3% 95% bootstrap CI) compared to 4- (11.6%, 9.4 - 14.0%) and 12-species plots (10.7%, 8.7 - 12.8%). We also tested whether survival varied by mycorrhizal association (i.e. functional diversity), and found that increasing abundance of arbuscular mycorrhizal associated trees in a plot was associated with a small but significant reduction in the survival of ectomycorrhizal associated species (χ2 = 3.89, p = 0.049). Ultimately, our results suggest that diverse forest restorations can restore diversity to the landscape with a lower risk of planting failure (i.e., the portfolio effect), but species selection still plays a large role in early forest restoration establishment.