Oceanographic coupling across three trophic levels shapes source-sink dynamics in marine metacommunities

摘要

A central goal of metapopulation ecology is to determine which subpopulations have the greatest value to the larger metapopulation. That is, where are the 'sources' that are most essential to persistence? This question is especially relevant to benthic marine systems, where dispersal and recruitment are greatly affected by oceanographic processes. In a single-species context, theoretical models typically identify 'hotspots' with high recruitment, especially high self-recruitment, as having the highest value. However, the oceanographic forces affecting larval delivery of a given species may also influence the recruitment of that species' predators, prey, and competitors. We present evidence from the Virgin Islands and Bahamas that oceanographic forces produce spatial coupling between the recruitment of planktivorous fishes, the recruitment of their predators, and the productivity of their zooplankton prey. We examined the consequences of this type of multi-trophic coupling using a simple analytical population model and a multispecies numerical simulation model with parameter values based on the Virgin Islands system. In both analyses, strong coupling caused planktivores at the highest recruitment sites to experience higher mortality (a consequence of higher predator densities) but faster growth and higher fecundity (a consequence of higher zooplankton densities) than planktivores at low recruitment sites. As such, the relative strength of oceanographic coupling between the three trophic levels strongly determined whether a particular reef acted as a source or sink. In the simulation model, density-dependent competition for zooplankton limited overall metapopulation biomass more severely than predation, so oceanographic coupling between planktivore larval supply and zooplankton productivity had a stronger effect on the metapopulation value of a patch. We argue that the potential for such tri-trophic coupling should be incorporated into future metacommunity models and has considerable implications for the design and evaluation of marine reserves.