Interacting coexistence mechanisms in annual plant communities: Frequency-dependent predation and the storage effect

Abstract

We study frequency-dependent seed predation (FDP) in a model of competing annual plant species in a variable environment. The combination of a variable environment and competition leads to the storage-effect coexistence mechanism (SE), which is a leading hypothesis for coexistence of desert annual plants. However, seed predation in such systems demands attention to coexistence mechanisms associated with predation. FDP is one such mechanism, which promotes coexistence by shifting predation to more abundant plant species, facilitating the recovery of species perturbed to low density. When present together, FDP and SE interact, undermining each other’s effects. Predation weakens competition, and therefore weakens mechanisms associated with competition: here SE. However, the direct effect of FDP in promoting coexistence can compensate or more than compensate for this weakening of SE. On the other hand, the environmental variation necessary for SE weakens FDP. With high survival of dormant seeds, SE can be strong enough to compensate, or overcompensate, for the decline in FDP, provided predation is not too strong. Although FDP and SE may simultaneously contribute to coexistence, their combined effect is less than the sum of their separate effects, and is often less than the effect of the stronger mechanism when present alone.

Introduction

The complexity of ecological systems suggests that few ecological phenomena have a single cause. Species coexistence is no exception. Although most theoretical studies tend to focus on a single cause, species coexistence is likely caused by multiple mechanisms in nature. Moreover, these mechanisms are quite unlikely to be additive when present in combination. To understand non-additive (i.e. interactive) situations, multiple mechanisms must be studied in combination. We consider here interacting effects of coexistence mechanisms associated with predation, competition, and environmental fluctuations.

Recent work has drawn attention to similarities in the way predation and competition affect species coexistence (Chesson and Kuang, 2008, Krivan, 2003, Kuang and Chesson, 2008, Kuang and Chesson, 2009, Leibold, 1996) bolstering earlier work (Holt, 1984, Kotler and Holt, 1989). These developments led to a perspective where coexistence and exclusion mechanisms can both be classified as competition based or predation based. Following long-standing ideas in ecology (MacArthur and Levins, 1967), a competition-based coexistence mechanism works by intensifying intraspecific competition relative to interspecific competition, promoting coexistence (Chesson and Kuang, 2008). Resource partitioning provides the classic example of a competition-based mechanism. In a similar way, predator partitioning, where predators are relatively specialized to prey on different prey species, leads to a predation-based coexistence mechanism, because it intensifies intraspecific apparent competition relative to interspecific apparent competition between prey individuals, again promoting species coexistence (Chesson and Kuang, 2008).

Along with this mechanism classification comes a new understanding of how competition-based and predation-based mechanisms can interact with each other (Chesson and Kuang, 2008). The strength of a competition-based coexistence mechanism, in terms of how strongly it promotes species coexistence, is proportional to the overall strength of competition, as discussed in Chesson and Kuang (2008). Thus, if the addition of predation weakens competition, it will undermine competition-based mechanisms. Likewise, the strength of a predation-based mechanism is proportional to the strength of apparent competition, which can be weakened with the addition of competition.

Chesson and Kuang (2008) studied mechanism classification and mechanism interactions in generic Lotka–Volterra models of resource partitioning and predator partitioning. Lotka–Volterra models are useful for establishing general tendencies, but their linear per capita growth rates restrict application in nature. Nonlinear models, and models with environmental variation, may introduce new phenomena modifying the predictions of Lotka–Volterra models. Thus, for understanding these issues, it is essential to study a variety of models applying to various important situations in nature. One system that has been much studied empirically and theoretically is the desert annual plant system (Chesson et al., 2004). For that system, we address here interactions between predation-based and competition-based mechanisms in a model that deviates in several ways from the Lotka–Volterra model of Chesson and Kuang (2008). First, in this model, environmental variation drives the competition-based mechanism, the storage effect. Second, here there is just a single predator, and so predator partitioning by prey is not possible. Instead, the predator has frequency-dependent behavior, which introduces major nonlinearities in the model. These nonlinearities introduce the potential for a single predator to act in a manner that is similar to the action of several predators that focus on different prey species, as discussed below.

This model has application to arid environments where high diversities of annual plant species often occur in areas of sparse perennial plant cover. Seed predation is an important phenomenon in desert annual communities (Davidson et al., 1985). However, the storage effect has been a favored hypothesis for desert annual plants (Chesson et al., 2004), as discussed below. In these systems, predation-based and competition-based mechanisms have been investigated separately as species coexistence mechanisms (Chesson et al., 2004, Davidson et al., 1985). In a recent publication, we showed how the addition of predation to a model of competing annual plant species undermines coexistence by the storage effect (Kuang and Chesson, 2009). However, in that study, there was a predator, but no predation-based coexistence mechanism. By reducing the strength of competition, predation reduced the strength of competition-based mechanisms, but there was no opportunity for competition-based coexistence to be replaced by predation-based coexistence. Thus, we could not address the interaction between competition-based and predation-based coexistence mechanisms in annual plant communities. We do that here by considering predators with frequency-dependent behavior.

Frequency-dependent predation (FDP), or “switching” (Murdoch et al., 1975), is a powerful coexistence mechanism. FDP promotes coexistence by shifting predation to the more abundant species, facilitating recovery of other species from low density. In effect intraspecific apparent competition is strengthened relative to interspecific apparent competition by frequency-dependent behavior, which automatically gives a species at low density an advantage. A similar outcome is promoted with multiple predators by predator partitioning in the Lotka–Volterra models of Chesson and Kuang (2008).

FDP can occur for many reasons. We divide them into two classes, which have different effects on prey populations: constraints on behavior, and optimal behavior. Constraints on behavior include the difficulties of learning about infrequently encountered prey, leading to poor search image development, poor foraging technique, and poor knowledge of prey location (Hughes and Croy, 1993, Murdoch et al., 1975, Persson, 1985, Warburton and Thomson, 2006). All these factors potentially result in reduced predation on any prey species if its relative abundance declines. With optimal diet selection (Charnov, 1976, MacArthur and Pianka, 1966), the predator seeks the most nutritious prey per unit of predation effort, only seeking less nutritious prey when the most nutritious prey are hard to find. Optimal diet selection leads to an asymmetrical form of FDP because less nutritious prey are never favored over more nutritious prey, regardless of their frequency. Optimal patch selection, where the predator remains longer in denser patches (Garb et al., 2000), combined with spatial segregation of prey species, also leads to FDP (Bonsall and Hassell, 1999, Marini and Weale, 1997, Murdoch, 1977, Murdoch et al., 1975), but in this case there is no necessary asymmetry between species. In this paper, for simplicity, we focus on symmetric FDP that would result from foraging constraints.

We consider the competition-based coexistence mechanism, the storage effect (SE), because it has been a prominent hypothesis for the coexistence of desert annual plants (Chesson et al., 2004, Chesson and Huntly, 1989, Pake and Venable, 1995). With SE, temporal environmental fluctuations promote coexistence when different species have different responses to those fluctuating conditions, leading to a form of temporal resource partitioning (Chesson et al., 2004). Specifically, when a species drops to low density its increase is not opposed by intraspecific competition, and because it may experience a favorable physical environment when its competitors do not, it may increase at such times unopposed by interspecific competition also. A seed bank, or other persistent stages in the life cycle, prevent strong negative growth at other times, leading overall to recovery from low density, which promotes species coexistence. This is the storage effect.

Before SE became a popular explanation of coexistence in annual communities, predation was a favored hypothesis (Davidson et al., 1985). However, the recent understanding of competition–predation interactions, discussed above, suggests that it is time to revisit this hypothesis and how it interacts with the storage effect. This work provides the theoretical underpinnings of these questions. Below we present our model of annual plant dynamics incorporating competition, predation and environmental fluctuations. We analyze our model using techniques that provide quantitative measures of the strengths of the coexistence mechanisms (Chesson, 1994). We use these measures to study changes in mechanism strength as system properties change. By this means, we explain mechanism interactions, and the sizes and shapes of coexistence regions in parameter space. We then consider the implications of these results for further theoretical and empirical work on the coexistence of desert annual plants.