Introducing resource-use complementarity
While multi-trophic interactions were intrinsic to our food-web model, incorporating resource-use complementarity required a modification of the producer-resource interaction. We introduced it based on two simple assumptions: First, resource-use complementarity can only occur if species differ in their access to resources, forming different resource compartments, for example, due to differences in chemical forms of resources used or their spatial distribution (e.g., access to different soil layers). Second, we assumed that resource-use complementarity is maximized if all species take up resources from distinct resource compartments.
We therefore introduced differences between producer species by limiting their resource-use to certain compartments of the resource pool to simulate resource-use complementarity (Fig. 1). Species that access the same compartments directly compete for the resources within those compartments. To investigate resource-use scenarios where all species utilize resources from different compartments (i.e., no competition), the number of resource compartments C was defined as the maximum of producer species richness considered in our design (i.e., 16). Further, we assumed that all compartments were quantitatively the same. By increasing the dissimilarity between resource-use strategies of the 16 producer species within a species pool, we created a gradual change from no complementarity (i.e., all species access all compartments) to maximum complementarity (i.e., each species has its own resource compartment; Fig.1). For this gradient of resource-use dissimilarity (RUD), we ensured that (1) all producer species had access to the same number of compartments at a given level of RUD and that (2) accessed resource compartments were the same for both resources considered.
At maximum producer richness, species within a community where RUD < 1 initially always compete for resources with at least two other producer species with overlapping compartments, up to having all species competing with each other when RUD = 0. The competitive outcome is determined by which species can lower the resources the most (’R*-rule’, Tilman 1982), whether resource competition can be weakened by trophic processes (Brose 2008), or both. To capture how the resource-use and thus productivity Y was distributed among coexisting producer species i, we calculated Shannon diversity Hexpas Hexp = exp( - ∑i pi ln(pi)), with pi = Yi / ∑i Yi. Hexp reflects aspects of richness (i.e., how many species coexist) and abundance (i.e., how much resources each species uses) alike and is maximized at the number of coexisting species if all species use resources evenly (Jost 2006). Lower values indicate an uneven distribution of resource-use. In comparison to RUD, Hexp is based on realized instead of fundamental resource niches.