Discussion

This meta-analysis explicitly revealed that both functional trait variability and identity influenced the effects of tree mixtures on forest productivity, both across and within species richness levels of tree mixtures in experimental tree communities at a global scale. Specifically, we found that the functional dispersion of tree mixtures increased the extent of the positive mixture effects on productivity overall, and within the two- and four-species mixtures. Moreover, the CWM of acquisitive traits of species mixtures enhanced the positive effects of mixtures on forest productivity. Our findings offer novel insights into the importance of plant functional traits in determining the magnitude (and even directionality) of the biodiversity-productivity relationships that have been under debate for more than two decades.
To date, a few experimental studies have segregated functional aspects from species richness to test the effects of functional diversity or individual traits involved in tree productivity (Tobner et al. 2016; Grossman et al. 2017). The previous studies
emphasized that a particular combination of functional attributes (e.g., deciduous and shade-intolerant species, high leaf-nitrogen, and calcium) (Tobner et al. 2016; Grossman et al. 2017; Huang et al. 2018), or shade tolerance heterogeneity between constituent species (Zhang et al. 2012), caused the observed species diversity effect. However, these researches also demonstrated that functional diversity per se could not explain the additional variation in ecosystem productivity across communities at a given species richness level. Beyond such conventional wisdom, our meta-analysis revealed that both functional dispersion and identity in tree mixtures determined the extent of diversity effects on productivity across and within the species richness levels. Our results, therefore, provided evidence that increased functional diversity should enhance ecosystem functioning through the coincidental dominance of influential species, or through niche partitioning (Tilman et al. 1997; Diaz & Cabido 2001; Loreau et al. 2001).
The degree of functional differences between species drive the effects of plant mixtures on ecosystem productivity, due to niche partitioning and positive interactions between constituent species at the community level (Tilman et al. 1997; Diaz & Cabido 2001; Loreau & Hector 2001). Plant leaf and wood economics traits are associated with plant resource acquisition, shade tolerance, hydraulic transport, mechanical support, and carbon storage (Reich 2014). Communities consisting of species with contrasting leaf and wood economics traits caused niche differentiation with respect to the utilization of light and water and facilitative interactions (Fichtner et al. 2017; Baez & Homeier 2018), which might increase the community-level acquisition and efficient use of light and water (Anderegg et al. 2018; Huang et al. 2018). Species-diverse mixtures with higher FDis, in turn, enhances the efficcacy of resource use in mixtures due to recourse niche differentiation (Tilman et al. 1997; Cardinale et al. 2011), thereby improving ecosystem productivity (Flynn et al. 2011). Moreover, we found that increasing FDis resulted in higher productivity, even at a given species richness level. These outcomes suggested that trees in the communities with the same number of species that occupied various positions in the leaf and wood economics spectrum, increased the efficacy of resource utilization, and tended to promote forest productivity.
Consistent with our hypothesis, we found that the higher CWM of LNC increased the positive effects of species mixtures on productivity, which indicated the important role of the N-acquisitive strategies involved in diversity effects for improving forest productivity (Fig. 4). The functional characteristics of plant species determined the interactions between constituent species in plant communities along abiotic stress gradients (Maestre et al. 2009). In this study, mixtures dominated by acquisitive species were found at experimental sites with warmer climates and higher precipitation (Fig. S3). Additional available resources allowed for intense species interaction caused by effective light acquisition of fast-growing species, and hence niche differentiation in communities (Bertness & Callaway 1994; Callaway et al. 2002). These interactive processes appeared to be more intense in mixtures that included acquisitive plants compared with conservative mixtures, which consequently improved the effects of mixtures on productivity (Tobner et al. 2016; Fichtner et al. 2017). It is noted that the impacts of species mixtures on productivity were enhanced with the CWM of leaf nitrogen but independent of wood density (Sakschewski et al. 2015). Wood density correlates to a large number of structural characteristics of wood plants (Chave et al. 2009), and species with high wood density generally represent the conservative-end of the fast-slow plant economics spectrum (Reich 2014). Communities being characterized by great CWM of wood density reflect the coincidental dominance of slow growing species for maintaining ecosystem productivity. In such case, the interactive processes should be weak in mixtures that dominated by slow growing plants, which consequently cannot enhance the effects of mixtures on productivity.
In conclusion, our meta-analysis integrated the functional differences of species between global scale tree diversity experiments and investigated how functional trait variability and identity determined the outcomes of tree mixture effects on ecosystem productivity. Our results revealed that the effects of tree mixtures on productivity increased with the functional dissimilarity of the leaf and wood economics traits, and the community-weighted mean of leaf nitrogen content overall and within the two- and four- species mixtures. These results revealed the key role of the functional dispersion and composition of species mixtures toward explaining the variations in the effects of plant mixtures on ecosystem productivity, both across and within the species richness levels. We anticipate that our analysis will stimulate future inquiries into the role of functional traits in the diversity-productivity relationships.