Trait gradients and functional diversity predict ecosystem-level litter decomposition pattern in a mixed conifer-hardwood forest

Background/Question/Methods

1. Litter decomposability is coordinated with the functional traits that determine plant resource economics. Community-assembly processes influence the distribution of economic traits, generating heterogeneity in the functional composition of litter. Community-weighted trait means (CWM) were used to predict the distribution of decomposition rates in a mapped conifer-hardwood forest plot (Haliburton Forest).
2. Function diversity (FD) is increasingly tested for alongside CWM to account for non-additive effects of trait dispersion on ecosystem processes. We briefly discuss how averaging techniques used to provide CWM can generate an aggregation bias, potentially leading to the false detection of a diversity effect if processes respond nonlinearly to trait values.
3. Four years of species-specific litter trap data were collected. Decomposition rates were estimated by balancing litter production with organic-horizon mass. We introduce a new method of estimating initial mass-loss rates inspired by a two-phase decay transition model. 

Results/Conclusions

1. Partitioning organic-horizon samples into humic and non-humic portions improved predictions in all cases. Decomposition was positively correlated with community-level foliar nitrogen and litter nitrogen, and negatively with leaf-mass per unit area and litter C:N. Slopes of log-log relationships indicated that decomposition responds to CWM nonlinearly. However, a negative FD effect was detected while the form of nonlinearity observed would have contributed to false detection of a positive FD effect. Incorporating FD had a much greater impact on predictions than species-level diversity indexes.
2. Coordination between CWM and litter decomposability suggests that community-level influences on biogeochemistry complements the current distribution of functional traits, acting to stabilize relative abundances along trait gradients. Negative diversity effects encountered here could enable competitors to interrupt nutrient-cycling feedbacks, facilitating establishment and species turnover. We recommend that ecologists interested in testing for CWM and FD effects on ecosystem processes carefully consider the hazards of aggregation bias. Marginal FD effects have potentially been overstated.