Alternative stable states in communities with intraguild predation

Abstract

Intraguild predation (IGP) is a widespread ecological phenomenon in which two consumers that share a resource also engage in a predator–prey interaction. Theory on IGP predicts the occurrence of alternative stable states, but empirical evidence of such states is scarce. This raises the question of whether alternative states are a rare phenomenon that is unlikely to be observed in nature. Here we analyze a model in which the resource exhibits logistic or chemostat dynamics and consumers have saturating (Type II) functional responses. We show that alternative states can arise for a wide range of biological scenarios and that environmental constraints can make their detection difficult. Our analysis identifies three possible combinations of alternative states: (i) IG prey or IG predator, (ii) coexistence or IG predator, and (iii) coexistence or IG prey. Bifurcation diagrams reveal that alternative states are possible over large regions of the parameter space. However, they can be limited to narrow ranges along the resource productivity axis, which may make it difficult to observe the occurrence of alternative states in communities with IGP. Microcosm experiments provide a promising avenue for detecting combinations of asymptotically stable states along a productivity gradient.

Introduction

Trophic levels in nature are not always clearly defined: two consumers can be competitors for a common resource and, at the same time, interact as predator and prey. This phenomenon is referred to as intraguild predation (IGP) because consumer species prey on their own “ecological guild”. Intraguild predation is a common occurrence in natural communities (Polis et al., 1989, Rosenheim et al., 1995, Arim and Marquet, 2004). In communities where two consumers compete for a shared resource, consumer coexistence is possible only if the IG prey is a superior resource competitor compared to the IG predator, because the IG predator has the advantage of being able to prey on its competitor (Polis et al., 1989, Holt and Polis, 1997). The competition—IGP trade-off is expressed at intermediate levels of resource productivity, when resource limitation favors the superior competitor but is not so strong as to exclude the IG predator. In highly productive environments, theory predicts that the IG predator excludes the prey (Holt and Polis, 1997, Diehl and Feissel, 2000, Mylius et al., 2001), though this is not always observed in nature (Mylius et al., 2001, Borer et al., 2003, Amarasekare, 2003, Amarasekare, 2007).

Alternative stable states arise when, for a given set of parameters, the community can exhibit multiple attractors. Analyses of the three-species IGP food web module have revealed that alternative states are possible, at least in theory (Holt and Polis, 1997, McCann and Hastings, 1997, Diehl and Feissel, 2000, Mylius et al., 2001, Revilla, 2002, Ruggieri and Schreiber, 2005, Takimoto et al., 2007). If alternative states exist, the long-term community composition is determined by the initial conditions; additionally, communities can jump between alternative states in response to perturbations (either rapid transient fluctuations in population size or gradual sustained changes in environmental conditions). Such abrupt transitions could affect the entire food web and have consequences for both community diversity and ecosystem stability.

Empirical evidence of alternative states in communities with IGP is scarce, largely because of the difficulty to detect them in nature. Laboratory experiments provide the best setting for testing theoretical predictions (Morin, 1999, Diehl and Feissel, 2001), although factors such as demographic stochasticity can cause discrepancies between theory and data even under controlled conditions (Montserrat et al., 2008). In a microcosm experiment of an IGP community, Morin (1999) verified that coexistence between IG predator and IG prey occurs, as predicted by theory, at intermediate levels of resource productivity; however a previous study with the same system (Lawler and Morin, 1993) found that the IG prey was excluded at lower levels of productivity. Montserrat et al. (2008) observed two different outcomes in four replicates of the same laboratory experiment on mites competing for pollen: when resources were abundant, the IG predator could exclude the IG prey (three replicates) or the IG prey excluded the IG predator (one replicate). Although many factors could explain these unexpected results, the possibility of alternative states cannot be ruled out. These two empirical examples underscore the importance of theoretical investigations that improve our ability to predict and diagnose alternative states (or, at the least, to be able to dismiss their possibility) in natural communities.

Most theory on IGP is based on population models that consider logistic or chemostat dynamics for the resource, and linear (Type I) or saturating (Type II) functional responses for the predation terms. These models exhibit three basic types of alternative states. First, IGP models with logistic resource dynamics and linear functional responses exhibit alternative states where either the IG prey or IG predator populations are stable but coexistence is not (Holt and Polis, 1997, Diehl and Feissel, 2000, Revilla, 2002). Second, IGP models with implicit resource dynamics (e.g., the Schoener model with IGP) exhibit alternative states with coexistence of IG prey and IG predator or IG predator only (Holt and Polis, 1997, Ruggieri and Schreiber, 2005). (Note that the resource function used in Ruggieri and Schreiber, 2005 corresponds to a quasi steady-state approximation to chemostat dynamics and linear functional responses.) These two types of alternative states also occur in models with chemostat resource dynamics, whether functional responses are linear (Takimoto et al., 2007) or saturating (Mylius et al., 2001). Third, IGP models with logistic resource dynamics and Type II functional responses in the consumers exhibit alternative states with IG prey only or coexistence of IG prey and IG predator. This has been previously reported by McCann and Hastings (1997) in the form of the stable resource—prey limit cycle occurring at the same time as a coexistence (chaotic) attractor. This combination also occurs in a 4-species version of the chemostat model with nonlinear functional responses (Kuijper et al., 2003) in the form of two stable equilibria. With the exception of the study by Kuijper et al. (2003), the alternative states scenario with IG prey only or coexistence has been largely overlooked in the analysis of alternative states.

While previous studies on alternative states have greatly enhanced our understanding of the dynamics of communities with IGP, the fact that different models incorporate different biological scenarios has made it difficult to derive predictions that can be empirically verified. This paper aims at developing a general conceptual framework for conducting a comparative analysis of alternative states in IGP models. We compare models with different types of resource dynamics (chemostat vs. logistic) and consumer functional responses (linear vs. saturating). Following Takimoto et al. (2007), we analyze the sequence of bifurcations along a resource productivity gradient that lead to the different combinations of alternative states. We derive analytical conditions for the existence of alternative states under different types of resource dynamics and consumer functional responses. We thus extend the work of Takimoto et al. (2007), which was limited to chemostat resource dynamics and linear functional responses. A novel feature of our work is the comparative framework, which allows us to identify the types of communities in which alternative states are most likely to be observed. Our results apply to a wide variety of communities, including insects host–parasitoid interactions, omnivorous bacteria, and fish and mammalian communities with IGP.