How best to predict the dynamics of species-rich ecological communities has been a long-standing goal for both applied and basic ecology.  Effective ecosystem-based management necessitates knowing the limitations of our predictive accuracy.  To quantify these limits, a NCEAS "predator effectiveness" working group I have investigated which approaches to measuring species interaction strengths are most successful at making predictions regarding how species will response to disturbances elsewhere in their community.  My work has focused in particular on the effects of network complexity and interaction strength estimation error on the utility of qualitative and quantitative matrix methods: Loop Analysis and the Community matrix.

On the rocky shores of the Gulf of Maine, the American lobster, Homarus americanus, the Jonah crab, Cancer borealis, the Rock crab, Cancer irroratus, and the Green crab, Carcinus maenas, compose a guild of highly mobile predators.  Although the species are potential competitors that consume the same prey and utilize the same shelters, lobsters also prey on crabs (i.e. lobsters are intraguild predators of crabs).  During daytime low tides, crabs are also preyed upon by Larus spp. gulls. In previous work, I investigated the importance of avian and intraguild predation in influencing the diel (day/night) and spatial (depth) patterns of decapod activity in the low intertidal and subtidal zones of the Isles of Shoals archipelago.  This work has suggested that the modern overfishing of coastal fishes has increased the importance of intraguild interactions between lobsters and crabs, which in turn has caused crabs to become active during the day (a novel behaviour to the Cancer genus).  Having thereby become more susceptible to predation by gulls, crabs now contribute a stronger link between marine and terrestrial ecosystems.

The study of predator foraging behaviours such as prey choice, relative prey preferences, and the manner in which predator feeding rates respond to changes in prey abundance (i.e. functional responses) has long been a mainstay of modern ecology.  The nonlinear nature of trophic interactions that such behaviours introduce has important implications for the dynamics of populations and the structure and stability of food webs.  Populations of specialist predators, for example, fluctuate more than do those of generalist predators, but the mechanisms promoting the stability of generalist predator-prey dynamics have rarely been investigated.  The saturating functional responses which most predators exhibit, for example, destabilize predator-prey dynamics in theory, begging the question of how whole food webs persist in nature.

    Using my data on New Zealand’s whelk food webs, I addressed this question by determining (i) to what degree the feeding rates of whelks are saturated with respect to the density of their prey, (ii) the extent to which prey-attributes can be used to predict prey-specific contributions to the nonlinearity of a predator’s functional response, and (iii) how a predator’s diet richness affects the degree to which its overall feeding rate is saturated.  By using my empirical data to parameterize an extension of the classic Rosenzweig-MacArthur model of predator-prey interactions, I asked whether the degree of saturation observed within New Zealand’s whelk populations is nonlinear enough to affect the stability of their predator-prey interactions.  Results indicate that (i) whelk feeding rates are not strongly saturated, (ii) that most prey species contribute little to their predator’s saturation, and (iii) that increasing diet richness has a non-additive effect on a predator’s saturation such that the addition of alternative prey has a stabilizing effect on predator-prey dynamics.  This work has thereby offered a new mechanism by which generalist predators may stabilize the dynamics of their species-rich food webs, and an explanation for why predator-removal experiments typically result in linear responses in prey populations despite the inherent nonlinearity of trophic interactions.

Trophic omnivores – species that feed at multiple trophic levels – are central to our understanding of the structure, dynamics, and functioning of food webs. Many analyses have now shown that omnivores are ubiquitous and often over-represented in ecological communities. Their presence in food webs complicates the predictive power of trophic cascades and undermines the utility of the trophic level concept itself.  This is particularly true when omnivores engage in intraguild predation (IGP) by feeding on a second consumer species with whom they also share a  prey. A now well-developed theory of IGP systems offers interesting predictions regarding the mechanisms governing species coexistence in omnivorous food webs and how species abundance patterns should change across gradients of system productivity.  However, although the IGP module is perhaps the best studied of all food web modules, its applicability to real, species-rich food webs remains largely unknown.

  I tested two key predictions of IGP theory by investigating species abundance patterns and the structure and interactions strengths of a series of species-rich omnivorous whelk food webs situated along a strong gradient of productivity present around New Zealand’s coastline. I found that the intermediate predator (Haustrum scobina) is the superior competitor for shared prey species, as predicted by IGP theory.  Counter to theory, however, I showed that it is the omnivore (Haustrum haustorium) that is the superior competitor when both shared and unshared prey are considered.  In further contrast to theory, I documented an increase in the abundance of the intermediate predator with increasing productivity. My data nevertheless reveal clear and remarkably regular cross-gradient shifts in the food web structure and strengths of species interactions and suggest that adaptive and optimal foraging behaviour, and interactions among basal prey species, may play an important role in structuring communities.  These empirical insights offer hope that future modeling efforts which incorporate such processes will lead to a theory that can predict the emergent properties of natural food webs.

Efforts to estimate the strength of species interactions in species-rich, reticulate food webs are hampered by the multitude of direct and indirect interactions such systems exhibit. They have also been limited by an assumption that pairwise interactions display linear functional forms. Empirical estimates of interaction strengths are nevertheless important for characterizing keystone species and parameterizing mathematical food web models. I am interested in developing methods that avoid these problems to further our understanding of the processes regulating communities.  For example, Tim Wootton and I have developed a logistically feasible observational method that avoids the indeterminacy of indirect effects and estimates the nonlinear per capita strength of trophic interactions in species-rich food webs. I have subsequently confirmed the empirical accuracy of this method by comparing its species-specific attack rate estimates with those obtained by independent experimental manipulations (i.e. caging experiments) of a New Zealand intertidal whelk, Haustrum scobina.

Not only does the rate at which species are going extinct make modern times the sixth mass extinction of the earth’s history, but the rate at which alien species are invading natural systems is also unprecedented.  There is growing appreciation that extinctions and invasions can have profound effects on the functioning of ecosystems and the services they provide, and that the order of the community (dis)assembly process is important.  Jon Moore and I have investigated ways to bridge random-assembly and single-species focused approaches to biodiversity-ecosystem function research by developing methods for discerning how community assembly and disassembly processes jointly contribute to the community structure of Californian stream systems. At ecological time-scales, species do not simply blink out of existence or become important players overnight.  Rather, their populations decrease or increase dynamically, and are affected not only by external drivers but by the interactions and feedbacks occurring between species themselves.  Incorporating these processes is likely to be important for forecasting the future ecosystem-scale effects of species loss and turnover.