Adaptive foragers and community ecology: linking individuals to communities and ecosystems

Abstract

Thirteen years ago, the journal Evolutionary Ecology ran a special issue commemorating 30 years of the influence of foraging theory on population and community ecology (Schmitz 1997). This era began with the classic foraging theories of MacArthur & Pianka (1966), Emlen (1966)Fretwell & Lucas (1969), and Charnov (1976a,b), and was followed by the seminal contributions of Pyke, Pulliam & Charnov (1977), Schoener (1971, 1973, 1976, 1983, 1987), Belovsky (1978) and Stephens & Krebs (1986). These contributions are noteworthy for their emphasis on making predictions about diet choice, habitat selection and territory size, as well as the degree of consumer specialization and generalism. Subsequently, a bout of empirical and theoretical work reinforced the notion that foraging behaviour, and more generally adaptive behaviour, is critical to understanding the structure of ecological communities and the dynamics of the populations within them (Schoener 1971, 1973, 1976, 1983, 1986, 1987, 1989; Holt 1983; Belovsky 1986a,b; Fryxell & Lundberg 1994). It is now well established that animals adaptively choose their diets, shift their habitat use and alter their morphology in response to temporal and spatial variation in resource availability, as well as the presence of predators and parasites. Optimal foraging remains the central theory linking resource and consumer traits to patterns of resource selection. Yet, even after 40 years of development, there are precious few advances towards truly synthesizing the connections between individuals, populations and large interconnected food webs (Schoener 1989; Schmitz 1997; Abrams 2008). Nevertheless, we suggest that food web biology will emerge in the next 10 years as a new focal point for understanding the interplay between adaptive behaviour, population dynamics and the complexity and structure of ecosystems. Food webs, the networks of feeding links between species, are central to our understanding of ecosystem structure, stability and function. Studies of food webs have informed us on issues ranging from how the loss of species may initiate cascades of extinctions to the types of structural regularities that may promote population and community stability. Moreover, the use of foraging rules and optimization in various large-scale models of food web structure and dynamics is growing again (Williams & Martinez 2000; Kondoh 2003, 2006, 2007; Loeuille & Loreau 2005; Beckerman, Petchey & Warren 2006; Brose, Williams & Martinez 2006b; Brose et al. 2006a; Loeuille et al. 2007; Petchey et al. 2008), perhaps inspired by the contributions in Evolutionary Ecology 13 years ago (Schmitz 1997). Despite this optimism, the typical empirical food web analysis today still falls into one of two categories, each with its own limitations. The first, the static category, aims to answer questions such as who eats whom, how many trophic links are there, and what are the patterns of trophic links? The second, the dynamics category, aims to address questions such as what happens when one species eats another species and, more generally, what are the effects of particular patterns of trophic interactions on population dynamics, stability and community structure (e.g. Gross et al. 2009). This static/dynamic dichotomy remains the norm and has contributed many insights into the structure and dynamics of real communities. However, through this Special Feature, we hope to reinforce that foraging behaviours, and the associated, broadly defined set of individual traits that make up behaviours of individuals, represent one of the most fundamental links between the static and dynamic approaches. To begin, we provide a general framework that draws together – and provides motivating conceptual links between – individual traits, adaptive behaviour, food-web structure, population dynamics, and community stability. Our rationale for proposing that adaptive foraging (traits/behaviour) and food webs must be linked is that foraging traits form the interface between selection pressures and population dynamics (Fig. 1). There is no linear way to present the dynamic feedback between biotic selection pressures on particular organisms, the traits of those organisms, the ensuing dynamics of both the traits and the populations, and the consequences for the complexity and structure of a community or ecosystem. Figure 1 does, however, present the key concept that phenotypes are both subject to natural selection and determine patterns of consumption, population growth and population dynamics. Food web properties rely on and are influenced by individual and species-specific traits. Traits are the phenotype that both undergo natural selection and underpin patterns of consumption, population growth and population dynamics. We highlight here a research programme that is enhancing links between selection, individuals’ traits and community structure, that sits at the confluence of ecological and evolutionary research. Resources ultimately drive patterns of physiology, morphology, behaviour, and life history, and thus generate strong and weak interactions among competitors and predators (Fig. 1, middle). Consumers generate comparable changes in prey/resource behaviour, morphology and life history that further affect interactions among the prey, competitors, and their own natural enemies (Fig. 1, middle). Here, we classify these changes in physiology, morphology, behaviour, and life history under the general concept of adaptive behaviour, where behaviour can be defined at multiple scales. These changes in resource and consumer traits simultaneously alter the pattern of who eats whom and the dynamics that result from the interactions (Fig. 1, middle to right). Evolution, driven via selection pressures, changes these traits. Yet is important to realize that the interactions per se and the dynamics that result from them are selection pressures (Fig. 1, left and right to middle). This simplified illustration of different levels of ecological organization (individual fitness, adaptively defined traits, species interactions, and community structure) underscores how and why traits related to foraging and eating are central to our understanding of food webs. Figure 2 expands this perspective to highlight various groups of traits that have played and continue to play an important role in food-web research. The left-hand column emphasizes the need for a broad definition of adaptive behaviour across three hierarchical scales – individuals, populations and small community modules and up to entire food-web networks and ecosystems. The techniques, theories and models for assessing the role of traits have been varied. Most techniques centre around integrating various concepts from optimal foraging theory, while some have leveraged generic optimization methods to explore the evolutionary ecology implications of embedding traits directly in dynamic models (e.g.Ives & Dobson 1987). A perspective on groups of traits that have and continue to play an important role in food-web research. The performance of individuals, the dynamics of populations, the structure of networks, and the dynamics of ecosystems are influenced by these various traits and the focus of large bodies of theoretical and empirical research. The research we highlight here suggests that enhancing this link between traits, theory and small and large networks will be very productive at providing a mechanistic explanation of stability, complexity and ecosystem structure. Perhaps most importantly, a few common channels of creativity are recognizable in the figure. Body size continues to feature heavily throughout research at different scales. It is represented even more often when one considers allometry as a specific focus on the ecological implications of body size (Schoener 1989; Cohen et al. 1993; Belovsky 1997; Jennings & Warr 2003; Berlow et al. 2004; Emmerson & Raffaelli 2004; Loeuille & Loreau 2005; Brose et al. 2006a; Petchey et al. 2008). Body size is further implicated in understanding the role of growth and ontogeny (e.g. life history) and in creating temporal, spatial and morphological refuges (e.g. species interaction traits) (Werner & Hall 1977; Sih 1982b, 1987; Werner 1992; Werner & Anholt 1993; Anholt & Werner 1995). These examples complement classical representations of habitat selection as a behaviour (Emlen 1966; MacArthur & Pianka 1966). Recent advances in understanding information quality and detection has stimulated productive lines of research into perception, signalling and learning as traits potentially influencing interactions and dynamics. However, all but a handful of this research stops at the foraging, habitat-selection and/or mate-choice scale, and does not extend to population dynamics and larger food webs. For example, there is a growing interest in how different types of information are accumulated (Valone 2006) and how this information affects habitat selection and the expression of defensive traits (Alonzo 2002; Luttbeg 2002; Stamps & Davis 2006; Turner, Turner & Lappi 2006). Further examples include the role and quality of alarm signals with respect to predation (Beauchamp & Ruxton 2007; Rowland et al. 2008; Chittka, Skorupski & Raine 2009; Higginson & Ruxton 2009). A notable exception is work by Luttbeg, which extends questions about information quality to dynamics games of more than two species (Luttbeg & Schmitz 2000; Luttbeg, Rowe & Mangel 2003). As Sih, Hanser & McHugh (2009) indicate in their recent perspective on linking social network theory to behavioural ecology, the links between network science, behavioural ecology and food webs can only grow. Interaction strengths, once a source of major debate (and perhaps still so, see Paine 1992; Laska & Wootton 1998; Abrams 2001a; Berlow et al. 2004) have emerged as a cornerstone of defining and measuring dynamics incorporating both species identity and more importantly, species traits (Paine 1992; Wootton 1997; Berlow et al. 1999, 2004; Ings et al. 2009). There is a long history of including various species traits in models of competition and predation (and disease), frequently inspired and motivated by optimal foraging theory (Schoener 1973, 1976; Sih 1981, 1982a; Holt 1983; Schoener 1983, 1986; Abrams 1987a, b; Schoener 1987, 1989; Sih 1987; Fryxell & Lundberg 1993, 1994). And not at all disconnected from this is the current challenge of partitioning direct and indirect effects, of varying strength, in food webs of varying size, to the categories of density or trait-mediated indirect effect (Schmitz, Beckerman & O’Brien 1997; Schmitz 1998; Peacor & Werner 2001; Schmitz, Krivan & Ovadia 2004; Bolnick & Preisser 2005; Preisser, Bolnick & Benard 2005; Abrams 2008). However, the large-scale food-web research programme has not embraced traits in the same way that research on smaller communities has. It should be no surprise to many people that, historically, the majority of this trait-inspired research on food webs and populations dynamics has been restricted to simple interactions between individuals within a single species or among fewer than four species, i.e. in small networks (see also Ives & Dobson 1987; Abrams 1996; Abrams & Matsuda 1996; Abrams 2000; Kokko & Ruxton 2000; Abrams 2001b). With a few exceptions, the focus of this research has been on understanding the role of traits in the stability and persistence of these small communities (but see Kondoh 2003; Loeuille & Loreau 2005). Even the briefest look at Fig. 2 confirms that only a handful of specific traits have been linked to large network/food-web questions (see below), despite the more general consideration of traits in small-scale food webs. One of the most encouraging patterns in the past 5 years is that two of the newest aspects of food web biology – binary-network analysis and the large-scale evolutionary–dynamic analysis – are increasingly making use of one of the oldest motivating theories in behavioural, population and community ecology: that of adaptive behaviour/optimal foraging. The cascade model (Cohen & Newman 1985), the niche model (Williams & Martinez 2000) and subsequent alterations of the niche model (Cattin et al. 2004) use probabilistic rules to encode established empirical and theoretical rules about foraging. The dynamic and evolutionary models of Kondoh (2003, 2006) and Loeuille and Loreau (2005) use simple rules of energy gain to add adaptive trait dynamics to explorations of stability and complexity in large communities. Our own work (Beckerman, Petchey & Warren 2006; Petchey et al. 2008) uses the contingency model of optimal foraging and allometry to predict both the complexity and structure of observed food webs. Where to now? With all of this history, do we really need a new vision? We think so. 1, 2 illustrate that a great deal of effort has been made to integrate traits into what we consider small-scale food webs. We advocate three over-arching objectives for the future. First, we must better synthesize and begin to capture generalities from across trait types among the small network studies (see Abrams 2009; Stouffer 2009). Second, we must further extend the use of traits and theories successful at providing insight at the small scale to the larger, network-size food webs (see Brose 2009; Kondoh 2009; Loeuille 2009 and Stouffer 2009). Finally, we must also recognize the opportunity to link advances from small-scale dynamics involving traits to the larger networks, understanding that the small, tractable communities are what network biologists know as modules or motifs in the larger networks (see Stouffer 2009). In essence, we expect that the details from the work at small scales will provide node-specific and biologically relevant information to advance the current phenomenological focus of networks (Proulx, Promislow & Phillips 2005; Bianconi, Pin & Marsili 2009; Sih, Hanser & McHugh 2009). There are currently precious few examples of these efforts. This Special Feature contains five papers that together support a vision for the future of food-web research centred on adaptive behaviour. They are authored by people who might be considered grand masters of the field as well as people just entering academia. Importantly, they provide perspective and insight while stimulating thought about the use and importance of traits across various scales. It is this emphasis that we champion and we hope it will lead towards a synthetic definition of food-web biology in which traits, dynamics and structure are linked together. The feature begins with a remarkable and notable piece of work by Peter Abrams (2009), who for the first time provides a review of his contribution to understanding the stability, dynamics and contingency of simple models of a few species. One of the core features of Abrams’ work over the past decades has been a reliance on what we all consider to be simple models and always clear assumptions. His attempts to include traits associated with foraging and trade-offs in the models have continued to develop our understanding of how traits matter a great deal and how dynamics, persistence and stability are contingent upon model assumptions. His overview provides insight into five over-arching themes and questions that drive all of his research. The article by Nicolas Loeuille (Loeuille 2009) provides another overview, this time from the perspective of large-scale population dynamics. Loeuille’s research with Michel Loreau is a fundamental contribution to understanding how evolution of foraging traits and constraint on species’ diets can shape the complexity, structure and dynamics of food webs (see also Kondoh 2003). His more recent forays into metacommunity ecology maintain the focus at a large scale, and his paper here helps draw together methods and insights that come from defining traits among many species at these large scales. The works by Uli Brose (Brose 2009) and by Michio Kondoh (Kondoh 2009) bring us directly to one of the most common traits used in food-web biology: size. Brose reviews the increasing body of empirical evidence about the allometry of foraging traits, such as handling times, maximum ingestion rates and attack rates, while highlighting some of the gaps. For example, there is relatively little information about how the shape of a consumer’s functional response (Type I, Type II, Type III, or somewhere in amongst these) is influenced by the size of prey and predator. Despite the need for more and better empirical information (as always!), Brose reviews models of how individual foraging behaviour, expressed both in prey to predator mass ratios and the distribution of trophic interactions, can resolve one of the longest standing puzzles in ecology: how very diverse and complex ecosystem can be dynamically stable. Although overall body size is of great importance for foraging behaviour, considerable variation around allometric relationships indicates the potential for other traits to have strong effects. Michio Kondoh presents an analysis investigating how cognitive capacity and the ability and rate at which organisms can adapt behaviourally (e.g. learning) can influence the relationship between predator and prey. In particular, Kondoh’s contribution documents an intriguing relationship that suggest predator–prey pairs (i.e. diet choice) are determined not only by overall organismal size, but also on relative brain sizes. While the underlying causes of these relationships remain open to speculation, he proposes that arms races in the cognitive ability of prey and predator are a possibility. Such arms races have the potential to strongly influence population dynamics and community structure. Finally, the work of Daniel Stouffer (Stouffer 2009) offers a critical review of many of the phenomenological and population-level models of food webs, with an eye towards asking two fundamental questions: do we need the population-level models and, if we do, do the small sets and patterns (motifs) of interactions among species (e.g. Abrams 2009) offer sufficient insight into patterns at the large scale captured more often by phenomenological models. Stouffer’s own research on motifs in food webs offers an intriguing link between small and large food-web networks, and this question of scale is critical to future empirical research and attempts to manage natural resources and communities. This collection of papers aims to build on the exciting recent progress in food-web biology by integrating research in behavioural ecology, allometry, population ecology and community ecology. The authors explore the extent to which the structure and dynamics of populations and communities can be understood by including aspects of individual behaviour. The papers report on foraging decisions, cognitive ability and body size in models of and experiments at small and large scales, across a breadth of different systems, and by observational, experimental and theoretical approaches. The papers will hopefully acquaint ecologists with the current state of knowledge regarding scaling adaptive behaviour to community and ecosystem patterns and motivate the next decade of research linking behaviour to food webs. The articles in this special feature arise from a Special Symposium at the 92nd Annual Meeting of the Ecological Society of America in San Jose CA. We thank Charles Fox, Frank Messina and Liz Baker for their insight and guidance at Functional Ecology.