Open access publications in the SLU publication database

Abstract

Trophic interactions are fundamental in the functioning of ecosystems. Predator-prey interactions between cod (Gadus morhua), herring (Clupea harengus) and sprat (Sprattus sprattus) in the Baltic Sea are typical examples of strong trophic interactions with consequences on the structure of this ecosystem. In addition to their ecological roles, cod, herring and sprat are also the main targets of the central Baltic Sea fisheries. Transition from a management based on a single species to an Ecosystem-Based Fisheries Management (EBFM), which accounts for trophic interactions, is needed to avoid overexploitation of these species and severe consequences for the ecosystem. The main aim of this thesis was to investigate ontogenetic and temporal patterns of prey selection by cod, using a unique dataset on stomach content, and to incorporate this information with various fisheries-dependent and scientific survey data, into a multispecies model with the potential to support EBFM. Cod stomach data showed clear ontogenetic shifts in the diet, with smaller cod predating mainly on benthos and larger cod on fish (Papers I and II). The multispecies model developed and implemented in this thesis was able to represent the main patterns in the prey species and length selection by cod (Paper I). This allowed to compare prey length selection by cod and the pelagic fisheries and to evaluate the competition between them. Considerable overlap between herring and sprat lengths targeted by cod and the fisheries leads to immediate competition for the same prey. In addition, cod predates on prey sizes smaller than those targeted by the fisheries, thus generating a delayed effect of competition in the form of a loss of potential future biomass available for the fisheries. During certain periods, the loss of future biomass was estimated to be comparable to the amount of biomass suitable for fisheries directly removed by cod predation (Paper III). A comparison of the model implemented in this thesis with two other multispecies models developed for the Baltic Sea on the same species allowed to evaluate their similarities and differences in simulated management scenarios (Paper IV and V). Despite the differences among the models, the results suggested that multi-model inference was still useful to evaluate the robustness of alternative simple fisheries management strategies to different models’ assumptions, which could contribute to the development of an EBFM for the central Baltic Sea fisheries.