Including Species Interactions in the Design and Evaluation of Marine Reserves: Some Insights from a Predator-prey Model

Conservation of marine species through fisheries management and no-take marine reserves have focused primarily on single species, but such protection may influence the target speciesʼ predators, prey, competitors, or mutualists. Conversely, successful protection may depend on responses of these other species. Empirical data and previous theory indicate that fisheries status and life-history attributes strongly influence speciesʼ responses to protection. Both direct effects and indirect effects of protection (through species interactions) have been documented. A predator-prey model depicting the dynamics of two species in a two-patch habitat (a no-take reserve and a fished area) revealed conditions under which the predator and prey may decline after reserve establishment. Not surprisingly, model results suggest that management scenarios and life-history traits leading to high predator population growth are more likely to produce prey declines following reserve establishment. Interestingly, trade-offs between enhancing predator and enhancing prey occurred at low fishing intensities regardless of the prey and predator life-history traits. At high fishing rates, reserve establishment generally outweighed predation effects and resulted in increased abundance of both predator and prey. Simple spatial models can help determine the range of possible responses of interacting species to protection and can yield some general insights for their management. Marine reserves, portions of the coastline or ocean set aside and protected from fishing and other human uses, have recently received much attention as a means of conserving marine biodiversity and restoring depleted fish stocks (Bohnsack, 1996; Palumbi, 2001). One of the advantages of marine reserves over traditional fishery management is that reserves protect not only target species but also habitat, nontarget species, and the suite of biotic interactions within the protected area (Roberts and Polunin, 1993; Agardy, 1997; National Research Council, 2001). Although this multispecies focus of reserves is frequently discussed, few empirical or theoretical studies have addressed effects of protection on species interactions explicitly. A majority of empirical studies of marine reserves have compared abundance, biomass, or diversity of multispecies assemblages in reserves and unprotected areas or in single areas before and after the establishment of reserves (Boersma and Parrish, 1999; Halpern, 2003). In contrast, few studies have focused on the effects of reserves on ecological processes and interactions underlying community structure and function (Boersma and Parrish, 1999; but see Castilla, 1999; Shears and Babcock, 2002; Fanshawe et al., 2003, and references therein). The bulk of the theory on marine reserves focuses on population dynamics from a single-species perspective (Gerber et al., 2003). Several investigators have evaluated multispecies issues indirectly by comparing effects of reserves on species with different life-history characteristics (e.g., DeMartini, 1993; Attwood and Bennett, 1995; Sladek Nowlis and Roberts, 1999), but their models do not explicitly include species interactions. A review of existing models pertaining to marine reserves (Gerber et al., 2003) found only one modeling approach that focused explicitly on multispecies interactions BULLETIN OF MARINE SCIENCE, VOL. 74, NO. 3, 2004 654 (Walters et al., 1997, 1999; Walters, 2000). These investigators use spatially explicit Ecosim models (Ecospace) to estimate changes in biomass after reserve establishment on the basis of trophic interactions. A key general prediction of Ecospace models is that prey densities tend to be low where predator densities are high, such as within protected areas. Moreover, potential benefits of reserves can be counteracted by high movement rates and by concentration of fishing effort along the reserve edges, which can create prey gradients that attract predators outward from the reserve (Walters et al., 1999). This general model suggests that dispersal, trophic responses, and spatial fishing-effort responses are all likely to reduce the efficacy of small marine reserves (Walters, 2000). The establishment of no-take marine reserves can influence populations of multiple species and the structure of whole communities through a suite of direct and indirect effects. Species may respond differently to protection depending on the intensity of exploitation they are subject to outside the reserve and prior to its establishment, their lifehistory characteristics, and their larval, juvenile, and adult dispersal ability (Gerber et al., 2002). Meta-analyses of studies of fish assemblages from marine reserves revealed that differential responses of fish families and species to marine reserve establishment were correlated with whether taxa were fishery targets, with body size, and with trophic level (Mosqueira et al., 2000; Micheli et al., in press). Fisher and Frank (2002) analyzed 31-yr time series of abundances of over 70 fish species within an area closed to fishing and an adjacent reference area on the Scotian Shelf, Canada. Fish community compositions were significantly different before and after the implementation of the fishing closure, and several species contributed to driving these differences. A preliminary review of life-history attributes for 16 species in this data set indicated that these different trajectories may be related to dispersal ability of the species (Fisher and Frank, 2002). Protection can also influence species indirectly, through its effects on habitat and species interactions. For example, the establishment of a marine reserve designed to protect a particular species may influence that speciesʼ predators, prey, competitors, or mutualists. Indirect effects of predation on community structure, including trophic cascades, have been documented extensively in the marine environment (Pace et al., 1999; Pinnegar et al., 2000), and such interactions can influence all types of communities and can occur as a result of protection in marine reserves (Pace et al., 1999; Pinnegar et al., 2000). The question remains, however, of what combinations of species traits, environmental conditions, and human interventions are more likely to lead to counterintuitive effects of protection, including species declines and losses following the establishment of no-take reserves. As a hypothetical example of how trophic interactions occurring within marine reserves can lead to counterintuitive effects of protection, let us consider two fish species with distinct dispersal abilities; species A has limited dispersal, and species B is highly mobile. The establishment of a reserve might be followed by an increase in abundance and a shift to larger size within the reserve for species A, but not for species B if the size of the reserve does not encompass the spatial extent of its movements and individuals are thus lost from the reserve. If species A and B do not interact, differential responses of species A and B to decreased fishing intensity within the reserve might have been accurately predicted on the basis of information about the life history and dispersal range of each species independently. If species B is prey to species A, however, its populations may undergo further decline because of increased predation intensity within the reserve. In addition, species A may also decline, with some time lag, because of feedback between the two species through the predator functional and numerical responses. Thus, the nonlinMICHELI ET AL.: SPECIES INTERACTIONS IN RESERVES 655 earity introduced by complex biotic interactions may lead to outcomes that are radically different from expectations about reserve management based on single species. Counterintuitive population declines following the establishment of marine reserves may pose trade-offs between the primary conservation and management goals of specific reserves. Many pairs of species valued by fisheries are known to be linked through trophic interactions—e.g., abalone and lobsters; shellfish and crabs, cephalopods and carangid fishes—where successful enhancement of one species or trophic level might result in declines and losses of other valuable or threatened species. For example, sea otters in coastal marine communities of the Pacific Northwest control invertebrate herbivore populations, increasing productivity and pathways through the food web and the structural complexity in the system (Estes et al., 1998), but they prey so heavily on abalone that fewer abalone may be present in reserves with sea otters than in areas where abalone are fished but sea otters are absent (Fanshawe et al., 2003). Conservation and management of multispecies assemblages requires better understanding of and greater predictive ability about how protection may directly and indirectly influence different species under different reserve configurations and environmental conditions. In the study reported here, we used a simple two-patch predator-prey model to examine the combined effects of reserve size and fishing intensity outside reserves on the population dynamics and long-term persistence of a prey and a predator characterized by different life histories and dispersal abilities. This simple heuristic model does not capture all the biological complexities of any two-species system, let alone of assemblages of multiple interacting species, but conceptually simple models can yield qualitative rules for the design and evaluation of reserves that can apply to a broad set of situations and can generate hypotheses to be tested empirically. In addition, these general models can represent an important first step in identifying the key variables and processes to be included in more complex and biologically realistic models. Our approach was to build a two-patch predator-prey model that incorporates two types of parameters. The first set of parameters includes life-history characteristics such as prey growth rate, predator attack rate, and larval dispersal. The second includes management criteria, such as the fishing rates on prey and predators outside the reserve and the fraction of the total area allocated to the reserve. We used this model to explore the implications of marine reserves from the perspective of consumer-resource interactions. In particular, we asked how consumer-resource dynamics vary with (1) the life history and dispersal characteristics of the species, (2) reserve size, and (3) fishing intensity outside the reserve. Our goal was to develop simple predictions about what combinations of these parameters may, in theory, lead to counterintuitive results of reserve establishment, such as decline or loss of one or both species linked though trophic interactions.