Introduction
A large body of research in the last several decades has investigated potential factors that can promote the structural and dynamical stability of complex food webs and their constituent populations. These factors include hierarchically ordered feeding ), characteristic predator-prey body mass ratios (Brose et al. 2006), allometric degree distributions of feeding links (Otto et al. 2007), compartmentalization (Stouffer & Bascompte 2011), weak interaction links including weak omnivory and reduced predation pressure at low densities
(, pairwise negative correlation between interaction strengths , and self-regulation (e.g., cannibalism, intraspecific interference, ), among others (reviewed by ). More recently, studies have started incorporating different types of interactions in complex food webs (multiplex or multi-layer networks; ) to account for multiple types of ecological interactions, such as mutualisms and parasitism, in which organisms simultaneously engage in natural communities.
A ubiquitous feature of natural systems is that almost all organisms grow in size during their lifetime and switch diets, trophic positions, species interacting with, and habitats as they grow . Such ontogenetic development introduces life-history stages and flows of biomass between the stages through growth and reproduction to food webs, collectively forming complex multi-layer ecological networks. Studies have shown that ontogenetic diet shifts have far-reaching effects on competitive and predator interactions, population dynamics, and community structure in small food web modules . The persistence of consumers can be enhanced in life-history structured communities through biomass overcompensation in consequence of ecological asymmetry between different stages (e.g., juveniles are better competitors than adults; ). Such asymmetry, however, can also be expected to destabilize populations by inducing cohort cycles or alternative stable states without a predator . Research on how these effects in small food web modules may scale up to an entire complex food web is still in its infancy, and so are the tools to generate life-history structured complex food webs in a biologically justifiable manner.
Studies have reported the mixed effects of including a stage structure on the stability of complex food webs . Rudolf & Lafferty (2010) found that, using static topological models of food webs, structural robustness to species removal was lower with a stage structure than without. They pointed out that species might be more sensitive to resource loss when ontogenetic stages were sequential resource specialists. Bland et al. (2019) used population dynamical models of complex food webs and showed that non-stage-structured food webs lost twice as many consumer taxa as stage-structured webs, while the variability of biomass dynamics did not differ. Mougi (2017) also used similar population dynamical models and concluded that species persistence (the fraction of species persisting in a food web) increased as the proportion of stage-structured species increased in the food webs and that the effect was more pronounced in food webs with a greater number of species and interactions. More studies are needed to elucidate the role of a stage structure on persistence and stability and how it may come about in complex food webs.
Rudolf & Lafferty (2010) and Bland et al. (2019) used the niche model (Williams & Martinez 2000) to generate network topologies and split a node into stages to create a stage-structured taxon (nodes represent taxa, and interacting taxa are connected by links in ecological networks). The niche model has a demonstrated ability to produce many observed structural properties of empirical food webs despite its simplicity and has been the most widely used food web structural model. Splitting a node, as in Rudolf & Lafferty (2010) and Bland et al. (2019), can nontrivially modify the food web topology generated by the niche model, likely compromising the desirable properties of the food webs. Therefore, it is unclear how realistic the modified food webs in these studies would still have remained after new nodes and links were added to incorporate a life-history structure. Firstly, minimizing the alteration of the network topology generated by the niche model is desirable because it is capable of producing realistic food web topology and because food web data resolved to life-history stages to verify the topology of food webs with stage-structured taxa are currently very scarce. Secondly, the niche model generates a “trophic species,” which is a functional group defined to consist of one or more taxa (e.g., species, genus, ontogenetic stages) that share the same sets of predators and prey . Life-history stages of a species are distinguished for their distinct ecological roles, at least partly by their characteristics related to feeding, so that a life-history stage can be considered as a whole trophic species (not a fraction of it). Based on these interpretations and the observation that ontogenetic diet shifts are widespread in nature (Werner & Gilliam 1984), a plausible alternative approach is to instead group nodes generated by the niche model to assemble a stage-structured taxon. This approach would allow preserving mostly the original topologies of the food webs from the niche model. No study has investigated this approach before.
We took the grouping approach to introduce a stage structure into complex food webs. Following Bland et al. (2019), we applied the allometric trophic network (ATN) model of biomass dynamics to the stage-structured food webs on which we linked stages by biomass flow via growth and reproduction. We motivated the food webs studied here from aquatic communities at temperate and northern latitudes. It is well established that consumer-resource interactions are hierarchically structured largely by body size in aquatic communities because of the indeterminate growth of fishes and gape-limited predation . We found that food webs with stage-structured consumers tended to be less dynamically variable and supported a greater number of species than food webs with non-stage-structured consumers.