A closer look at our research on foraging |
|
The importance of foraging |
|
All species eat. Consequently, foraging plays a central role in ecological interactions and in the success and failure of most species. Many of the topics that I investigate deal with foraging in some way. As a result, at the same time that my research has investigated aspects of biodiversity (e.g., species invasion and extinction), it has also provided insights into fundamental aspects of foraging. |
|
Effects of filter feeding burrowing shrimp on estuarine phytoplankton abundance |
|
Upogebia pugettensis is an abundant burrowing shrimp that forms extensive subterranean burrows on intertidal mudflats of bays and estuaries throughout the Pacific Northwest. Burrowing and suspension-feeding activities of this ecosystem engineer can have substantial ecosystem impacts such as
altering benthic community composition, altering
benthic structure through bioturbation, influencing nitrogen cycling, organic matter remineralization, solute
fluxes across the sediment–water interface, and filtering large portions of suspended particles from the water column. Oyster growers have previously contended that the shrimp compete with oysters for phytoplankton food. Griffen et al. 2004 sought evidence for this claim by measuring the size-specific functional response of the shrimp and using these measurements to calculate the proportion of the water column in Yaquina Bay, OR that is filtered daily by the shrimp. |
|
 |
 |
 |
Shrimp are not the only sink for phytoplankton as water flows over shrimp beds. Holes in the substrate (pictured above) demonstrate burrow openings from the extensive burrow beds that are common on intertidal flats. The picture to the left shows a fiberglass cast of a burrow that has been constructed. Attached to the cast are small bivalves (Cryptomya californica) that live commensally inside the burrow wall and extend their siphons into the burrow, filtering phytoplankton from the water that is pumped through the burrow by the shrimp. Thus, phytoplankton may be removed from suspension by directing consumption by shrimp, consumption by the commensal clams, and adhesion to the mucus-lined burrow wall. These three components together make up the shrimp-burrow complex. |
In addition to measuring the size-specific functional responses of shrimp in laboratory experiments, and size-specific functional responses of commensal clams, I also determined the amount of phytoplankton that is removed by the burrow wall with an experiment in which the length of the burrow was constrained by altering the volume of sediment that shrimp were provided (burrow casts from this experiment shown on right). By combining these three components I was able to determine the relative contribution of each to the total phytoplankton removal by the shrimp-burrow complex. |
|
I |
I then conducted a field experiment in which I measured the functional response of the entire shrimp-burrow complex together using benthic chambers placed in areas where shrimp were present at different densities (figure on left). Results of this field experiment were used to calibrate calculations of filtration by the aggregate shrimp-burrow complex as determined by laboratory experiments. These calibrated values were then scaled up to the entire area of the Yaquina Bay estuary to determine the proportion of the total volume of the estuary that is filtered daily by the shrimp-burrow complexes. The figure on the lower left from Griffen et al. 2004 demonstrates that the shrimp-burrow complex is capable of filtering the entire volume of the estuary daily. Finally, we also examined particle retention efficiency for shrimp, commensal clams, and oysters to examine overlap in particle size selection and thus the potential for food competition. The figure on the lower right demonstrates that while shrimp and their commensal clams seem to partition food based on particle size, shrimp and oysters consume the same sized particles. Use of the same sized food particles and the capability of shrimp to filter the entire volume of the estuary suggest that shrimp are indeed capable of competing with oysters for food. |
 |
 |
The importance of predator dependent foraging for determining consumer population size |
|
Early predation theory assumed that predation rate was a function of prey density only, leading to prey-dependent functional responses. The influence of predator interference on predation rates was subsequently noted, and it has been suggested that prey and predator densities may be equally important so that predation rates are a function of the ratio between the two. Under ratio-dependent foraging, predation rates generally increase with prey density up to some maximum level and decrease with predator density. Whether predator foraging is prey dependent or ratio dependent has ramifications for community and population dynamics. Predator–prey models based on prey dependent foraging predict that increased productivity of basal resources transfers through the food chain, alternately influencing the biomass of each trophic level depending on food chain length. In addition, these abundance changes are predicted to decrease the dynamic stability of the food web (i.e., the paradox of enrichment). In contrast, models based on ratio dependent foraging predict that increasing productivity at the resource level will in turn increase abundance at each level of the food chain, including basal resources, prey, and predators, and will not affect food web stability. There has been considerable debate about which of these two forms of the functional response, prey dependent or ratio dependent, should be used as the basis for predation theory. The general consensus is that we lack sufficient empirical support to determine which is more appropriate. I conducted an experiment with Carcinus maenas and Hemigrapsus sanguineus, two species of invasive crabs found intertidally on the New Hampshire coast, where I measured the functional response of these two species foraging at different densities. This allowed me to examine the appropriateness of the prey dependent and ratio dependent foraging model for each of these species. |
|
The figure on the right is from Griffen and Delaney 2007 and demonstrates that the strength of predator dependence differs for these two crab species. Specifically, interference has a larger impact on foraging by Carcinus than on foraging by Hemigrapsus. The number of mussels consumed by each individual crab decreased for each species as the density of crabs increased. However, this decrease was much more pronounced for Carcinus (as indicated by the clear separation of the curves) than for Hemigrapsus (where the curves continue to overlap. Prey dependence and ratio dependence are really two extremes along a continuum, and predators with different strengths of interference can fall at different places along this continuum. But where a predator falls along this continuum has important implications for predator population size. Simply stated, ecological theory predicts that on the same prey base, predator foraging that is only influenced by prey density can result in larger predator populations than predator foraging that is simultaneously influenced by both prey density and predator density. |
 |
 |
I then fit a model the data in the figure above. This model is a modified version of the Hollings type III functional response model where the prey density (N) is divided by the predator density (P). When the exponent m is equal to 0 (anything raised to the zero power equals 1) this yields the prey dependent, or standard, Hollings type III functional response model. When this exponent is equal to 1, this yields the ratio dependent model. This exponent may take on a range of values depending on the strength of predator interference. I fit this model to the data for each species to estimate this exponent for each species in order to determine where each species falls along the predator dependent spectrum. This analysis indicated that while neither species purely prey dependent or ratio dependent, they do forage on opposite ends of this continuum, with predator interference influencing predation by Carcinus much more than predation by Hemigrapsus. |
The second part of this study was to sample 30 sites throughout the invaded region of these two species. Carcinus was introduced in the early 1800s and spread slowly throughout New England and into Canada. Hemigrapsus was introduced in 1989 in New Jersey and has quickly spread and has replaced Carcinus as the dominant crab in rocky intertidal areas along the way. The figure on the right is from Griffen and Delaney 2007 and demonstrates that while Hemigrapsus dominates in southern areas and Carcinus in northern areas, overall across all of these sites, Hemigrapsus was on average 6 times more dense and had 4 times higher biomass than Carcinus. Two important conclusions come from this study. First, the answer to the prey dependent - ratio dependent debate is likely species specific. The importance of predator dependence can differ strongly, even for two seemingly similar species such as two intertidal crabs. Second, with the advancing invasion of Hemigrapsus, Carcinus is being replaced (click here to learn more about this species replacement). This species replacement is shifting foraging by the most abundant intertidal predator from the ratio dependent end of this spectrum to the prey dependent end of the spectrum. This shift then reduces the effect of predator interference, and as theory predicts, may result in increased population size at the predator trophic level. Therefore, this decrease in predator interference, an individual behavior, may explain at least in part the very high density and biomass of Hemigrapsus at sites that were once dominated by much lower abundances of Carcinus. |
 |
Consumer interference and ideal free distribution in invasive consumers |
|
The distribution of the invasive crab Carcinus maenas on a beach on the New Hampshire coast facilitates optimal foraging as a response to interference between crabs.
Several theories have been proposed to explain how consumers should distribute themselves across habitats based on resource abundance in each habitat. Probably the most influential of these theories is known as the ideal free distribution (IFD). The IFD assumes that as consumer density increases, interference between consumers will cause them to feed out across feeding patches so that the per capita consumption rate in each patch is identical. While intuitively appealing, this theory is based on some assumptions that clearly are not justifiable in most situations. First, the model assumes that consumers have an ideal, or perfect, knowledge of the quality of each available habitat and can make accurate assessments of the consumption rates that could be obtained in each. Second, the model assumes that consumers are free to move into any habitat without cost, or at least at equal cost for each habitat. Third, the model assumes that all consumers are identical in their ability to compete for resources. Several variations on the ideal free distribution have been developed that relax some or all of these assumptions. However, this baseline model is often the starting point for empirical studies on the distribution of consumers. |
|
 |
The picture to the left is a satellite view of Odiorne State Park on the New Hampshire coast. The red dots indicate separate sampling areas. This site is characterized by a series of coves that contain boulders covered by macroalgae. These coves are close enough together that movement between them is possible for crabs during high tide. Further, because of the orientation of this beach, there is a large difference in hydrodynamics across these sites, resulting in large differences in the abundance of mussels that recruit intertidally into the separate coves. Mussels are a favored food items for Carcinus. However, only crabs greater than 20 mm carapace width eat mussels. So when we use this size cutoff and look at the distribution of large and small crabs separately as a function of mussel density at each of these sites, we see that the density of large crabs asymptotically increases with mussel density (shown by the black circles), whereas small crabs don’t seem to be tracking mussel density at all (shown here by the white circles). This means that crabs are likely not settling as juveniles into sites with high mussel abundance, rather they must be aggregating to areas with lots of mussels when they reach the size at which mussel consumption begins. |
I examined whether crabs at this site follow ideal free expectations and if so, what mechanisms might lead to this distribution. This site clearly violates all of the assumptions of the IFD.
Crabs are very adept at sensing prey in their environment, but surely a crab at one end of our string of sites does not know the availability of mussel prey at a site a kilometer away on the other end. Additionally, the model assumes that crabs can freely move to any site, but given the linear array of sites along this beach, surely crabs can more easily move to adjacent sites than to sites at opposite ends. Finally, the IFD assumes that crabs are equally capable of competing for mussel prey. However, crabs display competitive hierarchies, with larger individuals outcompeting smaller individuals for prey. Therefore, if Carcinus distribution does match the IFD, even after violating the assumptions, then the mechanisms proposed by the model must not be required. |
|
I used individual-based simulation models to examine crab distribution. The first model was a deterministic model that met all the assumptions of the IFD. The model output was the density of crabs at each of the 13 coves across this site. The intention of this model was to examine what the IFD of large (>20 mm) Carcinus would be across the 13 coves. The figure on the right is taken from Griffen 2009 and demonstrates that this predicted IFD matches the observed distribution of large Carcinus, indicating that large crabs at this site are in fact distributed according to IFD predictions. |
 |
I then constructed two additional individual-based stochastic simulation models. These models did not assume ideal knowledge or free movement, but were instead built on known foraging patterns and characteristics of Carcinus. The first of these models assumed that crabs were equal competitors, the second assumed unequal competitors based on crab size. I compared these two models to a null (random) model. The results of these models are shown in the figure below, taken from Griffen 2009. These results demonstrate that once behavioral mechanisms of crab foraging are accurately included in the model,
the IFD does a good job of predicting the distribution of Carcinus, even though this species violates the theory’s fundamental assumptions. These results also suggests that Carcinus actively migrates to maximize foraging efficiency and that the processes modeled here
capture essential aspects of these foraging movements. |
|
 |
|
Current work in this area |
|
As stated above, most of my research incorporates aspects of foraging ecology while simultaneously addressing other ecological questions. I am currently investigating the role of individual diet specialization within populations and what role that plays in population dynamics. In addition, I am also examining how individual personality influences foraging and the role of foraging in community dynamics. |
|
