Introduction
Most research on biodiversity-ecosystem functioning (BEF) relationships has focused on effects of varying horizontal diversity (i.e., diversity within a single trophic level), most commonly of plants in controlled experimental communities (e.g., Isbell et al. 2015). However, natural communities are characterized by complex interaction networks that integrate diversity and its effects across trophic levels (Duffy et al. 2007; Brose et al. 2019), with their BEF relationships varying substantially in strength (Barnes et al. 2014; Duffy et al. 2017; van der Plas 2019). Recent research has aimed at resolving this separation between within-trophic level and multi-trophic approaches to BEF relationships (Loreau 2010; Brose & Hillebrand 2016). For example, the vertical diversity hypothesis links ecosystem functions of primary producers, and hence their diversity effects, to variance in vertical diversity (i.e., diversity across trophic levels), specifically the maximum trophic levels and body-masses of multi-trophic ecosystems (Wang & Brose 2018). It also points to other related aspects such as food-web structure (Thompson et al. 2012; Montoya et al. 2015; Brose et al. 2017) or animal diversity (Naeem et al. 1994; Schneider et al. 2016; Zhao et al. 2019) that influence ecosystem functions at the producer trophic level. Despite ample evidence for such top-down effects on producer BEF relationships, the underlying mechanisms have remained elusive.
The many biological mechanisms involved in creating positive diversity effects in producer communities can be broadly categorized into two classes (Loreau & Hector 2001; Loreau 2010). First, complementarity mechanisms occur when functionally different species use dissimilar niches, hence have a low interspecific competition. This low competition fosters coexistence, which simultaneously increases the ecosystem functioning of the whole community. Second, selection mechanisms favor species with competitive advantages. If such advantages are associated with particular functional traits such as a higher growth rate, selection can affect ecosystem functioning. Complementarity and selection are both enhanced by a larger species-pool that may provide more complementary species and strong competitors alike (i.e., sampling effect). However, they have opposite implications for realized diversity, which is maintained by complementarity but reduced by selection mechanisms. Even though the functional identity of the dominating species can be important depending on the ecosystem function in question (Loreau 2004; Hooper et al. 2005), most evidence points towards complementarity mechanisms as the dominant driver of BEF relationships (Hooper et al. 2005; Cardinale et al. 2007; Barry et al. 2018).
Complementarity between co-occurring producer species is most commonly associated with resource-use complementarity (synonymous with resource partitioning; Barry et al. 2018), expressing fundamental differences in resource-access of coexisting species. These differences can arise from varying aspects of resource-use such as differences in the chemical forms of resources used (McKane et al. 2002; Von Felten et al. 2009; Ashton et al. 2010), phenological asynchrony (Henry et al. 2001; Sapijanskas et al. 2014), or spatial separation, both above- (e.g., crown packing in Sapijanskas et al. 2014) and belowground (e.g., rooting depth in Mueller et al. 2013). Additional resource-based mechanisms such as facilitation (Wright et al. 2017) and niche plasticity (Von Felten et al. 2009; Mueller et al. 2013) can modify resource niches to decrease competition and increase complementarity among producers further.
In the presence of animal consumers, however, competition is not only resource-based (exploitative competition) but can be mediated by multi-trophic interactions (apparent competition; Holt 1977; Loreau 2010). When herbivorous feeding is complementary (i.e. herbivores have different resource-species), apparent competition between producer species is low, which fosters coexistence as it creates complementarity at the producer trophic level (Thébault & Loreau 2003; Brose 2008; Poisot et al. 2013; Wang & Brose 2018). As a result, simple herbivore communities alone are sufficient to create positive diversity effects on standing biomass and resource uptake (i.e., primary production) of producers, even without resource-use complementarity (Thébault & Loreau 2003). Increasing the vertical diversity in complex trophic networks can further enhance coexistence, indicating that complementarity scales with the diversity of the multi-trophic animal community (Wang & Brose 2018). Additionally, herbivorous feeding can amplify differences in the competitive abilities of some producer species and thereby introduce selection mechanisms that can affect ecosystem functioning both positively or negatively (Thébault & Loreau 2003). For example, large producers that have low mass-specific metabolic rates are more suited to cope with herbivory, thus are more competitive and maintain higher biomasses (Schneider et al. 2016). Investigating complementarity mechanisms without considering selection is therefore impossible when trying to understand what drives BEF relationships in multi-trophic ecosystems.
It is evident that resource-use complementarity and multi-trophic interactions can both shape BEF relationships at the producer trophic level. Complementarity from either source will favor a positive relationship between biodiversity and ecosystem functioning, while selection may interact in more complex ways, potentially having opposing effects. It is therefore important to investigate how these mechanisms most likely combine in realistic complex food-webs. Our study addresses this issue by integrating multi-trophic interactions and resource-use complementarity into a complex allometric food-web model to examine how they create and shape the positive effects of producer species richness on primary production (hereafter: net diversity effects). First, we investigate how resource-use complementarity amongst producers creates positive net diversity effects across levels of producer richness. The subsequent inclusion of multi-trophic interactions allows us to investigate how such effects are modified through changes to the producer community’s functional composition, potentially driving both selection and complementarity mechanisms. By varying the species richness of the multi-trophic animal community, we investigate how diversity across trophic levels influences the mechanistic interaction with resource-use complementarity and thus determines net diversity effects.